Methodology

• Kaptein, M. C., Nass, C., & Markopoulos, P. (2010). Powerful and consistent analysis of likert-type ratingscales. CHI ’10 (pp. 2391-2394). New York, New York, USA: ACM Press. doi: 10.1145/1753326.1753686.

• Wobbrock, J. O., Findlater, L., Gergle, D., & Higgins, J. J. (2011). The Aligned Rank Transform for Nonparametric Factorial Analyses Using Only A NOVA Procedures. CHI 2011 (pp. 143-146).

[Oulasvirta 06]

Oulasvirta, A., & Saariluoma, P. (2006). Surviving task interruptions: Investigating the implications of long-term working memory theory. International Journal of Human-Computer Studies, 64(10), 941-961. doi: 10.1016/j.ijhcs.2006.04.006.

• interruptions have no effect on memory when LTWM encoding speed is fast enough for task processing, regardless of pacing, intensity, difficulty;
• cost of interruptions = memory loss - omissions and errors;
• safeguarding task representations;
• individual and task differences in interruption tolerance;
• if safeguarding has taken place, resumption of a task should be quick and accurate regardless of processing demands of interrupting task;
• current theories of STWM, short-term traces are assumed to decay completely, in a time of about 30s, when no rehearsal can be carried out during interruption;
• theories of working memory based on transient activation of information in LTM cannot explain the resumption of activity once the information in working memory has be irretrievably lost;
• if an interruption takes place, task representation in LTWM remains in an interrupted but intact state; people can use the few seconds available upon task switching to encode some of those contents to LTWM; when encoding skills do not match the processing demands assumed by the task environment, resumption of the task requires alternative strategies, searching for cues in the environment or generating them based on semantic memory on the task;
• benefit of interruption lag / opportunity to rehearse disappeared with task practice;
• processing demands of interruption task does not affect memory for main task (in contrast to [Gillie 89]); rehearsal used as maintenance strategy; retrieval structures not used for meaningless stimuli;
• no evidence for transient activation, refreshing activation during interruption lag / interruption
• length of interruption can matter i conditions of massive interference and where rehearsal is rules our by experimental manipulation;
• memory load / task phase are not the factors determining disruptiveness to interruptions although they do coincide with it;
• extreme time pressure may lead to selection of stimulus-specific processes at the expense of processes relying on a more general representation;
• semantic access in contrast to surface access;

• Iqbal, S. T., & Bailey, B. P. (2010). Oasis. ACM Transactions on Computer-Human Interaction, 17(4), 1-28. doi: 10.1145/1879831.1879833.

• Bogunovich, P., & Salvucci, D. (2011). The Effects of Time Constraints on User Behavior for Deferrable Interruptions. Read, 3123-3126. Retrieved May 16, 2011, from http://www.cs.drexel.edu/~salvucci/publications/Bogunovich-CHI11.pdf.

• Dabbish, L., Mark, G., & González, V. M. (2011). Why do i keep interrupting myself?: environment, habit and self-interruption. CHI 2011 (p. 3127–3130). ACM. Retrieved May 16, 2011, from http://portal.acm.org/citation.cfm?id=1979405.

• Uttl, B. (2008). Transparent meta-analysis of prospective memory and aging. PloS one, 3(2), e1568. doi: 10.1371/journal.pone.0001568.

• Hasher, L., Stoltzfus, E. R., Zacks, R. T., & Rypma, B. (1991). Age and inhibition. Journal of experimental psychology. Learning, memory, and cognition, 17(1), 163-9. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1826730.

• Kim, S., Hasher, L., & Zacks, R. T. (2007). Aging and a benefit of distractibility. Psychonomic bulletin & review, 14(2), 301-5. Retrieved from http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2121579&tool=pmcentrez&rendertype=abstract.

• Lindenberger U, Baltes PB. Sensory Functioning and Intelligence in Old Age: A Strong Connection. Psychology and Aging. 1994;9(3):339-355. Available at: http://www.sciencedirect.com/science/article/B6WYV-46SGK21-1/2/6b37eac862452ac3771ca27846debd83.

• Craik, F. I., Byrd, M., & Swanson, J. M. (1987). Patterns of memory loss in three elderly samples. Psychology and aging, 2(1), 79-86. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3268196.
• Cabeza, R. (2004). Task-independent and Task-specific Age Effects on Brain Activity during Working Memory, Visual Attention and Episodic Retrieval. Cerebral Cortex, 14(4), 364-375. doi: 10.1093/cercor/bhg133.

• Cabeza, R., Anderson, N., Locantore, J., & Mcintosh, A. (2002). Aging Gracefully: Compensatory Brain Activity in High-Performing Older Adults. NeuroImage, 17(3), 1394-1402. doi: 10.1006/nimg.2002.1280.

• Einstein, G. O., & McDaniel, M. A. (1997). Aging and Mind Wandering: Reduced Inhibition in Older Adults? Experimental Aging Research, 23(4), 343-354. doi: 10.1080/03610739708254035.

• West, R. L. (1996). An application of prefrontal cortex function theory to cognitive aging. Psychological bulletin, 120(2), 272-92. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8831298.
  • [West 96] studied PFC change with age and effect on RM, PM, interference control/suppression of distractions, response inhibition (go/no-go), recall and recognition memory; evidence from patients with frontal lobe damage - neural correlates of distraction;

• Gluck, J. S. (2006). An investigation of the effects of matching attentional draw with utility in computer-based interruption. Citeseer. Retrieved February 1, 2011, from http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:An+Investigation+of+the+Effects+of+Matching+Attentional+Draw+with+Utility+in+Computer-Based+Interruption#0.

• Oulasvirta, A. (2005). Interrupted cognition and design for non-disruptiveness: The skilled memory approach. CHIʼ05 extended abstracts on Human factors in computing systems (p. 1124–1125). ACM. Retrieved January 30, 2011, from http://portal.acm.org/citation.cfm?id=1056843.

• Oulasvirta, A., & Ericsson, K. A. (2009). Effects of Repetitive Practice on Interruption Costs : An Empirical Review and Theoretical Implications. European Conference on Cognitive Ergonomics: Designing beyond the Product—Understanding Activity and User Experience in Ubiquitous Environments (p. 1–9). VTT Technical Research Centre of Finland. Retrieved January 30, 2011, from http://portal.acm.org/citation.cfm?id=1690508.1690548.

• Oulasvirta, A., & Saariluoma, P. (2004). Long-term working memory and interrupting messages in human - computer interaction. Behaviour & Information Technology, 23(1), 53-64. doi: 10.1080/01449290310001644859.

• Oulasvirta, A., & Saariluoma, P. (2006). Surviving task interruptions: Investigating the implications of long-term working memory theory. International Journal of Human-Computer Studies, 64(10), 941-961. doi: 10.1016/j.ijhcs.2006.04.006.

• Oulasvirta, A., & Salovaara, A. (2004). A cognitive meta-analysis of design approaches to interruptions in intelligent environments. Extended abstracts of the 2004 conference on Human factors and computing systems - CHI ’04, 1155. New York, New York, USA: ACM Press. doi: 10.1145/985921.986012.

[Hoey 10]

Hoey, J., Leuty, V., & Mihailidis, A. (2010). A Tool to Promote Prolonged Engagement in Art Therapy : Design and Development from Arts Therapist Requirements Categories and Subject Descriptors. ASSETS 2010.

• abstract : a novel tool to increase the capacity of creative arts therapists to engage cognitively impaired older adults in creative activities (touch screen interface)
• uses decision theory to reason about what actions the agent can take to optimize over a user-specified utility function; abstract model of automated assistance based on the partially observable Markov decision process;
• possible to give non-technical users control over an intelligent assistant;
• inertia can be provided by the device to help with loss of focus - a therapist's inertial tool;
• techniques for tracking engagement and mood; use of facial expressions and gaze;
• defines categories of system actions or prompts based on their influence on engagement, their cost ti the user in terms of interruption and disturbances; interactivity of an action as the amount of involvement it requires from a user; highly interactive vs. mildly interactive; trade-off is that a very interactive prompt may get a disengaged user involved, but may be a disruptive action for an already engaged user, causing them to disengage;
• the POMDP tracks a belief state, a continuous measure of a person's engagement; if they are responsive to interactivity level of the action, effect of the action is to increase their engagement; PODMP must trade-off involvement and engagement against interactivity; classified into levels of involvement: interactive, active, intermittent, inactive; engaged can be yes/no/confused; respond can be yes or no;
• therapists showed approval of mapping the generic interactivity levels to the specific application prompts;
• therapists requested ability to pause the AI decision making process or experiment wit different settings throughout a single session; some felt uncomfortable trusting a computer to accurately measuring engagement;
• request for multi-touch functionality, intuitive interfacing system would allow clients to use the system with much more ease and less anxiety;
• request for tracking biological data;
• customizability: clients may externalize their engagement at different levels; one can adjust/tweak parameters the passivity level of the system, the activity level of the client, and the eye-contact level of the client;

[Monk 08] IDRG [10.19.10]

Monk, C. A., J. G. Trafton, and D. A. Boehm-Davis. “The effect of interruption duration and demand on resuming suspended goals..” Journal of experimental psychology. Applied 14, no. 4 (December 2008): 299-313. http://www.ncbi.nlm.nih.gov/pubmed/19102614.

• abstract: duration and demand of interruptions / importance of goal rehearsal in mitigating delay

• [Altmann 02] - memory for goals model for studying interruptions - activation decay, longer task resumption times
  • interrupting task: mood checklist over a ToL main task
• [Hodgetts 06b] support for duration effect
• [Gillie 89] no support for duration effect; similar w/ [McDaniel/Einstein] - DV: performance on ProM task - digit monitoring interruption task
• global measures in interruption studies insensitive to goal resumption and resumption time; interruption durations in past experiments may have been reasonable in terms of face validity but may have masked duration effects (39s > 2.75min vs 3 > 18s)
• interruption demand also a factor / inhibiting goal maintenance > longer resumption times
• [Zijlstra 99] - document editing main task w/ menial interruptions (i.e. phone # lookup)
• [Cades 07] n-back better than shadowing for inhibiting goal maintenance;
• [Hodgetts 06b] addition (simple vs complex) interrupting tasks
• "complexity" overloaded term, replace w/ demand, processing demands on working memory that prevent or allow rehearsal of suspended task goals

• experiments: does memory for goals decay? with interleaved interruptions? requires a primary task w/ many subgoals to be performed linearly (i.e. VCR programming task)

• Exp 1: pursuit-tracking interruption task in split screen; 3 interruption durations
• errors: deviations from optimal path, potential for speed/accuracy tradeoff - no sig. difs
• both tasks equally important
• resumption lag compared against IAI - interaction intervals in uninterrupted condition - viewed as appropriate comparison
• main effect of interruption duration found - rationale: sensitive DVs and manipulation of duration
• performance on interruption task worse in short-duration interruption

• Exp 2: fitting resumption lags to log function of memory for goals model
• explains absence of duration effect in [Gillie 89]
• is 13s > 23s p[oint of asymptote?
• no effect of duration of tracking task performance
• implication: interleave tasks - strive to shift attention at least every 15s for optimal resumption times

• Exp. 3: 3 duration x 3 demand mixed design w/ interruption demand as BS factor;
• conditions: no interrupting task, tracking task, verbal n-back test
• sig. main effect of demand b/w n-back condition and others
• sig. main effect of duration in no-task and n-back conditions (not in tracking task - approaching sig.)
• linear contrasts sig. b/q n-back and no task
• decay will happen however we can mitigate its effect to some extent - steeper lags in longer, high demand conditions

• discussion: reminders for primary task state interleaved with interruptions
• shallow vs. deep goal rehearsal
• decay vs. interference models?
• more frequent interruptions > shorter resumption times?
• practical implications: flying, driving, ERs, IM, loss of productivity
• conclusion: mitigating decay processes by rehearsing goals during interruptions

• discussion points:

• in HCI research/applied scenarios- how long are typical interruptions? what other interactions are possible - similarity, social obligations, modality, interference
• face/ecological validity of past interruption studies and their tasks used
• my research: considering similarity as in [Gillie 89], modality as in [Storch 92]
• IAI vs ToT. what DVs are most ecologically valid and account for Zeigarnik effect? reflects long-term effects on productivity?
• how do we measure complexity in HCI research? can it always be empirically measured or do we pilot until several levels of complexity are determined?
• main tasks: nonlinear main tasks? programming and visuospatial tasks? puzzles?
• does log function of decay hold for older adults?
• non-computerised interruptions? effect of modality - interaction with age (cohort effect)
• how to include reminders of main task / ensure deep rehearsal in UI design? which environmental cues to use? recency, frequency
• social obligations of interruptions

• Carton, A. M., and J. R. Aiello. “Control and Anticipation of Social Interruptions: Reduced Stress and Improved Task Performance 1.” Journal of Applied Social Psychology 39, no. 1 (January 2009): 169-185. http://doi.wiley.com/10.1111/j.1559-1816.2008.00434.x.

• Harr, R., and V. Kaptelinin. “Unpacking the social dimension of external interruptions.” Proceedings of the 2007 international ACM conference on Conference on supporting group work - GROUP ’07, no. 1 (2007): 399. http://portal.acm.org/citation.cfm?doid=1316624.1316686.

• McFarlane, D. “Comparison of Four Primary Methods for Coordinating the Interruption of People in Human-Computer Interaction.” Human-Computer Interaction 17, no. 1 (March 2002): 63-139. http://www.informaworld.com/openurl?genre=article&doi=10.1207/S15327051HCI1701_2&magic=crossref||D404A21C5BB053405B1A640AFFD44AE3.

• Monk, C. A., J. G. Trafton, and D. A. Boehm-Davis. “The effect of interruption duration and demand on resuming suspended goals..” Journal of experimental psychology. Applied 14, no. 4 (December 2008): 299-313. http://www.ncbi.nlm.nih.gov/pubmed/19102614.

• Palanque, P., J. Ladry, E. Barboni, D. Navarre, and M. Winckler. “Une approche formelle pour i’evaluation de la tolérance aux interruptions des système interactifs.” Proceedings of the 21st International Conference on Association Francophone d’Interaction Homme-Machine - IHM ’09 (2009): 141. http://portal.acm.org/citation.cfm?doid=1629826.1629848.

• goal: design guidance for explicitly designing successful interruption mgmt
• summarises approaches to manage interruptions in multitasking environments;
• [Cohen 80] found that unpredictable and uncontrollable interruptions induce personal stress that can negatively affect performance after interruptions;
• [Speier 97] found a negative relationship b/w interruption freq and human performance on complex tasks;
• people show a measurable difference in their cognitive style relative to multitasking, called field dependence-independence; conservation tasks and reversal shift tasks predict multitasking performance;
• [Latorella 96b,98]: visual interruptions of auditory tasks resulted in slowest performance times in starting the interrupting task; auditory interruptions of auditory tasks resulted in the most errors on procedural tasks, visual interruptions of visual tasks resulted in best overall performance during interruptions; auditory interruptions of visual tasks resulted in most errors on interruption tasks;
• [Gillie 89]: disruptive effects of interruption on peoples' memories were not caused by an inability to rehearse memory prior to handling an interruption, negative effect was caused by memory interference created by interruption tasks that were complex or similar to pre-interruption task;
• interruptions cause initial decrease in how quickly people can perform post interruption tasks, reduce peoples' efficiency, increase stress;
• [Hess 94] found that training can suppress negative effects of interruption;
• [Malin 91] said that UI should be design to reorient users to previously interrupted activities when they try to resume them; users can explicitly mark the occurrence of an interruption; computer can then generate appropriate recovery support; [Rouncefield 94] also found marking as a useful strategy;
• five basid strategies to improve human performance on interrupt-laden task: training, incentives, personnel selection, completely replace person with automation, design HCI support;
• UI support for after switch phase:
  • enhance memory of interruption position by external markers or by allowing rehearsal; provide overview status of background tasks; a summary of amount of time spent away from original task;

• Rouncefield M, Viller S, Hughes J, Rodden T. Working with “constant interruption”: CSCW and the small office. The information. 1995;11(3):173-188. Available at: http://www.informaworld.com/index/921631428.pdf.

• Hess SM, Detweiler MC. Training to Reduce the Disruptive Effects of Interruptions. Human Factors and Ergonomics Society Annual Meeting Proceedings. 1994;38:1173-1177(5). Available at: http://www.ingentaconnect.com/content/hfes/hfproc/1994/00000038/00000018/art00006. (citation only)

• Latorella KA. Investigating Interruptions: An Example from the Flightdeck. Human Factors and Ergonomics Society Annual Meeting Proceedings. 1996;40:249-253(5). Available at: http://www.ingentaconnect.com/content/hfes/hfproc/1996/00000040/00000004/art00023. (citation only)

• Speier C, Valacich JS, Vessey I. The effects of task interruption and information presentation on individual decision making. In: Proceedings of the 18th International Conference on Inormation Systems. New York, NY, USA; 1997:21-36. Available at: http://portal.acm.org/citation.cfm?id=353071.353080.

• Hasher L, Zacks RT. Working Memory, Comprehension, and Aging: A Review and a New View Motivation GH. Psychology of Learning and Motivation. 1988;Volume 22:193-225. Available at: http://www.sciencedirect.com/science/article/B7J17-4RY6MF2-8/2/ff33e1ddf97b219c623fbbc1c13ebb70. (citation only)

• Lindenberger U, Baltes PB. Sensory Functioning and Intelligence in Old Age: A Strong Connection. Psychology and Aging. 1994;9(3):339-355. Available at: http://www.sciencedirect.com/science/article/B6WYV-46SGK21-1/2/6b37eac862452ac3771ca27846debd83. (citation only)

• Hasher L, Zacks RT. Working Memory, Comprehension, and Aging: A Review and a New View Motivation GH. Psychology of Learning and Motivation. 1988;Volume 22:193-225. Available at: http://www.sciencedirect.com/science/article/B7J17-4RY6MF2-8/2/ff33e1ddf97b219c623fbbc1c13ebb70.

• Lindenberger U, Baltes PB. Sensory Functioning and Intelligence in Old Age: A Strong Connection. Psychology and Aging. 1994;9(3):339-355. Available at: http://www.sciencedirect.com/science/article/B6WYV-46SGK21-1/2/6b37eac862452ac3771ca27846debd83.

On older adults and cognitive ageing, interruptibility and distractibility

[Anderson 00]

Anderson JR. Cognitive Psychology and its Implications. 5th Edition. New York, NY, USA: Worth Publishers; 2000:531.

• changes in cognition occur throughout adulthood: cognitive ability does not uniformly increase with added years; [Salthouse 92] used WAIS-R intelligence tests (verbal intelligence: vocabulary and language comprehension - maintained throughout adulthood); dramatic decrease in performance component (reasoning and problem solving); such tests typically speeded (indicative of slower response time); this could be environmental (older adults do better at tests related to their daily work (accumulation of knowledge), while younger adults who are used to tests do better on speeded tests;
• age-related declines in brain function; brain cells gradually die and some areas are particularly susceptible; hippocampus (memory) loses 5% of its cells every decade; other cells shrink and atrophy; some evidence also for compensatory growth;
• peak of intellectual performance in one's mid-thirties [Lehman 53] studies of works of artists, scientists, philosophers;
• [Salthouse 92] claims loss of working memory w/ age; slower in information processing than younger adults, thus inhibiting ability to maintain information in working memory;

PSYC 322 notes

• theories of cognitive ageing
  • [Salthouse 96] - processing speed thoery
    • changes in PS underlie other cog. changes; we have a limited WM capacity regardless of age, ageing compromises processing efficiency (resources not smaller, longer processing time) takes longer to complete multi-step problems if earlier steps take longer, constraining later steps (simultaneity) - info from earlier steps may have decayed at later steps; exposed with timed tasks;
    • measures of PS: digit symbol coding, simple reaction time test (avg 210ms - slows with age), Ravan's progressive matrices (reasoning)
    • evidence for PS w/ longitudinal studies and negative correlation on these measures
  • [Craik 82] - working memory theory
    • self-initiated processing declines in older age, manifested in working memory
    • measures: backwards digit span, n-back test (verbal or auditory) (predicts WM, intelligence, school performance), Corsi block task (spatial task) (http://cognitivefun.net
    • is this a cohort effect? familiarity w/ computers and games?
    • responses to survey questions w/ age shows greater primacy and recency effects for MC questions;
    • relying on priming improves WM performance; additional context improves recall, complexity of sentences and sentence/picture combination improves recall (more info available) [Park 90], [Cherry 96]
  • [Hasher 88] - inhibition theory (distractibility and interruptibility)
    • ability to shield out distractions/irrelevant information w/ age (info gateway); displacement occurs - displacing what is to be remembered in STM (i.e. remembering phone number while someone shouts random #s at you)
    • go-no go test: more likely to make mistakes when distracted, must inhibit response; older adults less slow after inhibition, more likely to make errors, unable to suppress other factors, reduced impulse control; more likely to speak impulsively;
  • [Lindenberger 94] - sensory function
    • visual/auditory acuity drops > reduced processing speed, crude measure of brain integrity / cog. resource; higher education > higher cog. resource;
    • able to compensate w/ decline in resources; negative correlation w/ age: visual and auditory acuity and other declines: perceptual speed, reasoning, memory, knowledge, fluency, intelligence

• perception of time [James] - influence of memory and novel activities - good memory of activity > perception of time passing slowly, otherwise time is perceived as moving quickly;

• attention and distractibility
  • drops in divided but not necessarily sustained attention ; inhibition theory of memory relevant here; unable to suppress distractions and enhance foci; compromised attention influences PS and memory
  • enhancement and suppression [Gazzaley 05] study w/ places and faces; older adults less able to suppress, less able to enhance;
  • PFC responsible for enhancement and suppression, goal activation, attentional modulation - all change w/ age
  • [West 96] studied PFC change with age and effect on RM, PM, interference control/suppression of distractions, response inhibition (go/no-go), recall and recognition memory; evidence from patients with frontal lobe damage - neural correlates of distraction;
  • measures of distractibility: difficult to design studies to measure this; hard to design training programs for mitigating distractions;

more...

• Salthouse TA. The processing-speed theory of adult age differences in cognition. Psychological Review. 1996;103(3):403-28. Available at: http://www.ncbi.nlm.nih.gov/pubmed/8759042.

• Vance DE, Keltner NL, McGuinness T, Umlauf MG, Yuan Y. The Future of Cognitive Remediation Training in Older Adults. Journal of Neuroscience Nursing. 2010;42(5):255-264. Available at: http://content.wkhealth.com/linkback/openurl?sid=WKPTLP:landingpage&an=01376517-201010000-00004.

Interruptions and Memory

[McFarlane 02]

McFarlane D, Latorella K. The Scope and Importance of Human Interruption in Human-Computer Interaction Design. Human-Computer Interaction. 2002;17(1):1-61. Available at: http://www.informaworld.com/openurl?genre=article&doi=10.1207/S15327051HCI1701_1&magic=crossref||D404A21C5BB053405B1A640AFFD44AE3.

On Prospective memory (PM) and interruption/distraction, age-related issues

• McDaniel MA, Einstein GO, Graham T, Rall E. Delaying execution of intentions: overcoming the costs of interruptions. Applied Cognitive Psychology. 2004;18(5):533-547. Available at: http://doi.wiley.com/10.1002/acp.1002.

• Einstein GO, McDaniel MA, Williford CL, Pagan JL, Dismukes RK. Forgetting of intentions in demanding situations is rapid. Journal of experimental psychology. Applied. 2003;9(3):147-62. Available at: http://www.ncbi.nlm.nih.gov/pubmed/14570509.

• related work: distraction conflict theory - influence of distractions on decision performance [Baron 86]; distractions are detected by a different sensory channel from those of primary task and may be ignored or processed concurrently with primary task; interruptions cannot be ignored (same sensory channel) - capacity and structural interference - capacity: # of incoming cues greater than what can be processed; structural interference: decision maker must attend to 2 inputs that require the same physiological mechanisms; distractions facilitate performance on simple tasks, inhibit on complex tasks (# of cues increases), decision maker's excess cog. capacity decreases, when processing complex tasks, people minimise expenditure of scarce cog. resources, uncritically examining both relevant and irrelevant cues;
• models of interruptions: interruption framework: cognitive processing (freq, duration, content, complexity, timing of interruption), social characteristics of interruption (form, generator, social expectations), processing mechanisms (sequential, preemptive, simultaneous);
• current focus on computer-generated interruptions; decision-making tasks using symbolic or spatial cues;
• theory of cognitive fit - how information presentation formats support decision making tasks; analytical processes used most effectively to act on symbolic formats and tasks; perceptual processes used most effectively on spatial formats and tasks;

• experiment: 2 (BS: interruptions, no interruptions) x 2 (BS: table, graph format) x 2 (WS: simple, complex tasks) x 2 (WS: symbolic, spatial tasks) design; unlimited time to complete tasks;
• to control for fatigue: those in non-interruption group performed all interruption tasks before or after experimental tasks;
• interruptions were information acquisition tasks (both spatial and symbolic); main tasks were aggregate planning tasks (spatial) and facility location task (symbolic) - both tasks involved examining cues and performing calculations and interrelated planning;

• results: MANCOVA - all main effects significant (interruptions, format, task complexity, task type);
• work environment: decision accuracy with interruptions for simple symbolic tasks higher than with interruptions, difference not sig., sig less time to complete with interruptions
  • accuracy higher for sim-spat sig. higher w/ interruptions, sig. less time to complete than without interruptions
  • above reversed for complex spatial and complex symbolic (differences significant)
• no interaction effect for accuracy or time between work environment (interruptions vs. no interruptions) and information presentation (tables, graphs) for simple tasks; there is an interaction for complex symbolic tasks: graphs resulted in faster decision times for complex symbolic tasks, with and w/o interruptions; spatial formats will best support complex-spatial tasks with and without interruptions;
• post hoc perceptions: interruption/table condition found amount of information more overwhelming; domain expertise also explained differences in decision accuracy;

• discussion: hypothesised a crossover effect for performance with tables on complex symbolic tasks; interruptions do influence the manner in which information is processed; effect of interruptions on complex-symbolic tasks is manifested in decision accuracy but not decision times; interruptions influence the manner in which problem solvers perceive and process information when performing complex intelligence decision tasks; for simple tasks, interruptions focus a decision maker's attention on important cues resulting;
• limitations: generalisability, interruptions used here were devoid of social characteristics;
• implications: interrupted work environments lead to lower work quality and reduced speed on complex intellective tasks; interruptions perceived negatively despite facilitation of simple tasks; work hassles and interruptions result in more negative moods and increased fatigue;
• there is evidence of a crossover point where analytical and perceptual processing is equivalent; presenting information in ways that enhance the use of perceptual processes facilitates the acquisition and processing of complex information; use graphical formats to minimise the effect of interruptions on complex decision making tasks;

• conclusion: decision makers who are interrupted when solving complex symbolic tasks are better supported by graphs than by tables;

• intro: [Kreifeldt & McCarthy]: memory load at the time of interruption is not a crucial factor in determining whether or not it will be disruptive;

• intro: [Kreifeldt & McCarthy]: memory load at the time of interruption is not a crucial factor in determining whether or not it will be disruptive;
• disruptiveness of an interruption may vary from one interface to another; and between types of interruptions;
• current experiment: interruptions dissimilar to main task (answering simple questions); 2 types of interfaces (BS variable - GUI vs CUI); believed that GUI and mouse interaction is more complex and subject to greater disruption;
• commonly believed that telephone interruptions are most disruptive;

• experiment: subjects in a mock-up office with furniture, door, phone;
• results: screen interruption was disruptive, walk-in interruption may have been disruptive, telephone interruption was not disruptive;
• # of correct entries tended to be higher in the CUI, but difference was nonsignificant;
• number of entries less sig. less after screen interruption (compared to control), other 2 conditions did not differ sig. from control; performance sig. less after screen interruption compared to telephone interruption, did not differ sig. from walk-in interruption;
• telephone interruption caused subjects to look offscreen after interruption more than other interruptions or control; telephone 80% > walk-in 75% > control 74% > screen 68%
• looking at screen and entering correct data related un such a way that a lower # of entries corresponds to looking more at screen, especially after a screen or telephone interruption
• questionnaire: CUI subjects thought that on-screen interruption reduced rate of working more than other interruptions, GUI subjects notices no difference (contrary to performance)
• subject performance during post-interruption period decreased in relationship to how much the subject looked at screen;

• discussion: could duration of interruption be a factor in its disruptiveness (all subjects required to respond to interruptions); longer interruptions weakly correlated with more disruption; unlike [Gillie 89], disruptions in this experiment were random rather than at a fixed place in the sequence of events;
• screen interruption differed from phone and walk-in in that it prevented completion of in-progress entry;
• GUI made it easier to work through interruptions during phone and walk-in interruptions;
• [Czerwinski 91]: better recall of primary task when interruption was dissimilar in format to primary tasks;
• current result contrary to popular belief that telephones are most disruptive fro computer users;
• similarity in modality of screen interruption to main task may have also factored into disruptiveness;

• abstract: influence of interruptions on different types of decision-making tasks, and ability of info. presentation formats to alleviate them; interruptions facilitate performance on simple tasks, inhibiting performance on complex tasks; spatial presentation formats able to mitigate effect of interruptions while symbolic formats were not;

• abstract: looked at how unexpected interruptions disrupt computer user performance, whether there is a significant effect due to style of user interface (GUI vs CUI); performance in data entry of personnel database, 3 different forms of interruption (telephone, visitor, on-screen interruption - subject must answer simple question); eye motion data analysed in terms of amount of time and avg. duration of eye fixations on screen; result: screen interruption was disruptive, telephone was not, walk-in intermediate disruption; lower performance directly related to looking more at screen; duration of interruption weakly correlated to performance; another possible explanation for disruptiveness of screen interruption vs. telephone interruption was similarity by modality; GUI makes it easier to work through non-similar (phone, walk-in) interruptions;

[Speier 03]

Speier C, Vessey I, Valacich JS. The Effects of Interruptions, Task Complexity, and Information Presentation on Computer-Supported Decision-Making Performance. Decision Sciences. 2003;34(4):771-797. Available at: http://doi.wiley.com/10.1111/j.1540-5414.2003.02292.x.

[Storch 92]

Storch N. Does the user interface make interruptions disruptive? A study of interface style and form of interruption. In: Posters and short talks of the 1992 SIGCHI conference on Human factors in computing systems. ACM; 1992:14. Available at: http://www.interruptions.net/literature/Storch-DE92011295.pdf.

• Baron RS. Distraction-Conflict Theory: Progress and Problems. In: Berkowitz L Advances in experimental social psychology.Vol 19. Advances in Experimental Social Psychology. Academic Press; 1986:1-40. Available at: http://www.sciencedirect.com/science/article/B7J09-4S865F1-5/2/69e6c1d7028b406ec123b89638447fd2.

• Cohen S. Aftereffects of Stress on Human Performance and Social Behavior: A Review of Research and Theory. Psychological Bulletin. 1980;88(1):82–108. Available at: http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Aftereffects+of+Stress+on+Human+Performance+and+Social+Behavior:+A+Review+of+Research+and+Theory#0.

• Kirmeyer SL. Coping with competing demands: interruption and the type A pattern. The Journal of applied psychology. 1988;73(4):621-9. Available at: http://www.ncbi.nlm.nih.gov/pubmed/3209581.

• O'Conaill B, Frohlich D. Timespace in the workplace: Dealing with interruptions. In: Conference companion on Human factors in computing systems. ACM; 1995:263. Available at: http://portal.acm.org/citation.cfm?id=223355.223665.

• Czerwinski M, Chrisman S, Schumacher B. The Effects of Warnings and Display Similarity on Interruption in Multitasking Environments. SIGCHI Bulletin. 1991;23(4):38-39.

• introduction: success of goal retrieval may be dependent of processes of decay or interference in memory; there is lack of consensus as to what makes interruptions disruptive;
• ACT-R traditional goal-stack construct: goals may be subject to the same limitations as ordinary declarative memory elements; participants do not choose to engage in any costly preparatory processes at the point of goal suspension; activation decay may also be responsible for forgetting in goal memory; goals suspended for longer were found to be more time consuming to retrieve;
• [Altmann Trafton 02] associative activation; processes operating at goal suspension and resumption; activation of a goal in this model is dependent on not only the length of time since that goal was last sampled (base-level activation) but also environmental factors; goal repeatedly sampled or strengthened, overcoming proactive interference; subsequent retrieval of this goal aided by priming; associative links must be formed b/w cue and target goal before it is suspended; rehearsal of current goal before it is suspended will increase base-level activation so that it is later more easily reactivated in memory (during interruption lag)
• current work examines preparatory processes at a more fine-grained level to test specific predictions rooted in cog. theory;
• in some studies that encouraged participants to rehearse goals or to make notes during interruption lag - performance actually worsened due to increased interference;
• prospective goal encoding during interruption lag;
• empirical studies concerning role of retrieval cues in interruption recovery are inconclusive;
• a long lag with no primary task information reduced typical performance owing to boredom and perhaps the activation of irrelevant thoughts;
• role of preparatory processes in supporting memory for task goals is an issue that warrants further investigation

• experiments: [Altmann/Trafton 02]: cues linking target goal to features in current context are encoded relatively automatically simply by co-occurrence;
• presence of an interruption lag manipulated by the transition in tasks;
• research suggests that participants can efficiently preplan up to two subgoals ahead at the beginning of a trial, and then execution of this plan is supported by a process of online monitoring and updating;
• errors in goal retrieval may be more frequent in those conditions in which insufficient contextual cues render the suspended goal less active than competing distractors
• experiment 1a: interruption of the ToL task did not appear to affect accuracy of problem solving:
  • subject to a 2 (interruption vs. control) x 2 (full-screen vs. corner interruption) RM-ANOVA; sig. main effect of interruption (takes longer in sessions following interruption); main effect of interruption type; sig. interaction; resumption of primary task faster when interruption task in corner; regardless of whether trial was completed in min of 6 moves
  • time spent on interruption was non-significant
  • task resumption times a lot quicker than the initial planning times, suggesting that some residual knowledge survives the interruption and that participants are actually retrieving old goals rather than simply planning anew;
  • in most computer-based interruption studies, the primary task is not visible during the secondary tasks; on-screen interruptions were more disruptive than telephone calls or walk-in visitors [Storch 92], result attributed to task similarity; perhaps performance suffers less when participant is allowed more control over integrating of the two tasks;
• experiment 1b: introduces 3rd condition: full-screen interruption preceded by 2-second pause with display freeze;
  • sig. effect of interruption condition obtained; 2s pause followed by full-screen interruption no different from corner interruption, sig. faster than full-screen interruption with no pause; determining the ease of goal retrieval is the availability of task-related cues at the specific point of goal suspension, demonstrates importance of interruption lag; even the briefest opportunity to encode associative cues and/or boost the activation of the target goal is beneficial;
• experiment 2A: how is performance affected if changes are made to the visual display after the interruption? one would expect task resumption to be impeded if these cues are no longer available
  • manipulated whether the display that participants returned to was the one they had left at the point of interruption or whether it was different and would therefore not prime retrieval (changing colours of disks in ToL task);
  • sig. differences b/w conditions; no difference in duration of interruption; quickest with no interruption, slower w/ interruption, slower if colours changed after break; changes in contextual cues at point of task resumption cause disruption to goal retrieval; possible that any change to visual display would disrupt performance and that disks are not necessarily critical cues to goal retrieval;
• experiment 2B: changing background peg colour and colour of pegs upon task resumption - no difference b/w conditions of peg changing and no changing; subjects warned that colours may change upon task resumption; the disks themselves act as cues;
• experiment 3: presence of an interruption lag to encode task cues before goal suspension should be beneficial only if these cues (disk colours) are still available and unchanged at task resumption;
  • 2 (display change) x 2 (interruption lag) RM ANOVA - sig. effect of display change w/ participants slower to resume the task when the colours of disks had been rearranged; main effect of interruption lag did not reach significance; sig. interaction b/w interruption lag and display change: participants sig. quicker to resumer task when the display was not subsequently changed, when colours of disks changed at resumption, benefit of interruption lag was removed; opportunity to encode retrieval cues before onset of an interruption will be useful only if the cues are still available after interruption to prime goal retrieval; time spent on interruption revealed no sig difference b/w conditions;
  • one must be cautious in concluding that the longer interruption lag reduced resumption times further;

• discussion: lag before the onset of the second activity found to reduce task resumption times by allowing an opportunity for participants to prepare goals prior to suspension; preparatory time is needed for encoding of retrieval cues, goal retrieval impeded when changes made to returning visual display; no benefit of interruption lag if potential retrieval cues had been altered;
• findings difficult to accommodate within ACT-R; changes to visual display would perhaps affect ease of goal reconstruction;
• context - at points of both goal suspension and retrieval;
• longer interruption lags more beneficial; what is cost-benefit ratio?
• less intrusive interruptions => quicker resumption times;

• Storch N. Does the user interface make interruptions disruptive? A study of interface style and form of interruption. In: Posters and short talks of the 1992 SIGCHI conference on Human factors in computing systems. ACM; 1992:14. Available at: http://www.interruptions.net/literature/Storch-DE92011295.pdf.

[Altmann 04]

Altmann EM, Trafton JG. Task interruption: Resumption lag and the role of cues. Citeseer; 2004. Available at: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.4.5438&rep=rep1&type=pdf.

• abstract: examined role of external cues associated with interrupted task on resumption lag; cues available immediately before an interruption facilitate performance immediately afterwards - people deploy preparatory perceptual and memory processes, apparently spontaneously, to mitigate disruptive effects of task resumption

• interruptions have been found to have detrimental effects on situational awareness in dynamic task environments; studies in this domain typically focus on "switch cost"; current focus on higher-level tasks with greater ecological validity;
• resumption lag is roughly 4s - a substantial disruption both in absolute and relative terms;
• interruption lag is brief transitional interval immediately preceding an interruption, operator knows of pending interruption but is not yet engaged by it; during this interval, there is brief opportunity to complete a thought; many real-world interruptions, even more urgent ones afford a brief interruption lag;
• two characteristics of the interruption lag are examined - whether or not primary task is perceptually available during this brief period, actual duration of this period;
• when resuming primary task, retrieval cues will be necessary to re-activate the relevant representations;
• interruption lag has a crucial role to play in facilitating resumption; an opportunity to prepare to resume; it should help to have the primary task display perceptually available during the interruption lag;
• current work examines what length of interruption lag would render cue availability effective in reducing the resumption lag;

• experiments - 4 experiments. with interruption lags of 2,4,6,8 seconds respectively; planning and resource allocation subject to constraints, state information to be represented cognitively; interruptions are 30-45 second long; IV within each experiment is whether or not cues were available during the interruption lag; main DV was the resumption lag, compare the resumption lag to the mean interval b/w primary-task actions;
• each experiment => 24 participants
• curing interruption lag, cursor was hidden and disabled to eliminate active window and mouse cursor as retrieval cues that participants could deploy strategically to remind themselves of what they had been doing before the interruption;
• cue/no-cue was BS, DV was resumption lag (RL); RL compared to inter-action interval (IAI);
• results: all 4 experiments - diff b/w RL and IAI was sig. effect of cue availability on 8-s interruption lag, reliable for 6s; omnibus ANOVA: cue-availability effect marginal, interruption lag effect not reliable, no interaction;

• disruptive effect of interruption obvious when comparing RL and IAI; doesn't agree w/ [Speier 03], [Zijlstra 99]
• total time on task highly variable and showed no trends;
• Zeigarnik effect: experimenter charged with judging when subject was engrossed in order to interrupt w/ greatest impact;
• cue availability during interruption lag affected RL for longer lags - various cog. operations required to locate and encode retrieval cues during interruption lag take b/w 6 and 8 s for this task environment; longer lags afford enough time to link cog. representations to external cues to facilitate retrieval later;
• cue availability and interruption lag did not interact, increase in no-cue resumption lags could be spurious;
• task interruption may increase arousal to improve overall performance, but perhaps a longer interruption lag may decrease arousal and negate this effect; or if changes in arousal occur during interruption (little evidence of this)
• memory or perceptual process mediate delayed effect on task resumption;
• future work: fully-crossed factorial design of interruption lag and cue availability;
• alternative characterisations: [Ballas 99] - automation deficit - measured in accuracy, encoding of perceptual info; cost of re-encoding the display; facilitated by cue availability during interruption lag; one could vary extent of the cog. representations required to perform primary task on one hand, perceptual complexity of display on the other;
• cue-availability effect should be linked to complex cognitive states;

[Hodgetts 06]

Hodgetts HM, Jones DM. Contextual cues aid recovery from interruption: the role of associative activation. Journal of experimental psychology. Learning, memory, and cognition. 2006;32(5):1120-32. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16938050.

• abstract: tower of London problems; retrieval of suspended goal facilitated when given opportunity to encode cues during an interruption lag; impeded when those visual cues were subsequently altered following interruption; context before and after interruption is critical to efficient interruption recovery;

• Speier C, Vessey I, Valacich JS. The Effects of Interruptions, Task Complexity, and Information Presentation on Computer-Supported Decision-Making Performance. Decision Sciences. 2003;34(4):771-797. Available at: http://doi.wiley.com/10.1111/j.1540-5414.2003.02292.x.

[Parnin 10]

Parnin C, DeLine R. Evaluating cues for resuming interrupted programming tasks. Proceedings of the 28th international conference on Human factors in computing systems - CHI '10. 2010:93. Available at: http://portal.acm.org/citation.cfm?doid=1753326.1753342.

• abstract: contextual inquiry for strategies for resuming interrupted tasks, effect of automated cues on improving task resumption; developers rely heavily on note-taking - compare two different automated cues to note taking; both cues performed well, twice the success rate of note-taking alone (task completion DV);

• intro: software developing doesn't have linear structure - recap or rescanning last few paragraphs of text not enough to allow developer to resume work; maintaining task context would help reduce mental workload;
• interruption often comes at inopportune moment and gives insufficient time for preserving mental state;
• taking prospective measures at time of interruption has both negative and positive effects on task resumption, primarily successful when users have environmental cues about working state;
• strong connection b/w internalising environmental cues and reducing resumption costs; environmental cues such as open tabs may be insufficient because spatial and textual cues are lost (not semantic or structural location when automatically encoding working state): screenshot vs. semantic-based cues
• high volume of prospective tasks individually resolved with short-term activities or delegation; developers often given low volume of highly complex work items;
• potential solution: augmenting temporal, spatial, contextual cues;
• potential cues: change summary, code thumbnails, activity explorer, snapshots/instant replay;

• controlled study: allowed subjects in all 3 conditions to take notes, within-subjects protocol; 2 cues feed from log of developer activities (DOI tree-view of code structure, content timeline - chronological list of developer activities and code snippets); one filters program parts by name, activities by code structure, other by content, activities by time;
• at point of interruption: subject would review cue and take notes for up to one minute, each given one cue once (counterbalanced); automatic mechanism for interrupting subjects with next task;

• note taking is situated (within source code) or un-situated (notepad or physical notes); most common way for developers to cope with task resumption
• subjects preferred content time line over DOi tree view - not granular enough; time line places temporary placeholders in code;
• subjects focused on reaching a good stopping point or leaving a quick note
• psyc. studies show that having a period for studying an environmental cue for during interruptions makes the cue for valuable during resumption;
• content timeline for task resumption may be as effect for general knowledge workers as it is for developers;

• Safer I, Murphy GC. Comparing episodic and semantic interfaces for task boundary identification. Proceedings of the 2007 conference of the center for advanced studies on Collaborative research - CASCON '07. 2007:229. Available at: http://portal.acm.org/citation.cfm?doid=1321211.1321235.

• Hodgetts HM, Jones DM. Contextual cues aid recovery from interruption: the role of associative activation. Journal of experimental psychology. Learning, memory, and cognition. 2006;32(5):1120-32. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16938050.

• Monk CA. The effect of frequent versus infrequent interruptions on primary task resumption. In: Human Factors and Ergonomics Society Annual Meeting Proceedings.Vol 48. Human Factors and Ergonomics Society; 2004:295–299. Available at: http://www.ingentaconnect.com/content/hfes/hfproc/2004/00000048/00000003/art00004.

• Altmann EM, Trafton JG. Task interruption: Resumption lag and the role of cues. Citeseer; 2004. Available at: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.4.5438&rep=rep1&type=pdf.

[Morris 08]

Morris D, Morris MR, Venolia G. SearchBar: a search-centric web history for task resumption and information re-finding. Proceeding of the twenty-. 2008:1207-1216. Available at: http://portal.acm.org/citation.cfm?id=1357054.1357242.

• CHI '08 video
• interruptions of web search tasks - searching over multiple sessions
• hierarchical web history organised by topic / web query / individual URLs
• no automatic identification of new topics (manual entry) or default topic
• automatic HTML summaries of topics including notes generated for each topic
• 2 sessions, 1 week apart, planning a travel itinerary, interrupted 3 times in each session with unrelated urgent tasks; 2nd session required reconstruction of work from first session - comparison b/w control and SearchBar group;

• Mancero G, Wong B, Loomes M. Radio dispatchers' interruption recovery strategies. of the 21st Annual Conference of. 2009. Available at: http://portal.acm.org/citation.cfm?id=1738826.1738845.

• Parnin C, DeLine R. Evaluating cues for resuming interrupted programming tasks. Proceedings of the 28th international conference on Human factors in computing systems - CHI '10. 2010:93. Available at: http://portal.acm.org/citation.cfm?doid=1753326.1753342.

[Zijlstra 99]

Zijlstra FR, Roe Ra, Leonora AB, Krediet I. Temporal factors in mental work: Effects of interrupted activities. Journal of Occupational and Organizational Psychology. 1999;72(2):163-185. Available at: http://www.ingentaselect.com/rpsv/cgi-bin/cgi?ini=xref&body=linker&reqdoi=10.1348/096317999166581.

• related: Zeigarnik effect (people not liking to leave task uncompleted): uncompleted tasks better recalled than completed tasks; not replicated by [van Bergen 68], interruptions have effect on participants' recollection of their environment;
• [Cellier 92]: interruptions distract operator in execution of primary task and cause a decrease in efficiency, increase in processing time, error rate
• [Kirmeyer 88] effects of interruptions on appraisal and coping (police dispatchers;
• [Cohen 78] interruptions can be conceived as uncontrollable, unpredictable stressors that produce information overload;
• [Hacker 84, Henning 89]: short breaks have positive effects on performance; appears only to be the case with work of low mental demands / high degree of repetitiveness

• present study: examine effects of interruptions during work; key assumption of theory is that of cognitive control of activity;
• occurrence of interruptions may affect regulation process in a number of ways - modification of action plan, additional demand on resources needed for action execution, regulation of the activity as a whole (including interrupting event), affect worker's subsequent readiness to perform, influencing psychological and psycho-physiological state of the worker;
• goal-directed activity: assume that action should run its course and that it is aversive when actions are interrupted;
• h1: interruptions have detrimental effect on performance - time and error - larger effects for complex interruptions than simple interruptions, larger effects with more frequent interruptions
• h2: workers develop strategies to counteract these effects
• h3: psychological state - reduced well being, negative emotions, anxiety
• h4: occurrence of interruptions is associated w/ higher levels of activation and effort; interruptions pose additional demands on the worker;

• method: office workers interrupted during text editing task; freq. and complexity as factors; interrupted by telephone calls at predetermined and comparable points from experimenter (acting as supervisor), demanded that they perform an additional task w/ priority
• subjects work in simulated office for 2 days, interval of 1 week, standardised text editing tasks; combined one full working day in laboratory;
• main task: edit a text of moderate complexity; undisturbed execution lasted 45-60 min; no time limits, participants free to work at their own pace/working method; interrupted nu phone call with request to perform interrupting task, 2 levels of complexity - simple: request for some irrelevant info, complex: more elaborate tasks - additional short editing task - required participants to leave current document
• participants worked twice - day as factor; freq. of interruptions a WS factor; complexity a within subjects factor in Russian experiment (WS design), BS factor in Dutch study (mixed design);
• DVs: performance characteristics (time - total work time TWT, time-on-task TOT, total interruption time TIT, types of actions and flow of activity, interruption handling strategies, freq. and use of different interruption handling strategies, number of errors and omissions), psychological state indicators (+/- affect, well-being, activity and mood, anxiety), psycho-physiological (mental effort, fatigue)

• results: subjects needed less time on second day (learning effect), no effect of day on TIT; TWT increased when interrupted, TOT sig. decreased when interrupted; no time limits imposed and subjects worked at their own pace; interruptions can lead to an improvement, rather than to a decline, in efficiency in the performance of tasks, contradicting h1
• increase in complexity of interruptions led to an increase of all three time-parameters, increase in TOT did not reach stat. sig.
• TOT not affected (and even seemed to increase) when interruptions became more complex, suggest it is not so much the amount of time that people spend on an interruption that causes them to speed up, but rather the freq. of being interrupted;
• # of errors and omissions appeared not to be affected by freq. or complexity of interruptions;
• Dutch: increase freq. of interruptions leads toward a shift in strategy; delayed strategy becomes more prevalent: controlling moment of intrusion of interrupting event so as to limit its effect on execution of main task; complexity of interruption appeared to have no effect in this respect;
• complex interruptions take more time to handle; effects on change-over are greater in the case of complex interruptions; it takes longer to disengage and reorient to main task after complex interruption; suggests an after-effect of interruptions, no sig. effect of factor 'complexity' on resumption time;
• changes in psycho-physiological state: no sig. differences in well-being, activity, mood; sig. less positive feelings as # of interruptions increased; decrease in pos. emotional feelings b/w condition of no interruptions and condition with one interruption - after introduction of interruptions - increase in number of interruptions has no more effect; increasing complexity of interruptions does not have an effect on psychological state; increasing complexity leads to noticeable changes in emotional and psychological state of participants;
• Russian study: increase in effort expenditure, result of increasing complexity - no effect in Dutch study; Dutch responded more favourably to complex interruptions, emotional state improved, effort decreased; opposite was found for Russian group - complex interruptions could be welcome distractions rather than stressors;
• more complex interruptions may evoke change in strategy, whereby task performance standards go up;

• discussion: interruptions lead to a change in work strategy - main task strategy changed that no deterioration in performance of main task takes place; this is compensated for - quality is unaffected and speed of performance increases; for complexity contrasting effects found b/w research groups;
• contradiction of results from [Cellier 92]; support for Zigarnik effect;
• not really clear how change of working strategy is produced;
• interruptions have negative emotional effect, but effect of complexity had divergent effects;
• interruptions cause a higher level of effort expenditure, without change in activation, additional demands on people's resources, effect of an increase in freq. of interruptions
• future work: examine the process of activity regulation when worker is confronted w/ interruptions; length of change-over interval increases as complexity increases; non-additive effect of multiple interruptions on resumption time is less easy to explain; people have to re-orient after interruption;
• [Cohen 78]: frequent changing of focus of attention causes cognitive / mental fatigue, a decrease in total available attention capacity
• Dutch study: interruptions can have positive effect, more positive feelings, less effort expenditure; - potential for inverted U shape curve for optimal effect of interruptions;

• Avrahami D, Fogarty J, Hudson SE. Biases in human estimation of interruptibility: Effects and implications for practice. SIGCHI conference on Human. 2007. Available at: http://portal.acm.org/citation.cfm?id=1240624.1240632

[Bailey 00]

Bailey B, Konstan J, Carlis J. Measuring the effects of interruptions on task performance in the user interface. SMC 2000 Conference Proceedings. 2000 IEEE International Conference on Systems, Man and Cybernetics. 'Cybernetics Evolving to Systems, Humans, Organizations, and their Complex Interactions' (Cat. No.00CH37166).:757-762. Available at: http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=885940.

• abstract: experiment measuring disruptive effect of an interruption on a user's task performance; six web-based task categories, two categories of interruption tasks; user performs slower on interrupted task than non-interrupted task; disruptive effect of an interruption differs as a function of task category; different interruption tasks cause similar disruptive effects on task performance;

• tasks: adding, counting, image comprehension, reading comprehension, registration, selection (15-40s); interruptions: news article, stock decision; (10-30s);
• Zijlstra et al found that interrupting a task caused user to complete task faster than when performing the same tasks uninterrupted; no time constraint in this study - difference in exp. design
• hypothesised that adding and counting would have highest memory load, selection and registration the lowest, image and reading comprehension in between;
• [Gille 89]: length of interruption not a factor, similarity and complexity are;
• experiment: full-factorial ANOVA with task and interruption category as factors; 3 tasks per category: one control, one interrupted with news task, one interrupted with stock task - to complete as fast as possible while maintaining accuracy; interruption was presented approx. halfway through the primary task; presentation order of task categories, tasks within each category, interruption tasks counterbalanced;
• measurements: time on task, time on interruption;
• results: more time to complete interruption task than non-interrupted task except registration task; main effect of interruption category not present in data, no sig. interactions;
  • interrupting some tasks more disruptive than interrupting others; main effect of task category;
  • no effect of similarity, no interactions - contradicts [Gillie 89]
• discussion: interruption causes sig. increase in task completion time for variety of web-based tasks;
  • increase in task completion time has pos. relationship w/ memory load of task at time of interruption
  • additional length of interruption has no additional disruptive effect on task performance - depends on memory load of task at time of interruption
  • rationale for tasks: generalisability ;

• Zijlstra FR, Roe Ra, Leonora AB, Krediet I. Temporal factors in mental work: Effects of interrupted activities. Journal of Occupational and Organizational Psychology. 1999;72(2):163-185. Available at: http://www.ingentaselect.com/rpsv/cgi-bin/cgi?ini=xref&body=linker&reqdoi=10.1348/096317999166581.

Literature Review Notes

• Bailey B, Konstan J, Carlis J. Measuring the effects of interruptions on task performance in the user interface. SMC 2000 Conference Proceedings. 2000 IEEE International Conference on Systems, Man and Cybernetics. 'Cybernetics Evolving to Systems, Humans, Organizations, and their Complex Interactions' (Cat. No.00CH37166).:757-762. Available at: http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=885940.

• abstract: influence of IM on ongoing computing tasks; generally disruptive effects of IM during fast-stimulus-driven search tasks; interruptions coming early during a search task result in user forgetting primary task goal;
• following from [Gillie 89: info should be presented in such a way as to reduce similarity interference;
• verified findings of previously reported work; main effects for presence of notifications and for gist search type; little to no benefit of having a marker present after a notification was received; reminders used more often if IM was received earlier in search trial; implications for automated systems for delivering notifications;
• summary: come task phases are less amenable to interruption than others; current study results demonstrate harmful effects of notification delivery on memory for the prior task early in a task's lifecycle;
• potential solution: use graphical and linguistic summaries of interrupted task? a single text sentence describing a previous task can be an affective tool to get users back on track after a notification;

• examines length, similarity, complexity of interruptions and how they disrupt ongoing tasks; adventure game paradigm;
• length of an interruption cannot be used to determine how disruptive it will be;
• interruption that is similar to the main task and demands immediate attentions is disruptive; opportunity to rehearse one's position in main task doesn't guarantee immunity from disrupting effect of interruption - similarity of interruption to main task is not crucial in determining whether or not an interruption will be disruptive;
• note on retrieval cues used in main task: cues not seen until after list of task items is presented;
• length of an interruption, opportunity to control point at which main task is stopped and interruption started are not important factors in determining whether or not interruption will disrupt performance on return to interrupted task; rather, nature of interruption (similarity to main task) and complexity of interruption / amount of processing or memory storage required seem to determine which interruptions will be disruptive and which will not;
• opposed to findings in Kriedfeldt and McCarthy (1981) - similar interruption task disrupts ongoing task (calculations in both); Field (87) - disruptive effect of interruptions on users' post-interruption activity - no comparison of interrupted and non-interrupted problems;
• Zeigarnik effect - interrupted tasks were recalled more often than uninterrupted tasks; asking wrong question - how easily do people resume what they were doing when interruption finishes;

• best paper award, CHI 05

• abstract: methods for inferring cost of interrupting users based on multiple streams of events;
• acoustical and visual sensors, monitor activity of a user interacting with different client devices, pending availability of microphone and camera, computing devices for reporting visual pose / head tracking, online appointment information; capture and share out the state of low-level and higher-level events; information about application in and out of focus
• focus on learning models that predict the state of interruptibility of users in office settings, pursue models that characterise the user's interruptibility
• related research: Coordinate, Seer
• summary: effort to build models that can predict cost of interrupting users; tools for logging and tagging a database of cases, probe sensitivity of classification;

• abstract: human coding of audio and video recordings; experience sampling; self-reports of interruptibility;
• could a robust estimator of interruptibility make people more efficient and/or more relaxed?
• related work: tasks that were interrupted were more likely to be recalled after a delay than tasks that were not interrupted [1920s German psych. research]
• recipients of interruptions often derive personal benefit from the interruption, often at the expense of the initiator;
• subjects asked to rate interruptibility on scale of 1 to 5;
• simulated sensors: occupant related (occupant presence, speaking, writing, sitting, standing, on the phone, touch/interaction with desk, table, file cabinet, food, drink, keyboard, mouse, monitor (gaze), papers), guest related (number of guests present, pose of each guest), environment-related (door open/close, day of week and time of day) - chosen because a priori belief that they would be correlated with interruptibility;
• strongest indicators of non-interruptibility would be those related to social and task engagement;
• info gain stats only consider the predictive power of features in isolation and do not take into account the overlapping nature of researchers' sensors;
• combined "anyone talking sensor" most predictive, telephone, keyboard, time of day keyboard also predictive
• conclusion: sensor-based estimators of human interruptibility are possible

[Fogarty 05]

Fogarty J, Ko AJ, Aung HH, et al. Examining task engagement in sensor-based statistical models of human interruptibility. Proceedings of the SIGCHI conference on Human factors in computing systems - CHI '05. 2005:331. Available at: http://portal.acm.org/citation.cfm?doid=1054972.1055018.

• task engagement as predictor of interruptibility rather than task engagement;

• simultaneous or repeated demands can quickly become disruptive
• prior studies of reliability of sensor-based statistical models of human interruptibility have primarily yielded results related to social engagement; present work examines task engagement in sensor-based models of human interruptibility;
• experiment: value was associated with interruptions by telling participants that they would lose $2 for any ignored or incorrectly answered interruption; participants required to do mental multiplication; several practice interruptions given prior to the 70 minute primary task, during a 10 minute session of web surfing;
• interruptibility measured in terms of difference b/w time when blinking taskbar notification was displayed, indicating that an interruption was pending, and the time that the participant acknowledged the interruption by clicking on the taskbar notification; participants were externalising their working memory into the state of their development environment before addressing the pending interruption;
• experimenters did not pursue investigation into whether interruptions resulted in introduction of actual errors into Paint program; negotiated coordination of interruptions allowed participants to carefully externalise their working memory in the dev. environment;
• validation of data collected: each pair of classified clusters sig. different;
• related work: field studies of interruptions and how people perceive their interruptibility have informed current work, but do not directly inform deployment of sensor-based statistical models of human interruptibility;

[Fogarty 04]

Fogarty J, Hudson SE, Lai J. Examining the robustness of sensor-based statistical models of human interruptibility. Proceedings of the 2004 conference on Human factors in computing systems - CHI '04. 2004;6(1):207-214. Available at: http://portal.acm.org/citation.cfm?doid=985692.985719.

• social engagement plays a major role in interruptibility (self-reports); four subjects had similar jobs as high-level staff in university; this approach to modelling applies to a wide variety of people;
• data collection: experience sampling technique (Hudson et al); beeper study; recording subject responses to interruptibility prompts, sensor to determine whether anyone was talking in the office; magnetic switches on the door frame, phone hook; microphone talking sensor, number of keyboard, mouse move, mouse click events in previous second, logged title, type, and executable name of active and non-active windows;
• model performance: first feature to be selected was whether phone was off the hook in the last 15 seconds; then talking for 3 of last 5 minutes; then if subject generated 30 mouse move events in last 15 seconds; similar for other subject groups (manager, researcher, intern); talking sensor used for all 3 groups, time interval over which to examine the talk sensor is different for the 3 groups;
• statistical models can adapt to nuances of interruptibility
• sensor combinations: built-in laptop microphone and computer activity - manager model has accuracy of 86.6%; social engagement critical to predicting the interruptibility of manager subjects; task engagement for researcher and intern; phone sensor next most valuable sensor;

[Hudson 03]

Hudson SE, Fogarty J, Atkeson C. Predicting human interruptibility with sensors: a Wizard of Oz feasibility study. on Human factors in. 2003;(5):257-264. Available at: http://portal.acm.org/citation.cfm?id=642611.642657.

[Horvitz 03]

Horvitz E, Apacible J. Learning and reasoning about interruption. Proceedings of the 5th international conference on Multimodal interfaces - ICMI '03. 2003:20. Available at: http://portal.acm.org/citation.cfm?doid=958432.958440.

[Cutrell 01]

Cutrell E, Czerwinski M, Horvitz E. Notification, disruption, and memory: Effects of messaging interruptions on memory and performance. Human Factors. 2001;(1999). Available at: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.26.418.

[Gillie 89]

Gillie T, Broadbent D. What makes interruptions disruptive? A study of length, similarity, and complexity. Psychological Research. 1989;50(4):243-250. Available at: http://www.springerlink.com/index/10.1007/BF00309260.

• Horvitz E, Apacible J. Learning and reasoning about interruption. Proceedings of the 5th international conference on Multimodal interfaces - ICMI '03. 2003:20. Available at: http://portal.acm.org/citation.cfm?doid=958432.958440.

• Avrahami D, Fogarty J, Hudson SE. Biases in human estimation of interruptibility: Effects and implications for practice. SIGCHI conference on Human. 2007. Available at: http://portal.acm.org/citation.cfm?id=1240624.1240632.

• Fogarty J, Ko AJ, Aung HH, et al. Examining task engagement in sensor-based statistical models of human interruptibility. Proceedings of the SIGCHI conference on Human factors in computing systems - CHI '05. 2005:331. Available at: http://portal.acm.org/citation.cfm?doid=1054972.1055018.

• Fogarty J, Hudson SE, Atkeson CG, et al. Predicting human interruptibility with sensors. ACM Transactions on Computer-Human Interaction. 2005;12(1):119-146. Available at: http://portal.acm.org/citation.cfm?doid=1057237.1057243.

[Zhai 04]

[Zhai 04]

S. Zhai, J. Kong, and X. Ren, "Speed-accuracy trade-off in Fitts' law tasks — On the equivalency of actual and nominal pointing precision," International Journal of Human-Computer Studies, vol. 61, 2004, pp. 823-856.''

• abstract: when operating with different speed or accuracy biases, performers may utilise more or less area than the target specifies; complex interaction effect b/w Iu and Id suggesting that a simple and complete model accommodating both layers of speed-accuracy tradeoff may not exist;

• intro: this second layer (or component) of speed-accuracy tradeoff is subjective and personal;
• as practiced in the literature, they are called effective distance and effective target width, although the implication of the effective term is biased one without proof;
• it is reasonable to expect that Fitts' law in its behavioural form, or at least some modification of the task form of Fitt's law in consideration of the deviation, be more useful

• related work: most Fitt's studies instructed the participants to perform "as fast as possible and as accurately as possible" but do not systematically vary or control experimental participants' actual precision relative to the nominal task precision; it is difficult to evaluate the influence of the subjective layer of the speed-accuracy tradeoff and the correction using Ide.
• focus of this study is on modelling aspects of speed-accuracy tradeoff in aimed movements;

• experiment 1: each participant was instructed to repeat the experiment three times with different operating conditions biased toward accuracy or speed: accurate, neutral, and fast; balanced by a Latin square pattern across the 12 participants
• 28 trials removed of 2564 total (accidental clicks removed in data processing)
• instructions for operating strategy in the 3 experimental conditions had an obvious impact on participants' target utilisation and error rate; average error rate in A,N,F, conditions was 3.2%, 10%, 19,4%; higher than hoped (expected N condition to be 4%);
• 95-99% of variance in T could be accounted for by change of Idn; Fitts' law regression coefficients a and b using the Idn model were influenced by operating biases (b varied 62% from condition F to A in this experiment); compared to 17.6% when compared to Ide;
• r^2 b/w T and Ide decreased from corresponding r^2 between T and Idn (83-87% of T variance vs. 95-99% of variance); shrinkage of range of independent variable from Idn to Ide; Ide compensated operating biases (more converging a and b parameters), but not completely; robustness as a determinant of pointing time as measured by r^2 decreased within each operating condition but increased across conditions;

• experiment 2: 2 conditions (A,F); "move as fast as possible, a few errors OK" (soft ding on errors) and "try to avoid any errors" (loud ding on errors); 160/2178 trials discarded; avg error rates were 0 and 0.5% for A and F conditions; participants clearly more conservative in A than F;
• Idn is a slightly more robust pointing time determinant than Ide within each condition; Ide more robust than Idn across conditions; range of Idn shrank from that of Idn but only from the low end of the index of difficulty; Ide yielded more converging a and b parameters b/w the conditions than Idn hence it compensated for the different target utilisation levels in different operating bias conditions;

• experiment 3: greater range of target utilisation required; each participant repeated experiment 5 times with different operational strategies ranging EA, A, N, F, EF - results: 0%, 1%, 4%, 9%, 22%;
• Idn was shown again to be a remarkably robust determinant of the mean pointing time within each condition;
• Ide is a stronger determinant than Idn when data from all conditions were merged in one regression; Ide accounted for 78% of variance of mean trial completion time; Idn could account for 46%
• Ide could at least partially overcome the different levels of target utilisation due to operating biases and produce more stable estimates of Fitts' law coefficients;
• nominal task precision and performer's bias in over or under utilising the given task precision tolerate change pointing completion time; index of difficulty Idn is a remarkably robust predictor of completion time within each operating condition;
• We as a posteriori adjustment does not fully account for the time performance difference caused by the second and subjective layer of speed-accuracy tradeoff - performance's incompliance with the task specification resulting in an overall faster or slower speed;

• experiment 4: comparing the target width compliant scheme and the target width in-compliant scheme we could directly examine whether a d how well We reconciles the two layers of speed-accuracy tradeoff; Ide only partially compensates for the time variance caused by different operating biases; The impact of We adjustment lagged behind the impact of W changes in the same amount;

• discussion: use of We is an imperfect and insufficient adjustment of W, although the direction of adjustment was empirically correct;
• performer introduced accuracy layer causes a discrepancy b/w nominal task precision and actual behaviour precision (different a and b coefficients)
• the task form of Fitts' law T = f(Idn) is a very strong model; nominal Idn is an impressive determinant of mean movement time; accounting for up to 99% of time variance within an experimental condition;
• implication of this study depends on purpose of use; two layers of speed-accuracy tradeoffs, objective and task layer, subjective and behaviour layer have different impact on task performance;
• study shows that even for a low level tapping task, both external visual feedback and internal bias settings contribute to control process;
• pointing tasks performers could adjust their overall bias towards speed or accuracy and integrate such an internal high level setting with the external low level visual feedback to manage a pointing process;
• Iu should be kept as close to zero as possible and its variance should be kept as low as possible; adjustments consistently yield more logical Fitts' law parameter estimates, a and b, incomplete compensation for the subjective layer of speed-accuracy tradeoff;

more...

[Adamczyk 04]

Adamczyk PD, Bailey BP. If not now, when?: the effects of interruption at different moments within task execution. of the SIGCHI conference on Human. 2004;6(1):271-278. Available at: http://portal.acm.org/citation.cfm?id=985692.985727.

• abstract: measure effects of interrupting a user at different moments within task execution in terms of task performance, emotional state, social attribution; different interruption moments have different impacts on user emotional state and positive social attribution; design of an attention manager;

• intro: poorly time interruptions can adversely affect task performance; authors identify moments for interruption utilising task models based on event perception research;
• results show that predicted points for interruption consistently produced less annoyance, frustration, and time pressure, required less mental effort, deemed by user more respectful of their primary task;

• discussion: predicted worse condition produced the worst results on the same measures, causing users to rate their experience even more poorly than the random condition
• implications for design: designers should consider alternate modalities for interruption; multimodal interruption schemes;
• interruption lag can cue a rehearsal process before onset of an interruption to help in resuming primary task; however, training users in a rehearsal strategy actually decreased task performance [Miller 2007]. (Window of opportunity: Using the interruption lag to manage disruption in complex tasks. HFES '02.)

[Gluck 07]

Gluck J, Bunt A, McGrenere J. Matching attentional draw with utility in interruption. Proceedings of the SIGCHI conference on Human factors in computing systems - CHI '07. 2007:41. Available at: http://portal.acm.org/citation.cfm?doid=1240624.1240631.

• CPSC 544 notes: One aspect that I felt this article was lacking was the emphasis on the need for interruptions and notifications to be consistent across all applications on a system, and within each application a subset of low, medium, and high-priority interruptions. In addition to this, all applications may not be equally important, which will vary from user to user. For instance, User A may rank their email application program as a high-priority application, while their office suite is ranked as medium-priority, and their media player as a low priority application (a medium-priority interruption from the email application (i.e. incoming mail with no high-priority flag set by the sender) would be more important than a medium-priority interruption from the media player (i.e. notification of a song change)). Meanwhile, User B might not care about email at all, and could have an entirely different ranking of applications (and thus a different ranking of interruptions) - (i.e. they want to receive Facebook Wall Post alerts before anything else...). To a certain extent the application GROWL (Mac / Windows) allows users to set these priorities, and make the interruptions consistent in terms of timing and look-and-feel across all applications.

more...

• Findlater L, McGrenere J, Modjeska D. Evaluation of a role-based approach for customizing a complex development environment. Proceeding of the twenty-sixth annual CHI conference on Human factors in computing systems - CHI '08. 2008:1267. Available at: http://portal.acm.org/citation.cfm?doid=1357054.1357251.

• Hudson SE, Fogarty J, Atkeson C. Predicting human interruptibility with sensors: a Wizard of Oz feasibility study. on Human factors in. 2003;(5):257-264. Available at: http://portal.acm.org/citation.cfm?id=642611.642657.

• Fogarty J, Hudson SE, Lai J. Examining the robustness of sensor-based statistical models of human interruptibility. Proceedings of the 2004 conference on Human factors in computing systems - CHI '04. 2004;6(1):207-214. Available at: http://portal.acm.org/citation.cfm?doid=985692.985719.

• Gluck J, Bunt A, McGrenere J. Matching attentional draw with utility in interruption. Proceedings of the SIGCHI conference on Human factors in computing systems - CHI '07. 2007:41. Available at: http://portal.acm.org/citation.cfm?doid=1240624.1240631.

[Iqbal 06]

Iqbal ST, Bailey BP. Leveraging characteristics of task structure to predict the cost of interruption. Proceedings of the SIGCHI conference on Human Factors in computing systems - CHI '06. 2006:741. Available at: http://portal.acm.org/citation.cfm?doid=1124772.1124882.

• goal: create parsimonious model that uses characteristics of task structure to predict COI, a model that can be expediently applied to many goal-directed tasks

• systems could predict a more accurate COI if they also considered characteristics related to the structure of a task - subtasks, boundaries within task decomposition
• prior work showed that interrupting at subtask boundaries results in much lower COI than non-boundary moments
• users performed representative primary tasks and were interrupted at various boundaries with peripheral tasks; resumption lag, time to resume primary tasks after an interruption, was used to provide ground truth for COI
• benefit of COI model is that it can be expediently applied to approximate COI at subtask boundaries in many goal-directed tasks, without a physiological measure of workload

• empirical COI: interrupting at random moments can cause users to take up to 30% longer to resume the tasks, commit up to twice the errors, experience up to twice the negative affect than when interrupted at boundaries
• interrupting at earlier phases in task structure had lower cost;
• current work focus on subtask boundaries because systems can detect them, further investigates which characteristics of a task's structure best predict the COI at subtask boundaries;
• predicting COI: build a probabilistic model w/ input cues including desktop activity, visual and acoustical analysis, scheduled activities of the user

• experiment 1: uses a within-subjects design
• initial models based on authors' own understanding of tasks' execution; each model achieved more than 90% accuracy with no obvious patterns in the errors
• peripheral tasks took ~20s to perform and must be responded to as quickly as possible, while returning to the primary task as quickly as possible
• moments for interruption chosen from sample of ten representative subtask boundaries from corresponding GOMS model
• measurements: resumption lag = COI
• candidate factors for predicting COI: level, presence (not saliency) of visual resumption cue, percent of task complete at boundary, percent of parent subtask complete at boundary, difficulty of preceding subtask, difficulty of next subtask, carry over at boundaries;
• determining most predictive factors of resumption lag: at least one of the factors has a non-zero coefficient: level, carry-over, difficulty of next subtask; COI depends more on characteristics that reflect current and prospective allocation of mental resources than on those that reflect temporal position or cue availability in the task
• decided to cluster resumption lag into classes and determine cost for these interrupt-ability classes; total # of correct predictions 63.2%, much higher than chance (33%)

• experiment 2: eval. of the COI model; assignment of subtask boundaries to 3 COI clusters - correctly predicted 53% of COI values, much better than chance (33%); COI model can be reasonably generalised to other goal-directed tasks
• validates that a system can and should use our model to differentiate among subtask boundaries, enabling more effective decisions about when to interrupt

• discussion: model can differentiate subtask boundaries based on workload;
• limitations: model considers only one component of task structure - subtask boundaries, because systems can detect them
• familiarity with a task seems to have little effect on how users perceive its hierarchical structure; a possible solution is to extend COi model to include skill level as a predictor
• COI model best suited for tasks that are high freq., routine, safety critical

• future work: investigate automated methods for building task models and predicting COI
• extend COi model to include non-boundary points
• implement our COI model within an existing interruption reasoning system

• Adamczyk PD, Bailey BP. If not now, when?: the effects of interruption at different moments within task execution. of the SIGCHI conference on Human. 2004;6(1):271-278. Available at: http://portal.acm.org/citation.cfm?id=985692.985727.

[Moffatt 10]

K.A. Moffatt, "Addressing Age-Related Pen-Based Target Acquisition Difficulties," Ph.D thesis, 2010.

• [Zhai 04] documents user performance differences between five wordings of speed/accuracy task instructions for a Fitt's-like task.

• 3. Baseline study: identifying pen-based target acquisition difficulties across the adult lifespan
  • 3.2.2. Participants - 3 groups, 12 people each: young (19-54; actual: 19-53, M 31.7), pre-old (55-69; a: 55-68, M 62.8), old (70+; a: 73-85, M 76.9)
  • The justification for these groupings rests on the age-related changes that occur in cognition [Craik 92: The Handbook of Aging and Cognition], notably that higher cognitive function remains relatively stable up to about age 55, after which there is a small decline, followed by a much steeper one after 70.

more...

• abstract: resumption of a task after interruption is facilitated during the interruption lag; people adapt to partially disruptive forms of interruption

• experiment conducted that supports the prediction and helps to show how people prepare prospectively, by encoding specific goals to achieve at time of resumption, and retrospectively, by rehearsing state information from the point of interruption;

• task analysis: interruption lag (time b/w alert and start of secondary task); resumption lag (time b/w leaving secondary task and beginning primary task
• behaviourally, there will always be a lag in shifting from secondary task back to primary task; basic opportunity is the interruption lag, a brief window in which to lay the cognitive groundwork for returning to the primary task and therefore reducing the resumption lag; make use of interruption lag to reduce resumption lag;

• theoretical framework for interruptions: people do in fact retrieve goals even when reconstruction is a viable option;
• ACT-R cog theory; to capture relevance of a particular item, memory system computes that item's activation from both the item's history of use and from its associations to cures in the current mental or environmental context; the cog. system should be able to exploit the predictive computations of the memory system to overcome decay and keep certain information active for use in the future
• two ways people might use interruption lag to prepare the primary task: prospective goal encoding and retrospective rehearsal (state information from the primary task at moment of interruption) - deliberate use of memory
• [Gillie and Broadbent 89] - having the opportunity to rehearse the point of interruption does not automatically offer protection against the disruptive effect of an interruption;
• activation of a goal is priming from contextual cues to which a goal is linked
• we do not manipulate the availability of cues in this particular experiment, but do expect the verbal protocol data to reflect emphasis on perceptually available info

• experiment - manipulate opportunity to prepare to resume after an interruption; expected that participants in immediate condition would resume more slowly because it would take them longer to retrieve the relevant goal, or more goal reconstruction; experimenters provided no instruction that preparation might be a useful strategy
• used talk-aloud protocol; design: BS - warning or immediate condition, WS - session
• computed resumption lag as interval from the moment the primary task interface was restored following the interruption to the first mouse click or key press a participant made to resume the primary task; inter-click lags stable within participants; computed disruption scores by computing differences of inter-click lag from resumption lag
• used talk-aloud protocol to determine amount, timing, type of preparation during interruption lag
• results: 62% involved prospective goal encoding, 38% used retrospective rehearsal of state information
• warning participants prepared sig. more than immediate participants
• waring participants used interruption lag to prepare
• resumption lag was longer than inter-click lag; general practice effect found
• main effect of having a warning was not reliable for performance on secondary task, nor was interaction b/w session and condition; interruption lag did not have generalised effect on performance overall
• warning condition and session also interacted in effect on disruption score, disruption score was unaffected in warning condition
• participants were able to improve their ability to resume the task, but only in immediate condition
• for session one, the warning condition had a lower disruption score than immediate condition; no effect of warning for session two or three
• summary: interruptions were disruptive, participants did use interruption lag to prepare to resume, producing smaller disruption scores; preparation boosted activation of whatever information was target of the preparation, facilitating retrieval of this information from memory during resumption

• discussion: focus is on prospective actions to be taken at task resumption - as opposed to state information from point of interruption - involved directly available perceptual information (as opposed to intermediate products)
• brief interruption lag of a few seconds can facilitate resumption; interruption lag was 8 seconds, however a shorter lag of 1-2 seconds should serve equally well, especially if operators become skilled in using interruption lag to full advantage
• retrieval cues are necessary to be able to retrieve information about primary task at point of resumption; people do make use of environmental cues in preparing to resume
• lessons learned: provide operator w/ brief opportunity to prepare to resume interrupted task later, w/ salient retrieval cues associated with primary task information
• authors do not claim that simplest alert is always the best, but tradeoff exists, time to parse a complex alert will be deducted from time invested in preparing to resume
• training can compensate for lack of interruption lag

[Iqbal 07]

Iqbal S, Horvitz E. Disruption and recovery of computing tasks: Field study, analysis, and directions. Proceedings of the SIGCHI conference on Human factors in computing systems. 2007:686. Available at: http://portal.acm.org/citation.cfm?id=1240624.1240730.

• abstract: field study - suspension and resumption of tasks;

• intro: characterise task suspension and recovery among information workers in the course of their normal daily computing tasks; pursued patterns and understanding of user behaviour before and after interruptions
• related work [Czerwinski],[Cutrell],[Horvitz] has studied effects of external interruptions, impact of varying timing and type of interruption; investigated effects interruption on error rates;
• inopportune interruptions can increase task performance time, users have to reallocate resources to the suspended task, which becomes increasingly difficult if the resource demands were high to begin with

• study: when users suspend a task because of an alert or for other reasons, they may take advantage of the break in the execution of the primary task offered by the switch to interact with other peripheral application, and perhaps turn into other tasks;
• study took place over course of 2 weeks; findings corroborated through interviews

• phases of interruption lifecycle: preparation (user subconsciously or consciously performs activities that leave primary task in a more stable state before switching to alerting application), diversion, resumption (if spent more than 15 seconds on suspended application), pre-interruption;
• sought to better understand difficulty that people had with resuming applications that has been suspended as a result of responding to an alert, quantify difficulty in terms of time taken to return to suspended application and to restore context and state;
• deployment: 27 people, various job descriptions, IM/email alerts enabled, 974 sessions of ~2hr

• results: responding to an alert: 5 min to suspend application for email, 8min for MSN, 7min for communicator, 34s for Win. Messenger;
• users typically prefer to complete conceptual and/or motor subtasks before switching and do so quickly before responding to an alert;
• users may selectively work to externalise pending copy-paste goals
• no sig. dif. b/w saving operations in two phases
• intuition was that higher switch rates might indicate lower levels of focus on any one task;
• diversion phase was significantly higher than the switch rate in the pre-interruption phase and te switch rate during the resumption phase was higher than the switch phase in the pre-interruption and diversion
• findings reinforce hypothesis that breaks in activity associated with alerts from an application provide opportunities for performing communication operations with others, at significantly higher rates than usual
• findings suggest that visibility of windows may serve as a reminder to users to break out of the diversion chain and return to suspended applications
• results indicate that the diversion can result in substantial lag in resumption of primary tasks

• interviews: interviewees did not save documents categorically before switching since they were in the habit of saving periodically; auto-save upon switch feature would be helpful;

• lessons learned:
  • users view alerts as an awareness mechanism rather than a trigger to switch tasks, but the alerts often cause them to do otherwise
  • immediate responses indicate alert-driven interruptions and delayed responses are indicative of self-initiated interruptions
  • users spend more time than they realise responding to alerts
  • importance of suspended task is associated with early recovery
  • visibility of suspended application windows is associated with faster recovery

• design implications:
  • provide visual indicators of occluded application windows to assist with recovery of suspended tasks
  • automatically save task context on suspension
  • provide easy access to suspended task context
  • provide playback of actions within task contexts

• future work: influence of face-to-face and phone-based interruptions on task disruption and recovery (vs. computer-based alerts)

[Trafton 03]

Trafton GJ. Preparing to resume an interrupted task: effects of prospective goal encoding and retrospective rehearsal. International Journal of Human-Computer Studies. 2003;58(5):583-603. Available at: http://linkinghub.elsevier.com/retrieve/pii/S1071581903000235.

• Iqbal ST, Bailey BP. Leveraging characteristics of task structure to predict the cost of interruption. Proceedings of the SIGCHI conference on Human Factors in computing systems - CHI '06. 2006:741. Available at: http://portal.acm.org/citation.cfm?doid=1124772.1124882.

• Iqbal ST, Bailey BP. Understanding and developing models for detecting and differentiating breakpoints during interactive tasks. Proceedings of the SIGCHI conference on Human factors in computing systems - CHI '07. 2007:697. Available at: http://portal.acm.org/citation.cfm?doid=1240624.1240732.

• Iqbal S, Horvitz E. Disruption and recovery of computing tasks: Field study, analysis, and directions. Proceedings of the SIGCHI conference on Human factors in computing systems. 2007:686. Available at: http://portal.acm.org/citation.cfm?id=1240624.1240730.

• Trafton GJ. Preparing to resume an interrupted task: effects of prospective goal encoding and retrospective rehearsal. International Journal of Human-Computer Studies. 2003;58(5):583-603. Available at: http://linkinghub.elsevier.com/retrieve/pii/S1071581903000235.

[Marsh 98]

Marsh RL, Hicks JL, Landau JD. An invesitgation of everyday prospective memory. Memory and Cognition. 1998;26:633–643. Available at: http://www.psychology.uga.edu/hcpl/pub_pdf/16.pdf.

• PM better when retention interval is shorter b/w forming an intention and the time it is to be completed; incentives are offered to promote remembering; prospective tasks are more important;
• factors unique to PM: how comfortable the individual feels about completing the intention, strategies used to remember to perform an action in the future; how PM skills might develop from childhood, whether age-related differences exist;
• PM supported by set of human planning processes
• authors' approach was to freely admit that sacrificing some laboratory control in favour of ecological validity might nevertheless yield important insights regarding how people remember to accomplish their intentions. assumed that cognitive capacities other than RM were critical to PM performance
• most lab studies investigated a single intention performed once; or same intention repeatedly performed; paradigm investigates intentions that participants did not establish themselves;
• authors see reason to address planning and re-prioritisation processes that people use over extended periods of time;
• from the time of formation, most intentions are delayed for some extended period; intentions are then executed as environmental and cognitive demands permit; approach investigates the outcome of the natural processes of updating, abandoning, and revising one's plans as they occur simultaneously with all the various and typical human difficulties that people have managing their time and behaviour;
• experiment 1: assess how many plans people established for themselves over the course of a week and to determine what proportion was actually accomplished; why some intentions went unfulfilled; use of memory aids, native differences in attentional and memory capacities examined;
  • recorders and non-recorders (use of memory aids) had equivalent overall completion rates
• participants remembered to fulfil prearranged appointments and those that involved a commitment to another person; less diligent in fulfilling intentions to arrange appointments, to take or return things
• people overtly forget very few of their plans; re-prioritisation of intentions occurs as current demands dictated and as other opportunities and obligations arose;
• recorders neither had nor did they complete larger numbers of intentions in comparison to non-recorders
• natural recorders seemed to have deficits in both attentional and memory capacities
• recorders thought less about their obligations
• mental rehearsal of one's obligations beneficial to non-recorders, detrimental to recorders
• although recorders might realise benefit of a written record of things to be accomplished, non-recorders can benefit by changing their strategies to include writing things down;
• comprehensive theories do not exist for how people contend with the intentions they establish for themselves;
• PM relies on how people contend with intentions they establish for themselves;
• PM remembering clearly relies on a multidimensional set of cognitive processes;
• one cog. resource (good RM) augments another (more freq. monitoring) for non-recorders
• planning as it relates to re-prioritisation of old and newly established intentions - unrelated to RM, but important factor in PM remembering

• TWTE: test-wait-test-exit [Miller 1960]; no continuous monitoring but rather periodic evaluation for target time for intended action (time-based PM task); glancing at a clock;
• TWTE freq. and amount decreases w/ age; older people less vigilant / less evaluations
• supervisory executive system stimulates testing / initiates monitoring? depends on judgments of time derived from biological/cognitive internal clocks
• personality variables such as compulsiveness and conscientiousness may play a role; state-oriented > action-oriented personalities (higher PM performance) - more compulsively vigilant
• could monitoring be stimulated by environmental context?
• monitoring / occasional evaluation by random train-of-though process in multidimensional semantic space?
• lab studies tend to support TWTE

• alternative: preparatory attentional processes [Smith 04] (PAM)
• continuous monitoring - distinguishing target from non-target events; rehearsal of target?
• no behavioural movement indicative of monitoring (unlike TWTE)
• lab studies add ongoing task to share attentional resources
  • those that use articulatory loop of working memory => no effect on PM performance
  • those that use central/supervisory attentional system => neg. effect on PM performance (i.e. generating random order of digits)
• attentional demands may disrupt other tasks (other than monitoring)
• cost of PM task monitoring should be evident on non-target trials of ongoing task: those w/ PM intention vs. those without
  • i.e. lexical decision task, category judgment, sentence item completion, colour matching
• is monitoring functionally related to PM?
• more monitoring / emphasis on PM task => increased PM performance;

• PAM unlikely in everyday circumstances in which formation of intention and opportunity to perform intended action may be hours or days apart => PAM implies an attentional deficit to ongoing daily activities to support monitoring for PM event window
• what about in context? i.e. picking up milk on way home after work only becomes active upon driving home, but not throughout the day;
  • tested by [Cook 05], [Marsh 06]; inverse effect shown when target event occurs in inappropriate context
• is intention to monitor for appropriate context to perform action in itself another PM task?

• alternative: a PM retrieval mode - an increased sensitivity to target event, but no/little demand on attentional resources / deficit to ongoing activity performance;
• for retrospective memory (RM): divided attention (DA) at encoding => hit to RM; DA at retrieval => no hit to RM; could same exist for PM performance?

[McDaniel 07]

McDaniel MA, Einstein GO. Prospective Memory: An Overview and Synthesis of an Emerging Field. 1 ed. Thousand Oaks, California: Sage Publications, Inc.; 2007:264.

Ch.1

Prospective Memory: A New Research Enterprise

• Parameters of PM tasks:
  • execution of intended action is not immediate
  • PM task is embedded in ongoing activity
  • time frame for response execution is limited
  • there must be an intention

• event-based PM paradigm:
  • participants busy w/ ongoing task
  • at outset, participants asked to perform another task at appearance of target
  • participants given a distraction task prior to ongoing task so that PM intention is not maintained in working memory
  • performance measured by proportion of trials in which participants remember to execute PM task

Ch.2

Monitoring in Prospective Memory

• Schulze H. MEMOS: An Interactive Assistive System for Prospective Memory Deficit Compensation. Proceedings of the 6th international ACM SIGACCESS. 2004:79-85. Available at: http://portal.acm.org/citation.cfm?id=1028645.

• Carmien SP, Cavallaro FI, Koene RA. Ê»Senior Momentsʼ : Loss and Context. In: Proceedings of the 2nd International Conference on Pervasive Technologies Related to Assistive Environments. Corfu, Greece; 2009:0-3.

• Voinikonis A, Irmscher K, Schulze H. Distributed processing of reminding tasks within the mobile memory aid system, MEMOS. Personal and Ubiquitous Computing. 2005;9(5):284-290. Available at: http://www.springerlink.com/index/10.1007/s00779-004-0332-5.

[McDermott 04]

McDermott K, Knight RG. The effects of aging on a measure of prospective remembering using naturalistic stimuli. Applied Cognitive Psychology. 2004;18(3):349-362. Available at: http://doi.wiley.com/10.1002/acp.977.

• abstract: determine effect of age on noticing and memory search stages of a PM task; use of naturalistic stimuli; older participants had disproportionate reduction in recall of actions relative to cues notices compared to both younger and middle-aged participants; noticing complex and naturalistic visual cues in a shopping environment may rely primarily on automatic memory processes, whereas memory search depends on more strategic and conscious memory processes;

• intro: aimed to construct a measure that would avoid the ceiling effects of many single-item tests of PR;
• familiarity with naturalistic tasks mobilise a high degree of motivation in patients w/ brain injuries and older adults;
• to enhance distracting nature of street scene, participants asked to keep count of number of strollers and bicycles; naturalistic equivalent of ongoing task used in laboratory measures of PM;
• important distinction b/w automatic memory processes relevant to noticing stage and conscious and strategic processing involved in search stage; depending on characteristics of task, likely that noticing may involved both automatic and conscious search processes; evidence that automatic processing is relatively invariant with age, whereas capacity for conscious memory search declines;
• in present study, unfamiliar environment used to minimise participants' opportunity to search actively for relevant cues; retrospective recall of associated action is assumed to require a self-initiated search process that is primarily conscious; possible that some actions are recalled automatically;
• age-related decline in event-based PM has been demonstrated consistently
• Rendell and Craik raised possibility that PM would be unaffected by age on tasks completed under real-life conditions; older participants' performance was superior to that of the younger group; although younger people generally have better PR skills, older people may show superior performance on naturalistic tasks where they are likely to be highly motivated to succeed and have sufficient time available to focus on remembering intentions

• method: participants additionally required to report how many bicycles and strollers seen during video, make cell phone call to make doctor's appt 20 min after video started;
• results: sig. main effect of age, sig. main effect of proportion (cues/response) correct, significant interaction; simple main effects of age with both proportion scores; proportions of cues and tasks remembered did not differ between younger and middle-aged group; older group differed from both other groups on proportion of tasks recalled; older adults had proportionally more difficulty correctly recalling the action to be taken than noticing visual cues for action than did other two groups;
• consistent decline in performance on recalling cue and task for each item across age groups; regardless of nature of item, older individuals less likely to complete instructed action;
• counting bicycles and strollers: group means did not differ significantly;
• time-based recall: sig. group difference; older differed from younger, not from middle-aged;

• discussion: participants disproportionately less able to perform memory search than to notice relevant cue compared to younger and middle-aged groups; recognising a cue but being unable to remember what it means becomes more common in old age;
• very little is known about affect of middle-age PM; performance of this group did not differ from younger group in any significant way; substantial decline in memory processes associated with strategic memory search for recall of actions is not likely to occur until later life;
• failure of older persons on time monitoring tests has been observed in previous studies;
• number of instructions given to participants clearly outside range that anyone would normally attempt to complete in everyday life; previous research shown that recognising cues in environment is relatively easily accomplished;
• one feature of PM task is that it necessitates an interruption to one's activity in order to carry out the intention; without such an ongoing task many PM tests in laboratory settings would be tests of vigilance and few failures of PM would occur; naturalistic delay periods typically filled with natural distractions and internal cognitions;
• previous studies demonstrated age invariance w.r.t. automatic processes in presence of marked deterioration in controlled or strategic processing;
• older persons took longer to notice cues; reduced efficiency in processing significance of instance of a cue; implicates self-initiated retrospective recall as most sig. reason for failure of PM among elderly;
• test items not screened for age-specific content beforehand; differences in content of items are unlikely to have contributed substantially to PM decline among older people;
• higher level processing may also have a role in decline of PM; age does not necessarily diminish ability to perform vigilance tasks, doing so may affect capacity of older participants to complete a concurrent PM task; this may be strategic decision based on their subjective estimation of likelihood of success on memory task or it may be that focusing on monitoring task diminishes attentional capacity to notice cues or recall intentions;

• Jennings J., Jacoby LL. An opposition procedure for detecting age-related deficits in recollection: telling effects of repetition. Psychology and aging. 1997;12(2):352-61. Available at: http://www.ncbi.nlm.nih.gov/pubmed/9189995.

• McDaniel MA, Einstein GO. Strategic and automatic processes in prospective memory retrieval: a multiprocess framework. Applied Cognitive Psychology. 2000;14(7):S127-S144. Available at: http://doi.wiley.com/10.1002/acp.775.

• Ellis J, Kvavilashvili L. Prospective memory in 2000: Past, present, and future directions. Applied Cognitive Psychology. 2000;14(7):S1-S9. Available at: http://doi.wiley.com/10.1002/acp.767.

• Rendell RG, Craik FI. Virtual week and actual week: Age-related differences in prospective memory. Applied Cognitive Psychology. 2000;14(7):S43-S62. Available at: http://doi.wiley.com/10.1002/acp.770.

• Alderman N, Burgess PW, Knight C, Henman C. Ecological validity of a simplified version of the multiple errands shopping test. Journal of the International Neuropsychological Society : JINS. 2003;9(1):31-44. Available at: http://www.ncbi.nlm.nih.gov/pubmed/12570356.

• Park DC, Hertzog C, Kidder DP, Morrell RW, Mayhorn CB. Effect of age on event-based and time-based prospective memory. Psychology and aging. 1997;12(2):314-27. Available at: http://www.ncbi.nlm.nih.gov/pubmed/9189992.

• Maylor EA. Age-related impairment in an event-based prospective-memory task. Psychology and aging. 1996;11(1):74-8. Available at: http://www.ncbi.nlm.nih.gov/pubmed/8726372.

• Marsh RL, Hicks JL, Landau JD. An invesitgation of everyday prospective memory. Memory and Cognition. 1998;26:633–643. Available at: http://www.psychology.uga.edu/hcpl/pub_pdf/16.pdf.

• 2 x 2 x 2 mixed ANOVA - not a reliable effect of length of delay (surprising); nor of activity during the delay; interaction between age and activity during delay was reliable - filled delays negatively effected PM for younger subjects, slightly increased PM for older adults (not reliable); having a break vs. filled interval didn't make a difference for older adults; results not consistent with view that age-related decrements on RDPM task are due to inability of older adults to actively maintain intention in WM until appropriate time;
• reduced inhibition of distracting thoughts produces difficulties in remembering the retrieved attention
• WM results: younger had sig. higher operation span than older adults;
• hierarchical regression analysis to determine the extent to which the relation b/w age and PM performance could be understood in terms of WM ability; age-related decline in PM remained significant; interesting to note that the one condition in which the relation b/w age and PM was no longer stat. sig. was the condition that paralleled the conditions in Exp. 1. (filled, long delay); for shorter intervals and for unfilled intervals, age-associated factors other than working memory also play a sig. role in RD tasks
• other measures: younger answered more comprehension q's correctly than older

• experiment 3: explore generality of older adults' rapid loss of PM after a brief delay; the sense that task is easy (salient target) could encourage them to forego effortful strategies for maintaining intention in WM throughout the delay; robust drop in performance from immediate to delayed execution for older adults should be eliminated for PM cues that do not support fairly effortless retrieval of the PM intention;
• method: 2 (target-action interval - delay/no-delay) x 2 (attentional demands (standard, divided) mixed factorial design, non-salient targets;
• sig. decline in PM under divided attention condition relative to standard attention condition; effect in delay and no-delay groups; more difficult than high-salient cue conditions in Exp. 1, as divided attention did not disrupt PM in no-delay condition in Exp. 1; result of importance: no decline in PM performance when there was a short delay b/w low-salient cue and instant at which the PM tasks was to be performed;
• retrieval of the PM intention not related to WM capacity; correlations for delayed cells failed to show associations b/w PM memory and WM; in response to difficult initial retrieval, participants may adopt a more strategic approach to maintaining the intention over the delay interval; 60% of participants remembered on some but not all trials;

• discussion: substantial increase in PM forgetting when execution of intention had to be briefly delayed after initial retrieval of intention; decline in PM after brief delay was especially pronounced for older adults, even at short (10s) unfilled delays; no memory loss in Exp. 3 with non-salient target event;
• maintaining an intention over a brief interval is highly susceptible to forgetting difficulty of maintaining an intention over brief delay is sig. amplified for older adults;
• older adults do not have the WM capacity to actively maintain intention over delay interval while performing BTs; declines for older adults were substantial regardless of whether or not the interval was filled, and declines tended to be somewhat more pronounced with shorter 10s intervals than for the longer 30s intervals;
• on the basis of the strength of this memory at the moment of retrieval older adults may be deceived into assuming that controlled processes (rehearsal) are unnecessary, successful maintenance over delays as brief as 10s would require strategic rehearsal; older adults are less likely to initiate strategic rehearsal of the intention because they are unaware of the temporal limits of WM for passive storage of information;
• older adults are thought to have greater difficulty binding features of an event to the spatial and temporal context in which that event occurs, older adults may have fewer spontaneous overt reminders of the PM task over the delay interval; filled delays with a greater similarity with the general contextual features of the PM task would produce more reminding and better PM memory;
• frontal systems are though to maintain associations b/w environmental conditions and intended goals; with frontal decline, these associations are maintained in an deteriorated state, such that current goals are not as tightly connected w/ enabling conditions; salient cue of no greater negative effects of filled delays for older adults and of reliable correlations b/w WM measures and PM in retrieve-delay conditions; maintenance of intention is not strategic and instead dependent on ability of WM to maintain one's current goals in a sufficiently activated state;
• evidence that in naturalistic settings, people will reformulate their prospective plans in response to new environmental constraints;
• although a large proportion of age-related variance in PM was associated w/ WM, there was also sig. age-related variance after controlling for WM; mechanisms that are not obviously related to WM capacity like differences in metamemory and inhibitory control are needed in addition to WM capacity for a complete understanding of age-related decrements in RDPM tasks;
• when initial retrieval is difficult, older participants are more sensitive to challenge of maintaining intentions over brief intervals and ma¥ rely on controlled processes (rehearsal) to keep intention activated over the delay interval; possibility that the difficulty of maintaining an intention over a delay interval may be inversely related to difficulty of initial retrieval;
• easily maintained intentions may be very difficult to maintain over brief intervals;

On Prospective memory (PM) and interruption/distraction

• abstract: age-related PM decrements often dramatic when retrieval was facile; when initial retrieval difficult, older adults showed no forgetting over retention interval

• intro: researchers have found reliable and even substantial age differences in retrieve-execute type of PM task (REPM)
• one variable that seems to affect magnitude of age differences on PM task is demands of background task (BT); resource demands of BT can sig. affect PM, especially for older adults
• age-related deficits in remembering to perform an action could be due to older adults having fewer resources available for the PM task and not to an inherent deficit in retrieving PMs
• when retrieval cue is salient, an automatic retrieval process is stimulated that obligatorily delivers intended action to consciousness
• unclear what to expect when one has to delay execution of intention after retrieval (retrieve-delay PM task / RDPM)
• purpose of present research to examine RDPM performance
• reasonable to propose that reformulating a plan while performing an ongoing task and strategically maintaining an intention would be heavily dependent on working-memory resources
• consistent evidence that increasing working-memory demands of BT decreases PM performance, especially for older adults; relation b/w WM and PM - very little evidence for an association between the two;

• experiment 1: created a delay of 30-39s b/w occurrence of PM target and when it was appropriate to perform intended action; also measured WM capacity; varied presence or absence of secondary digit-detection task;
• highly salient targets used; given possible similarities and differences b/w REPM and RDPM, we could anticipate absence or presence of age-related deficits
• to equate conditions, needed to present the older adults with less material and at a slower rate;
• design was 2 (age) x 2 (delay/no delay) x 2 (attentional demands, WS); reading 3-sentence paragraphs, followed by reading-comprehension questions, two trivia questions; had interest in ability to remember to perform future actions; if they encountered a target word on any of the 3 sentences, they were to immediately press the F1 key (REPM), or to delay until the trivia phase (RDPM); those in divided attention condition given digit-monitoring task, see [Einstein 97];
• 20 paragraphs presented either w/ or w/o digit monitoring task, counterbalanced in each exp. condition, first prospective target appeared in 3rd paragraph, PM targets not presented in consecutive paragraphs; target words always appeared in 2nd sentence of a paragraphs; following this, subjects given vocab. test and WM memory span test;
• results: introducing a delay significantly lowered PM performance in no-delay and delay conditions; in no-delay condition, main effect of age was marginally significant; PM performance not affected by divided attention, no interaction b/w age and divided attention; highly salient target -> near perfect performance in no-delay condition regardless of age and divided attention;
• in delay condition, main effect of age: younger participants had higher PM than older participants, PM performance sig. worse in divide attention, decrement more pronounced for older adults (but no interaction); highly salient targets for REPM to not depend on availability of attentional resources, RDPM tasks poses special difficulties for older adults;
• smaller working memory spans for older adults; WM resources related to maintaining a PM intention over delay interval; WM likely a major factor in mediating relation b/w age and PM; analyses suggest age-related effect in PM primarily associated w/ WM processes;
• other measures: dividing attention interfered w/ comprehension, younger adults had higher scores than older adults, older adults more difficulty w/ digit monitoring

• experiment 2: important purpose of second experiment was to examine reliability of this rapid loss of PM and age-related differences after a delay; under these conditions, maintaining intended action in working memory after retrieval would be more critical for successful PM;
• varied demands of delay activity and length of delay (10, 30s) - delays were filled/unfilled w/ synonym questions;
• older adults do not have working memory resources needed to actively maintain intention through rehearsal over retention interval while simultaneously performing BT; performance on RDPM should be affected by demands of BT, during unfilled delays all participants should be affected by demands of the activities occurring during the delay; during unfilled delays all participants should be able to focus their WM resources on task of maintaining intention, performance should be high for older adults and younger adults;
• [Hasher and Zack 88,94]: ageing disrupts efficient functioning of inhibitory processes, perhaps more distractible, difficulty inhibiting thoughts unrelated to the cognitive demands of maintaining an intention until it is performed; older adults may show large amounts of forgetting even with unfilled delays;
• pilot work indicated that older adults had great difficulty reading items, generating answers, circling answer within allotted time; called out answers instead to experimenter
• 2 (young/old) x 2 (delay, WS) x 2 (filled/unfilled delay, WS) mixed factorial design; delay was synonym task or break for 10 or 30 s; target always occurred in last sentence of paragraph, thus delays b/w offset of sentence containing target and occurrence of target phase were wither 10 or 30s; each participant received 5 short-unfilled, 5 short-filled, 5 long-unfilled, and 5 long-filled delay periods, 4 counterbalanced orderings, equal # of participants assigned to each delay order;
• given WM operation span test afterwards to establish more general support for relation b/w WM and performance on RDPM tasks;

[Einstein 90]

• abstract: subjects given prospective memory test, 3 tests of retrospective memory; no age deficits in prospective memory, no reliable relation between the prospective and retrospective memory tasks; external aids and unfamiliar target events benefit prospective memory performance; results suggest some basic difference between prospective and retrospective memory

• intro: prospective memory (PM) is memory task that requires the greatest degree of self-initiation; typical procedure used to study PM has involved asking subjects to perform some action in naturalistic settings; age-related deficits in PM may not emerge in these settings because subjects may use external cues to help guide retrieval

• experiment 1: lack of an accepted experimental paradigm for studying PM; laboratory method needed for studying PM; compared PM of young/old subjects, varied whether or not they were allowed to create/use an external memory aid
• 2 (age, BS) x 2 (opportunity to form an aid) BS factorial design; PM test embedded within short-term memory task; subjects in external-aid condition given 30s to formulate some type of aid;
• 2x2 BS-ANOVA: retrospective memory, free recall - main effect of age, no effect of memory-aid, no interaction; similar results w/ recognition test; in STM performance, older subjects recalled a greater proportion of word sets than did younger subjects (main effect of age), no effect of memory-aid, no interaction
• PM performance - no effect of age on PM, main effect of memory aid, using an aid -> higher PM score, no interaction;
• most of those in memory aid condition used piece of paper taped to CRT
• relations b/w RM and PM: multiple regression analysis; PM related to memory aid, not related to performance on any RM task
• subjects tended to think about PM task more often during word set presentation than in other phases, elderly subjects admit that they thought less about the PM task than young subjects, this behaviour highly correlated with prospective remembering
• results run counter to view that PM should be especially problematic for older subjects; either memory deficits in elderly are not entirely due to problems in self-initiated retrieval, or certain kinds of prospective memory situations incorporate or have retrieval cues embedded in them

• experiment 2: test again effects of age on PM, test idea that retrieval cues are present and play a role in certain PM tasks
• successful PM depends on triggering of action by target event; PM should be better with an uncommon and unfamiliar target event than with a common and familiar target event; 1/2 subjects asked to press a response key whenever a familiar target event occurred, other 1/2 to press on unfamiliar target event (familiar/unfamiliar words)
• 2 (age) x 2 (familiarity) between-subjects factorial design
• RM performance: free recall reliable effect of age; STM performance - main effect of age;
• PM performance - PM was nearly 3 times higher w/ unfamiliar target than w/ familiar; no effect of age or interaction; PM performance not reliably associated with performance on any RM tasks;
• subjects thought more about PM task during presentation of word sets than during other phases, familiarity x phases interaction; unfamiliar target events increased monitoring ratings relative to familiar target events but only during word-set presentation; neither main effect of age or interactions reliable
• correlations performed b/w monitoring ratings during each phase and PM performance; monitoring ratings during presentation of word sets highly correlated w/ PM performance

• discussion - major finding was lack of age-related deficits in prospective memory; PM seems to be an exciting exception to typically found age-related decrements in memory; reliable age differences in RM
• problem with studying PM in laboratory setting is that demand characteristics could lead subjects to continuously rehearse PM task, an unrealistic PM situation; PM performance was not at ceiling (61% across all subjects)
• no indication that successful PM came at expense of short-term memory performance;
• failure to find age-related differences in PM is that elderly were more motivated than young to perform memory tasks; nonetheless, possible that elderly, relative to the young, saw the prospective memory task as more important and thought about it constantly; [Craik 86] theory that ageing mainly disrupts self-initiated retrieval processes; PM should be especially difficult for elderly because it requires person to remember to remember;
• questionable assumption that all PM tasks are alike in the sense that they are high in self-initiated retrieval, PM tasks vary in the degree to which they require self-initiated retrieval processes for successful memories;
• event-based PM tasks might not produce large age-related effects because they contain external cues; time-based PM tasks that involve more self-initiated retrieval than event-based tasks, might be more likely to produce large age-related deficits;
• no obvious relations b/w RM and PM;
• event-based target events not likely to be thought of as PM stimulus;
• w/ familiar target events, competing associations would be more likely to interfere with PM;

[Einstein 00]

http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Prospective+Memory+and+Aging+:+Forgetting+Intentions+Over+Short+Delays#0

more...

• McDaniel MA, Einstein GO, Stout AC, Morgan Z. Aging and maintaining intentions over delays: do it or lose it. Psychology and aging. 2003;18(4):823-35. Available at: http://www.ncbi.nlm.nih.gov/pubmed/14692867.

• McFarlane D, Latorella K. The Scope and Importance of Human Interruption in Human-Computer Interaction Design. Human-Computer Interaction. 2002;17(1):1-61. Available at: http://www.informaworld.com/openurl?genre=article&doi=10.1207/S15327051HCI1701_1&magic=crossref||D404A21C5BB053405B1A640AFFD44AE3.

• Einstein GO, McDaniel MA. Normal aging and prospective memory. Journal of experimental psychology. Learning, memory, and cognition. 1990;16(4):717-26. Available at: http://www.ncbi.nlm.nih.gov/pubmed/2142956.

• McDaniel MA, Einstein GO, Stout AC, Morgan Z. Aging and maintaining intentions over delays: do it or lose it. Psychology and aging. 2003;18(4):823-35. Available at: http://www.ncbi.nlm.nih.gov/pubmed/14692867.

• Einstein GO, McDaniel MA, Manzi M, Cochran BM. Prospective memory and aging: Forgetting intentions over short delays. Psychology and Aging. 2000;15(4):671-683. Available at: http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Prospective+Memory+and+Aging+:+Forgetting+Intentions+Over+Short+Delays#0.

• Einstein GO, Smith RE, McDaniel MA, Shaw P. Aging and prospective memory: the influence of increased task demands at encoding and retrieval. Psychology and aging. 1997;12(3):479-88. Available at: http://www.ncbi.nlm.nih.gov/pubmed/9308095.

• Henry JD, MacLeod MS, Phillips LH, Crawford JR. A meta-analytic review of prospective memory and aging. Psychology and aging. 2004;19(1):27-39. Available at: http://www.ncbi.nlm.nih.gov/pubmed/15065929.

• McDermott K, Knight RG. The effects of aging on a measure of prospective remembering using naturalistic stimuli. Applied Cognitive Psychology. 2004;18(3):349-362. Available at: http://doi.wiley.com/10.1002/acp.977.

more...

• Tran Q, Calcaterra G, Mynatt E. How an Older and a Younger Adult Adopted a Cooking Memory Aid. Proc. Human-Computer Interaction International (HCII). 2005. Available at: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.78.3784&rep=rep1&type=pdf.

• McFarlane D, Latorella K. The Scope and Importance of Human Interruption in Human-Computer Interaction Design. Human-Computer Interaction. 2002;17(1):1-61. Available at: http://www.informaworld.com/openurl?genre=article&doi=10.1207/S15327051HCI1701_1&magic=crossref||D404A21C5BB053405B1A640AFFD44AE3.

• M. Czerwinski, E. Horvitz, and S. Wilhite, "A diary study of task switching and interruptions," Proceedings of the 2004 conference on Human factors in computing systems - CHI '04, vol. 6, 2004, pp. 175-182.

• D.D. Salvucci, N.A. Taatgen, and J.P. Borst, "Toward a unified theory of the multitasking continuum: From concurrent performance to task switching, interruption, and resumption," CHI 2009, 2009, pp. 1819-1828.

• Nagel KS, Hudson JM, Abowd GD. Predictors of availability in home life context-mediated communication. Proceedings of the 2004 ACM conference on Computer supported cooperative work - CSCW '04. 2004;6(3):497. Available at: http://portal.acm.org/citation.cfm?doid=1031607.1031689.

On Prospective memory and interruption/distraction

• Einstein GO, McDaniel MA. Normal aging and prospective memory. Journal of experimental psychology. Learning, memory, and cognition. 1990;16(4):717-26. Available at: http://www.ncbi.nlm.nih.gov/pubmed/2142956.

• McDaniel MA, Einstein GO, Stout AC, Morgan Z. Aging and maintaining intentions over delays: do it or lose it. Psychology and aging. 2003;18(4):823-35. Available at: http://www.ncbi.nlm.nih.gov/pubmed/14692867.

• Einstein GO, McDaniel MA, Manzi M, Cochran BM. Prospective memory and aging: Forgetting intentions over short delays. Psychology and Aging. 2000;15(4):671-683. Available at: http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Prospective+Memory+and+Aging+:+Forgetting+Intentions+Over+Short+Delays#0.

• West R, Craik FI. Age-related decline in prospective memory: the roles of cue accessibility and cue sensitivity. Psychology and aging. 1999;14(2):264-72. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10403713.

• Einstein GO, Smith RE, McDaniel MA, Shaw P. Aging and prospective memory: the influence of increased task demands at encoding and retrieval. Psychology and aging. 1997;12(3):479-88. Available at: http://www.ncbi.nlm.nih.gov/pubmed/9308095.

• Morrow DG, Leirer VO, Altieri PA. Aging, expertise, and narrative processing. Psychology and aging. 1992;7(3):376-88. Available at: http://www.ncbi.nlm.nih.gov/pubmed/1388858.

• Plude DJ, Hoyer WJ. Age and the selectivity of visual information processing. Psychology and aging. 1986;1(1):4-10. Available at: http://www.ncbi.nlm.nih.gov/pubmed/3267377.

• Henry JD, MacLeod MS, Phillips LH, Crawford JR. A meta-analytic review of prospective memory and aging. Psychology and aging. 2004;19(1):27-39. Available at: http://www.ncbi.nlm.nih.gov/pubmed/15065929.

• Carlson MC, Hasher L, Connelly SL, Zacks RT. Aging, distraction, and the benefits of predictable location. and Aging. 1995;10(3):427-436. Available at: http://www.psych.toronto.edu/users/hasher/PDF/Carlson et al 1995 Psych and Aging.pdf.

• McDermott K, Knight RG. The effects of aging on a measure of prospective remembering using naturalistic stimuli. Applied Cognitive Psychology. 2004;18(3):349-362. Available at: http://doi.wiley.com/10.1002/acp.977.

• study 1: Einsteing et al 97, Light et al 00, McDermot 04: older adults reduced capacity to make effortful searches through memory; memory search sometimes automatic when a strong contextual connection exists b/w cue and instruction (i.e. buy coffee at starbucks); sometimes cue is identified without conscious awareness;

• possible that unfamiliar target would be easier to detect than familiar one; Einstein 90, McDaniel 93: unfamiliar words have been found to produce more reliable prospective remembering than familiar ones;

• McDaniel 03, Einstein 00 - brief delays b/w cue recognition and opportunity to act impair older persons' performance; older persons report difficulty remembering what they were doing before they were interrupted

• study 3: examined effect of distractions and interruptions on the shopping errands task; prompted by findings that older persons are more affected by irrelevant stimuli than younger persons in complex visual search environments (Carlson 95, Morrow 92, Plude 86, Rabbitt 65) - inability to inhibit irrelevant inputs (Hasher 88)
• West 99 - momentary lapses in intention; brief lapses in vigilance rather than an inability to remember task instructions; brief delays in opportunity to execute an intention caused deficits in performance of older adults
• interrupted by having a conversation with another person - introducing 4 verbal fluency tasks during the movement down the street; memory load has been shown to affect memory for intentions in older persons; hypothesised that older persons would display a deficit in the reinstatement of working memory for the shopping task
• 2-factor, condition (distraction, interruption, control) x age group ANOVA; significant age and condition main effects - ability for older group to detect cues in interruption condition was significantly reduced; only significant differences were for the older group, between the control and interruption conditions and the distraction and interruption conditions; ANOVA conducted for probability of recalling the tasks correctly - only sig. main effect was age; time/page data ANOVA - age was significant; no evidence that older group slowed their time of inspection per page to compensate for presence of any additional distraction; age by condition interaction - control group accessed more pages for each location correct than either the distraction or interruption groups;
• irrelevant visual and auditory distraction to the situation had little effect; relatively easy for participants to inhibit this auditory modality, since it is clearly not relevant to the shopping task, which relies on visual cues
• interrupted subsequently less likely to notice cues in the street; condition-specific impact on recall; if cue was noticed then probability of remembering the associated action was unchanged; recognising cues on the other hand is dependent on conditions that arise in the environment;
• McDaniel 03, Einstein 00 - brief delays b/w cue recognition and opportunity to act impair older persons' performance; older persons report difficulty remembering what they were doing before they were interrupted

• discussion - older persons have difficulty remembering intentions, even when environment and tasks are familiar and they can move at their own pace;
• older persons were specifically impaired in the memory search component of the task relative to detecting cues; recall more effortful than attentional processing in this situation; self-initiated memory search is impaired in old age;
• familiarity had no effect on accuracy or speed of movement for young/old persons
• older persons not specifically disadvantaged when asked to recognise a relatively small number of distinct cues in an environment that, although unfamiliar, is within their range of experience
• older persons responded by blocking out distractions, where the distraction could not be ignored, ability of older persons to notice cues was significantly disadvantaged, although meta-strategy was intact; this had the effect of disrupting the prospective vigilance or monitoring component of the task, while having no specific effect on the memory search component
• older persons may be aware that memory generally declines with age, translating this into a realistic expectation of performance on a specific or a novel task to be difficult
• limitation: time frame of the task (1 hr);
• future work: allow strategic planning, move backwards and forwards
• circumstances where attentional and memory strategy are self-determined, likely that kind of difficulties that come with age or are inflicted by brain injury will become more pronounced

[Farrimond 06]

Farrimond S, Knight RG, Titov N. The effects of aging on remembering intentions: performance on a simulated shopping task. Applied Cognitive Psychology. 2006;20(4):533-555. Available at: http://doi.wiley.com/10.1002/acp.1202.

• abstract: task measuring memory for delayed intentions - memory search but not cue detection affected in older persons, no age specific effect on cue detection, capacity for self-initiated reinstatement of working memory is reduced in old age

• intro: Henry et al 2004 confirms that prospective memory is impaired in old and and that is is particularly evident on tasks that required effortful or strategic processing
• older persons specifically impaired on memory and attentional tasks that rely on effortful, conscious or controlled processing;
• Henry et al 04 found that older groups superior to younger groups on naturalistic time-based tasks, decline on laboratory tasks
• examined three aspects hypothesised to be sensitive to age: opportunity to learn instructions, familiarity with environment, presence of distractions; distinction b/w process of recognising a cue and recalling associated action (relies on self-initiated memory search)
• number of instructions and length of testing sessions kept at reasonable level

• study 1: Einsteing et al 97, Light et al 00, McDermot 04: older adults reduced capacity to make effortful searches through memory; memory search sometimes automatic when a strong contextual connection exists b/w cue and instruction (i.e. buy coffee at starbucks); sometimes cue is identified without conscious awareness;
• ability to control activity may compensate for reduced ability of older group to complete prospective tasks
• either one or three learning trials (cue and associated task) - hypothesised that 1-trial group would have disproportionate negative effect for older participants on more effortful process of recalling actions relative to their ability to recognise cues;
• virtual street used - speed with which they move along street recorded, and time on each page; cannot go backwards
• expected to recall task aloud in response to cue - correct action recorded
• 3 factor ANOVA - age x # learning trials x task component (RM - cue and task); sig. main effect for age, # learning trials, component, learning trial and component interaction, component x age interaction, age x learning trial x component interaction approached significance, time significant for age, not slower in 3-trial condition, no significant age x learning trial interaction, both groups more efficient when they had 3 learning trials
• older group less likely than younger group to recall task associated with cue, older group more affected than younger group when only had 1 learning trials; reduced opportunity for encoding, self-initiated memory search more difficult
• superiority of memory recall in younger persons, enhanced ability to learn instructions, older persons can perform as well as younger group when given enhanced opportunity to learning instructions.
• older persons tended to be about 3-4 min slower

• study 2: determine whether recognising cues and associated task is more difficult in unfamiliar environment for older people than younger people; assess how ageing affects vigilance and remembering in unfamiliar naturalistic environments
• possible that unfamiliar target would be easier to detect than familiar one; Einstein 90, McDaniel 93: unfamiliar words have been found to produce more reliable prospective remembering than familiar ones;
• also interested in # of instructions they were likely to remember, older people have more realistic appreciation of what they are likely to remember
• procedure used was same as in first study, except no practice session
• 3-factor ANOVA - age x familiarity x component (cues vs tasks RM); significant age and component main effects, no effect of familiarity, significant component x group interaction; older persons have more difficulty recalling actions than cues relative to younger group
• significant age effect for time/page, no familiarity effect, no age effect for page/location; suggested that participants tended to be less efficient in their movements around the unfamiliar street
• participants in both groups underestimated how much they would remember, trend towards the participants predicting a poorer performance in the unfamiliar street
• older persons not affected by unfamiliarity of the environment, neither speed nor accuracy of cue detection affected; environment might not have been sufficiently alien to cause differences in attentional strategies
• younger group was significantly more conservative in estimates of what they would actually achieve than the older adults, expectations of remembering shopping instructions do not decline with age

• study 3

• Tran Q, Calcaterra G, Mynatt E. How an Older and a Younger Adult Adopted a Cooking Memory Aid. Proc. Human-Computer Interaction International (HCII). 2005. Available at: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.78.3784&rep=rep1&type=pdf.

• McFarlane D, Latorella K. The Scope and Importance of Human Interruption in Human-Computer Interaction Design. Human-Computer Interaction. 2002;17(1):1-61. Available at: http://www.informaworld.com/openurl?genre=article&doi=10.1207/S15327051HCI1701_1&magic=crossref||D404A21C5BB053405B1A640AFFD44AE3.

• Findlater L, McGrenere J, Modjeska D. Evaluation of a role-based approach for customizing a complex development environment. Proceeding of the twenty-sixth annual CHI conference on Human factors in computing systems - CHI '08. 2008:1267. Available at: http://portal.acm.org/citation.cfm?doid=1357054.1357251.

[Mihailidis 07]

• prompting device that uses AI to automatically monitor an older adult in a self-care facility - four integrated systems: tracking, state monitoring, Markov decision process (MDP) policy, prompting modules
• ref: Issue 10 of IEEE transactions on information technology in biomedicine
• levels of prompting specificity were minimal ("use the towel"), moderate ("use the blue towel"), specific ("John, use the blue towel on your right")
• inclusion criterion was clinical diagnosis of moderate to severe dementia - modified withdrawal-type single subject research design
• DVs: number of hand-washing steps that the participant completed without any interaction with a caregiver, number of interactions between the human caregiver and the participant for successful completion of hand-washing
• observed that this participant's level of dependence on the caregiver decreased when the device was introduced, increased when device was removed: potential decreased in caregiver workload
• participant ignored a large percentage of prompts from the device, not able to fully understand directions being given - responded positively to only 6% of the them, five or more prompts from the system were often required before she responded properly or the caregiver was required to enter and assist her, especially during the "use the soap" step

more...

• Boger J, Hoey J, Poupart P, et al. A planning system based on Markov decision processes to guide people with dementia through activities of daily living. IEEE transactions on information technology in biomedicine : a publication of the IEEE Engineering in Medicine and Biology Society. 2006;10(2):323-33. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16617621.

[Salvucci 10]

D.D. Salvucci, "On reconstruction of task context after interruption," Proceedings of the 28th international conference on Human factors in computing systems - CHI '10, 2010, p. 89.

• resumption not as a primarily memory-based process, but rather one of reconstruction - perceptual re-encoding of the information necessary to perform the task

• research grounded in cog.psyc. aimed to elucidate cog. underpinnings of task suspension and resumption, focusing on resumption as memory-based process: memory for goals, interrelated encoding and retrieval processes
  • measure using resumption lag - time needed to resume primary task - higher in applied studies, often making memory-based accounts difficult or implausible, almost certainly involving more than a single memory retrieval
• theoretical framework for thinking about task resumption as process of reconstruction, requiring visual re-encoding to reconstruct task context immediately prior to interruption

• ACT-R model - integrates various sub-theories of human cognitive, motor, and perceptual processes - knowledge as declarative chunks of information, procedural rules, framework to develop computational models of user behaviour and to measure behaviour and performance; mental resource known as problem state; most tasks do require the creation and updating of problem state during the execution of a task - a distinct mental resource
  • implications for understanding of task interruption and resumption; problem state can only be maintained for a single task at a time, at interruption, user stores away the current task's problem state and replaces it with a new problem state for the interrupting task - eventually retrieve the original task's problem state
  • memory decay will, over time, make problem-state information difficult or impossible to recall - failure to recall problem state leads to reconstruction of the problem state or a replication of the lost problem state, involves re-reading text, scanning paper notes, phoning someone for information, domain dependence of reconstruction process

• illustrative models of reconstruction - two tasks used in recent studies on interruption - route-planning and document-editing, interrupted by news-reading task in which they read a news article and decided on an appropriate title for the article
• determined "best" and "worst" possible points of interruption
• task analysis was conducted to determine the reconstruction and action steps are needed to recover the pre-interruption problem state; ACT-R used as it allows for better generalisation to broader theories of multitasking
• resumption lag represents the time needed to perform the first observable action in the primary task, and thus includes reconstruction as well as subsequent actions until the point of this first observable action
• straightforward analysis of reconstruction steps leads to a good estimate of behavioural measures: reconstruction can be a large piece of the total resumption lag

• discussion: domains involving a long interruptions and/or complex mental states are the most likely to rely on reconstruction; but no easy domain-general way of specifying the reconstruction processes
• memory and reconstruction based resumption processes are not mutually exclusive - reconstruction process may involve cueing of relevant memories by the external environment

• Nagel KS, Hudson JM, Abowd GD. Predictors of availability in home life context-mediated communication. Proceedings of the 2004 ACM conference on Computer supported cooperative work - CSCW '04. 2004;6(3):497. Available at: http://portal.acm.org/citation.cfm?doid=1031607.1031689.

• Jones K. ENABLE Report: enabling technologies for people with dementia: Report of the assessment in ENGLAND. Quality. 2004;(January).

• Holthe T, Engedal K. ENABLE Report: enabling technologies for people with dementia: National Report on Results From Norway. Information Technology. 2004;15(3).

• Topo P, Saarikalle K, Maki O, Parviainen S. ENABLE Report: enabling technologies for people with dementia: Report of Assessment Study in Finland. Proceedings of the... annual Conference on Rehabilitation Technology. 2004.

HCI & Universal Usability

On universal usability (general)

[Inkpen 01]

K.M. Inkpen, "Drag-and-drop versus point-and-click mouse interaction styles for children," ACM Transactions on Computer-Human Interaction, vol. 8, 2001, pp. 1-33.

• referred to by Joanna; potential application for PwDs with C-TOC

• abstract - drag/drop + point/click interaction styles compared via field study analysis, controlled experiment, state-transition diagrams, GOMS analysis; point/click superior in terms of speed/error/preference

• intro - some prominent user interface designs for adults may not be appropriate for children; children have difficulty w/ marquee-type selection, difficulty maintaining pressure on the mouse for extended periods; gender difference often exist with respect to interactions with computers;
• prev-work: dragging task was slower than a pointing task and more errors were committed during a dragging task than during a pointing task;

• exp 1 setup: IBM version of game used point-and-click, Mac version used drag-and-drop - muscular tension req'd to keep mouse button down while dragging; both use visual feedback of object being moved during move stage; drop errors in drag-and-drop move object back to starting location; drop-errors w/ point-and-click drop at current location;
• goal of exp 1 - interaction style's effect on game achievement (total # of puzzles solved) and motivation (whether or not game was played for full 30min session)
• sig.dif. in proportion of girls who were unable to solve any puzzles using the point-and-click interface and the drag-and-drop interface - 25% vs. 49%
• sig.dif. in # of puzzles solved in each of the experimental condition: P/C: 1.66, D/D: 1.15; S/C girls more successful playing the game;
• sig.dif in motivation in favour of S/C vs. D/D
• discussion: enhanced realism (real game), implementation details inconsistent - may have interacted w. results

• exp. 2 examines P/C and D/D in isolation of a game - simplified software environment, w/o possibly confounding factors, less realistic;
• 8 dif. possibilities of distance b/w source and (2) target x (2) source size x (2) target size x (2) interaction style - each combination appeared once / trial, 16 trials / block, order of appearance of each combination was random within each block;
• when source box picked up, visual feedback given as in exp. 1; in drop errors: S/C - source remains picked up, D/D - source returns to original position, forcing another pickup action;
• order of interaction style counterbalanced within each gender to evenly distributer practice effects; instructed to performa task as quickly as possible w/o any mistakes;
• dep. factors: MT, E, preference;
• sig. main effect for interaction style in MT (S/C faster than D/D); no effect of gender; main effects for target size and target distance also significant, source size also significant;
• sig. main effect for interaction style in E, (S/C less errors than D/D); sig. main effect of error type (mire pickup errors than drop errors), sig. interaction effect b/w type of error and source or target size, w/ size of source box sig. effecting # drop errors; no sig. main effect of errors;
• for pickup errors: sig. main effect for interaction style and size of source; sig. interaction of distance, only for long, not for short distances;
• no sig. main effect for interaction style for drop errors, target size was sig., gender interacted w/ target size - interaction style was only sig. factor on # drop errors for girls, target size was sig. factor on # of drop error for boys;
• Fitts' law analysis + linear regression: D/D - larger MTs than P/C;
• no sig. interaction of gender b/w preferences of girls and boys for interaction style - sig. more children preferred P/C over D/D - found P/C easier; D/D made fingers and hands tired; those who preferred D/D were familiar from using it at home; preferred the tactile feedback it provided; Buxton: kinaesthetic connectivity can help to reinforce conceptual connectivity of subtasks within a compound gesture;
• analysed P/C and D/D using state-transition diagrams; also GOMS style analysis in terms of goals (moving source to target), operators (mouse movements and clicks), methods (interaction style), selections; D/D would be slower than P/C; D/D has an additional mouse button operation and two extra mouse movement operations;
• also used state-transition and GOMS models to explain observations from Exp. 1; errors made in D/D result in more operations than in P/C

• conclusion - P/C more effective than D/D; Fitt's shows that dragging is slower and more error-prone than pointing, also effected by size of target and distance to target; children may have difficulty due to physical requirements needed to maintain constant pressure on the mouse button; research on adults: dragging task is slower and more errors made as compared to pointing task;

[Shneiderman 00]

http://portal.acm.org/citation.cfm?id=332833.332843&dl=GUIDE&dl=ACM&idx=332833&part=periodical&WantType=periodical&title=Communications

• Comment: This article's contribution to the field is far-reaching and addresses many forms of usability concerns. In the past 10 years, we have seen the rise of broadband internet access, the One-Laptop-Per-Child initiative (i.e. the XO-1 laptop), and the creation of many diverse online communities cited in the article. In addition, we cannot forget the rise of Web 2.0 and social networking, which have undoubtedly adhered to principles of universal usability. However, with the technical divide shrinking, a set of universal usability problems still persist today, such as concerns regarding net neutrality.

[Ho-Ching 03] - Auditory Impairments

Ho-Ching FW, Mankoff J, Landay JA. Can you see what i hear?: the design and evaluation of a peripheral sound display for the deaf. In: Proceedings of the SIGCHI conference on Human factors in computing systems. ACM New York, NY, USA; 2003:161–168. Available at: http://portal.acm.org/citation.cfm?id=642641.

• Comment: I expect that efficient wireless microphone arrays, especially those requiring little calibration, could eventually fall within a reasonable cost range, thus enabling users to use sound visualization systems such as the Ripple display. On the other hand, I expect that ubiquitous ambient detection and notification devices may prove to be more viable, and enjoy greater use among both hearing and non-hearing users. For example, imagine a kettle equipped with the capability to send a notification over a local network to a desktop or mobile device, informing the user that water has boiled. Alternatively, sensors under doormats or inside a door detect and notify when a person approaches. If we consider Moore's law and the decreasing costs of such devices, I wouldn't doubt the possibility of a house or office containing many devices with detection and notification abilities.

[Wobbrock 03] - Motor Impairments

Wobbrock JO, Myers BA, Kembel JA. EdgeWrite: a Stylus-Based Text Entry Method Designed for High Accuracy and Stability of Motion. In: Proceedings of the 16th annual ACM symposium on User interface software and technology.Vol 5. ACM; 2003:70. Available at: http://portal.acm.org/citation.cfm?id=964696.964703.

• Comment: I curious as to whether a similar system exists that will support non-Roman alphabets; the Chinese alphabet for instance contains thousands of unique characters. From my understanding of written Chinese, many characters can be divided into halves or quarters (left and right, top and bottom), with more simplified glyphs in each half or corner. Perhaps an variation of EdgeWrite with a 2 x 2 grid of adjacent writing squares could facilitate input for Chinese and other languages.

[Jacko 02] - Visual Impairments

Mohan Raj, Presenter

Jacko JA, Scott IU, Sainfort F, et al. Older Adults and Visual Impairment: What Do Exposure Times and Accuracy Tell Us About Performance Gains Associated with Multimodal Feedback? CHI. 2003;(5):33-40.

• Comment: It is no surprise to me that redundantly coding feedback across multiple modalities for drag-and-drop interactions is beneficial to all types of users (normal vision and impaired vision) - redundancy encoding has also been discussed at length in CPSC 533C (Information Visualization). Inspired by what was learned in that course, I am curious as to how visually impaired individuals (i.e. those with AMD) respond to animation as a form of redundant feedback. For instance, during a drag-and-drop operation of a file to a folder, it is typical for a small animation (i.e. the folder opening) to occur when the file icon is placed over the target folder. Could animations such as this, or easily-detectable animation such as a flashing icon be justifiably added to multi-modal feedback patterns for visually-impaired users?

more...

• K.A. Moffatt, "Addressing Age-Related Pen-Based Target Acquisition Difficulties," Ph.D thesis, 2010.

• S. Zhai, J. Kong, and X. Ren, "Speed-accuracy trade-off in Fitts' law tasks — On the equivalency of actual and nominal pointing precision," International Journal of Human-Computer Studies, vol. 61, 2004, pp. 823-856.''
  • comment: from Karyn Moffatt's thesis: documents user performance differences between five wordings of speed/accuracy task instructions for a Fitt's-like task.

• Emery VK, Edwards PJ, Jacko JA, et al. Toward achieving universal usability for older adults through multimodal feedback. ACM SIGCAPH Computers and the Physically Handicapped. 2002;(73-74):53. Available at: http://portal.acm.org/citation.cfm?id=957214.

On designing technology for older users

CPSC 544 - Universal Usability: Healthy Older Adults

• 544_Presentation_Universal_Usability_Health.pdf: CPSC 544 Research Topic Presentation (10/13/09)
• 544_A1_brehmer.pdf: CPSC 544 Research Topic Paper

Additional 544 References (not presented)

• Ellis RD, Kurniawan SH. Increasing the Usability of Online Information for Older Users: A Case Study in Participatory Design. Computer Law. 2000;16(3):180-186.

• Goodman J, Dickinson A, Syme A. Gathering Requirements for Mobile Devices using Focus Groups with Older People. In: Designing a More Inclusive World, Proceedings of the 2nd Cambridge Workshop on Universal Access and Assistive Technology (CWUAAT), Springer. Citeseer; 2004:1-10. Available at: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.130.7025&rep=rep1&type=pdf.

• Kurniawan SH. An exploratory study of how older women use mobile phones. UbiComp 2006: Ubiquitous Computing. 2006:105-147;122. Available at: http://www.springerlink.com/index/R7G4V0145304J187.pdf.

• Rosson M, Carroll J, Seals C, Lewis T. Community design of community simulations. In: Proceedings of the 4th conference on Designing interactive systems: processes, practices, methods, and techniques. ACM; 2002:75–83. Available at: http://portal.acm.org/citation.cfm?id=778726.

• Davis H, Vetere F, Francis P, Gibbs M, Howard S. “I Wish We Could Get Together”: Exploring Intergenerational Play Across a Distance via a 'Magic Box'. Journal of Intergenerational Relationships. 2008;6(2):191-210. Available at: http://www.informaworld.com/openurl?genre=article&doi=10.1080/15350770801955321&magic=crossref||D404A21C5BB053405B1A640AFFD44AE3.

more...

• N. Selwyn, S. Gorard, J. Furlong, and L. Madden, "Older adults' use of information and communications technology in everyday life," Ageing and Society, vol. 23, 2003, pp. 561-582.

• P. Gregor, A.F. Newell, and M. Zajicek, "Designing for dynamic diversity - interfaces for older people," Proceedings of the fifth international ACM conference on Assistive technologies - Assets '02, 2002, p. 151.

• P. Gregor and A.F. Newell, "Designing for dynamic diversity - making accessible interfaces for older people," Proceedings of the 2001 EC/NSF workshop on Universal accessibility of ubiquitous computing providing for the elderly - WUAUC'01, 2001, p. 90.

• A.F. Newell and P. Gregor, "Design for older and disabled people – where do we go from here?," Universal Access in the Information Society, vol. 2, 2002, p. 3–7.

• J. Rowe and R. Kahn, "Successful Aging," The Gerontologist, vol. 37, 1997, pp. 433-440.

• N. Charness, "Aging and human performance," Human factors, vol. 50, 2008, p. 548.

On designing technology to address users w/ cognitive/memory impairments

[Lee 07]

• abstract: improved understanding of caregiver support for EMI for design of life-logging technologies, reduce burden on caregiver, allow for better care to be provided;

• rich multimedia memory cues to help users remember their personal experiences better; performed observational study;

• related work
  • focus on how caregivers support need for person w/ EMI for information about their past experiences;
  • reminiscence therapy for EMI; burden caregiver w. recording photos/videos for memory cues
  • lifelogging technologies - video, audio, photos; continuously record a comprehensive account of the user's experience; captured media can be indexed w/ other contextual sensor data such as GPS, accelerometer, body sensors; record when triggered by sensor - systems generate a lot of data for caregiver to sort through that may not be helpful for supporting recollection

• study 1 - use of memory cues - 5 dyads of patients and caregivers, shadowed 4 of them - ranged from MCI to mild AD to moderate AD; attended significant events (i.e. weddings, memorials, etc.)

• results - use of memory cues
  • caregivers engaged in dialog using cues, only when there is a reasonable chance of successful cued recall, sometimes too laborious to correct mistakes or too distressing for person w/ EMI to be corrected; caregivers usually reveal cues in piecemeal fashion until persons w/ EMI can recall rest of episode w/ their own memory; clinical evidence that engaging in such cognitively stimulating mental exercise can slow progression of cog. decline
  • question of whether technology can assist overburdened caregiver in identifying and patiently presenting good cues;

• study 2 - identifying good memory cues
  • w/ an understanding of what good cues are, these extraction techniques can be better designed to automatically select good cues and reduce burden on caregiver
  • card-sorting technique w/ photos automatically taken by wearable digital camera; validated in a follow-up session at least one month after each experience
  • psychological models of autobiographical memory shown that info about an experience such as participants, locations, and time periods can be used as cues to retrieve generalised experience's from one's history
  • interviewed participant and asked them to reflect on what they though were good cues for memory

• results - identifying good memory cues
  • cues must either be memorable or at least recognisable; distinctive or personally significant: people, objects, places, actions
  • qualitative differences in selection strategies and abilities b/w participants and w/ and w/o EMI
  • distinction b/w person, object, place, and action based cues and experiences - most important details of a particular experience can be represented by one type of cue, majority type cue; incidental type cues (weather, clothing worn, daily mood) not found to be important
  • detail recognised or recalled from an experience can be used as a cue to retrieve other temporarily or semantically related details of the experience
  • distinctive details, especially those different from people's normal expectations, usually more memorable than less distinct details - consistent w/ psychological understandings of distinctiveness effects in memory; can be unusual or unexpected details of the experience, helpful for distinguishing a particular experience from similar experiences that fit the same general schema; details that stand out indicating that the unexpected details are good for triggering memory; do not always have to be unusual but can be prototypical of a distinctive experience, used to recall other prototypical aspects of the experience and to remember a particularly distinctive experience
  • personal significance - richer, more deeply encoded memory trace in brain, easier for subsequent retrieval - rich recollection of experience
  • differences b/w those w/ and w/o EMI - those w/ EMI likely to be more cognitively overloaded, limit @ of cues, only use most effective cues, basic orienting cues before more specific cues, those w/o EMI can use cues that are more subtle / less central to the experience

• discussion
  • problems: lack of rich cues in addition to caregiver's descriptions, recurring need to engage in cueing process
  • technology to reduce burden of repetitive support for episodic memory, allowing caregiver to select multimedia cues from experience and create digital narrative that person w/ EMI can review on their own to help recollect;
  • lifelogging tech usually errs on side of comprehensiveness, limited cog. resources to devote to reviewing
  • automatic summarisation techniques to reduce amount of data by extracting most effective cues; can they be identified automatically?
  • knowing the type of experience can help determine what types of cues are most appropriate to be captures and presented to best support recollection of the experience; systems should automatically determine these; resulting in more lightweight, less invasive capture systems, less overwhelming accounts of people's experiences that are still / more effective;
  • personally significant details difficult for technology to identify - needs history of interests - caregiver intervention? sensing physiological arousal? GSR? brain-waves? potentially invasive or embarrassing to wear - not socially acceptable
  • design to leverage expertise of caregiver in identifying good cues
  • caregivers can use their own memory of experience to draw out important cues and retrieve from system by searching / browsing;
  • lifelogging to benefit those w/ EMI before condition worsens - intelligent s¥stems to use personal history to figure out what details of new experiences are distinctive or personally significant;
  • provide caregivers or users ability to delete any captured data;
  • methodology issues: variability or real-world experiences

• conclusion
  • what is most effective modality for capturing data and presenting cues?
  • social and aesthetic issues for logging / presenting

[Sevilla 07]

• abstract: need for a cognitively accessible web; recommendations about cognitive disability are extremely difficult (if not impossible) to test; proposal of a simplified web browser and adequate web design;

• intro: according to web content accessibility guidelines (WCAG) of the WAI, disorders such as dyslexia, dyscalculia, ADHD, intellectual impairments schizophrenia are included - participants w. barriers to participation and learning- those with intellectual disorders emphasised; within a web page with AAA qualification there is no need to use special elements which indicate to web users that they have to make a choice and, however, many people with cognitive disabilities need this info to be explicit
• need for universal design (flexible and inclusive design, not one-size-fits-all)- same means of use for all users, provide adaptability to user's pace eliminate unnecessary complexity, use different modes for redundant presentation of essential information
• guidelines for easy navigation design: content and navigation - avoiding excess information, provision of linear navigation, stable navigation menus, quick navigation style; supports and help - use adapted language, provide solution for errors, provide alternative supports for making comprehension easier; design and style - clear typography, alternative texts for images, control over multimedia elements
• not to present more than 10-15 items on the same screen, make it easier for user to return to starting point
• make content text legible and understandable, make placement and functionality of content predictable

• cognitively accessible web proposal - focus on choice-making and browsing
• choice making: essential for autonomous navigation and gaining access to wide and growing range of web content
• browsing: visually showing all the stages that should be carried out to perform a task, w. image of starting and finishing point always present during performance
• individual preferences: adapting the content to individual preferences, reducing complexity
• design proposed: designed a specific web navigation and vis. system
  • elimination of browser menu and controls
  • use of scroll avoided
  • icons are large, descriptive text on mouse-over, horizontal and vertical alignment, alternative image on mouse over, alternative mouse pointer on mouse over
  • selection pages: show list of alternatives w/ descriptive text at top of page, audio instructions, allow user to distinguish actual options from page controls, options represented by page centred images, images are distributed around a 'selection' pictogram, large size, max # of choices = 5 (with page controls, max is 7), descriptive text of 12pt font below option images, every control has alternative text on mouse-over; selection pictogram indicated both the action to be done (choice) and the kind of page we are on - no alternative image or mouse pointer and is not clickable
  • browse pages - graphic aid emphasises the step-by-step structure idea and helps the users in the navigation process - first, last, and active step highlighted; action buttons complemented w/ additional descriptive audios created for supporting those who cannot read

• experimental design / development - generated adapted web pages automatically by making use of web ontology parser
• hypotheses: proposed design is usable and accessible by people w/ a range of impairments; underlying arch. of the design is appropriate for giving steps towards a testable protocol on cog. accessibility; level of usability of this adapted design is higher than the level of a typical web
• experimental design: 2 consecutive phases - conventional vs. adapted web - N = 1 experimental design w/ 20 replications
• specified goals with effectiveness, efficiency, and satisfaction - ad hoc measure of usability that includes measures of its components - also measured the interaction of the individual with cognitive impairment w/ accompanying teacher in order to complete the comparison study between the adapted and conventional version
• participants - those w/ mental retardation - cannot be generalised to whole spectrum of individuals with cog. deficits;
• results: adapted version found to be much more usable, no difference in satisfaction, no difference in interaction w/ teacher
• correlation of -0.473 between IQ and scores on web understanding, w/ adapted version no correlation b/w IQ and scores of any measures of usability

• discussion - is it necessary to have different levels of cog. accessibility?
• all participants had better performance on adapted version - level of mental retardation not most appropriate scale for determining levels of cog. accessibility
• adapted version very repetitive and boring due to reduced # of choice
• further investigation required into specific cog deficits - matter learning to navigate the web / matter of performing the navigation

[Hawkey 05]

• abstract: alleviating repetitive questioning behaviour

• intro: current guidelines - writing the answer down and directing the PwAD to the written answer; having them find the answer for themselves; providing reorientation with written reminders, signs, clocks, calendars, whiteboards, as well as through familiar objects and memories; misleading or distracting them if the true answer could cause distress
• explored dimensions of repetitive questioning in an effort to provide a user and needs analysis for a proposed technological intervention for the problem; interested in whether or not those w/ AD would be willing to use technology, the barriers to use the technology for this population, effectiveness of technology at relieving some of the caregiver stress associated with repetitive questioning behaviour;

• repetitive questioning study: what are the information needs of AD patients? what are their best modes of interaction? how can we provide information in a manner that is timely, informative, relevant? how can caregiver input the info w/ less stress and effort than answering the PwAD's questions?
• multi-method approach: diaries, semi-structured interviews;

• results: primarily inductive analysis approach
• information needs: time (concept of time for AD patients seems to change); schedule (TV schedules not used), reminders limit independence, but ay be helpful, limit advance notification of an event to reduce questioning - family gatherings should be known far in advance; current event details, information (i.e. misplaced items, recent history, brief biographies w/ photos of people; opinions / feedback - reinforcement questions;
• interaction abilities: not able to use a computer by themselves, mechanical interactions and remotes too complex, difficulty w/ the telephone; problems initiating activities, linguistic difficulties - reading absorption and comprehension, following a plot; stopped writing or are writing unintelligibly; generally found to have intelligible speech - difficulty coming up with right word, forming complex questions, difficulty understanding, problems w/ pronunciation;

• discussion: fluid reality of AD patients; denial vs. face-saving; competing interests of PwADs and CGs; PwAD not only trusts the device to give wanted information, but also must understand the information; those w/ mild AD or MCI would have need for more detailed info
• feasibility: need to satisfy various info needs in various settings; interaction abilities constrain the device; should be able to answer most questions of PwAD; time and schedule info; context about event underway (context and location-aware; item location (sensor technology), moveable within the home / outside the home? opinions and feedback, rules that require reasoning, list of steps to complete common items;
• mechanical skills - eliminate chance for errors; most basic of interactions, direct input (touch), speech may be difficult (declining language abilities, difficulty in remembering words), barriers to complexity and anxiety about technology is common among elderly, exacerbated among PwDs; perhaps introduced during earlier stages of memory decline? important that device does not resemble a computer; needs to always be on / ready to use; method of gaining attention required; interface design will be challenging; willingness of caregiver to maintain? effective enough at lessening the questioning behaviour to justify burden of maintaining the information (CG);

• conclusion: info applicant to reduce repetitive questioning likely to fail - must introduce at appropriate phase, keep level of detail appropriate for that phase;

544 Topic: Cognitive Impairments

• Back M, Szymanski MH. The AirBook: force-free interaction with dynamic text in an assistive reading device. In: CHI'01 extended abstracts on Human factors in computing systems. ACM; 2001:251–252. Available at: http://portal.acm.org/citation.cfm?id=634216.

• Dawe M. Desperately seeking simplicity: how young adults with cognitive disabilities and their families adopt assistive technologies. In: Proceedings of the SIGCHI conference on Human Factors in computing systems. ACM; 2006:1152. Available at: http://portal.acm.org/citation.cfm?id=1124772.1124943.

• Lee ML, Dey AK. Lifelogging memory appliance for people with episodic memory impairment. Proceedings of the 10th international conference on Ubiquitous computing - UbiComp '08. 2008:44. Available at: http://portal.acm.org/citation.cfm?doid=1409635.1409643.

• LoPresti EF, Mihailidis A, Kirsch N. Assistive technology for cognitive rehabilitation: State of the art. Neuropsychological Rehabilitation. 2004;14(1):5–39. Available at: http://www.informaworld.com/index/789064993.pdf.

more...

• R.M. Baecker, "Designing technology to aid cognition," Proceedings of the 10th international ACM SIGACCESS conference on Computers and accessibility - Assets '08, vol. 2020, 2008, p. 1.

• M. Balaam, A.M. Hughes, S. Rennick-Egglestone, and T. Nind, "Rehabilitation Centred Design," CHI 2010, Atlanta: 2009, pp. 1-4.

• J. Bauchet, H. Pigot, S. Giroux, D. Lussier-Desrochers, Y. Lachapelle, and M. Mokhtari, "Designing judicious interactions for cognitive assistance," Proceeding of the eleventh international ACM SIGACCESS conference on Computers and accessibility - ASSETS '09, 2009, p. 11.

• S. Carmien, "End user programming and context responsiveness in handheld prompting systems for persons with cognitive disabilities and caregivers," CHI '05 extended abstracts on Human factors in computing systems - CHI '05, 2005, p. 1252.

• M. Pollack, "Autominder: an intelligent cognitive orthotic system for people with memory impairment," Robotics and Autonomous Systems, vol. 44, 2003, pp. 273-282.

• H. Kautz, L. Arnstein, G. Borriello, O. Etzioni, and D. Fox, "An overview of the assisted cognition project," AAAI-2002 Workshop on Automation as Caregiver: The Role of Intelligent Technology in Elder Care, 2002, p. 60–65.

• M. Morris, J. Lundell, and E. Dishman, "Catalyzing social interaction with ubiquitous computing: a needs assessment of elders coping with cognitive decline," CHI'04 extended abstracts on Human factors in computing systems, ACM, 2004, p. 1154.

On designing technology to address users w/ dementia

[Wherton 08] - IDRG 05.18.10

J. Wherton and A. Monk, "Technological opportunities for supporting people with dementia who are living at home," International Journal of Human-Computer Studies, vol. 66, 2008, pp. 571-586.

• abstract
  • grounded theory analysis of interview transcripts with dementia patients + caretakers; discussed ADLs: dressing, taking medication, personal hygiene, preparing food, socialising; design challenges for flexible prompting systems sensitive to intentions, capabilities, values of users;

• intro
  • problem domain - users w/ impaired episodic memory, parallel deficits in exec. function; past efforts to address this area in pervasive computing + sensing have involved computer vision, RFID tags; affecting ADLs
  • exec functioning: sequencing, omissions, action additions, concurrent task performance;
  • impairment to attentional control processes - problems more likely w/ novel procedures; occupational therapy solution: procedural memory stimulation, environmental cues to support action, independence achieved through environmental manipulations, making task-relevant items more visible
  • stress need of naturalistic methods of assessment
  • temporal and spatial orientation: visual cues - flowcharts, calendars; reality orientation therapy - use of large signposts, signs on drawers and cupboards + verbal prompts + cues; extended w/ use of technology - large digital clocks; ENABLE project: activity compass, portable orientation + GPS, verbal prompts to return home;
  • prompting actions: prompt at designated times to support schedule keeping, sensors to guard against accidents;
  • less clear what people w/ dementia and their caregivers want from the technology - what tasks are most important to them? what difficulties are most hard to find non-technological workarounds for? what tasks would their caregivers rather not do for the people they care for?

• study rationale + method
  • 2 small sample interview (opportunistic) - #1 w/ OTs and care workers (20 participants - 11 in focus group, 9 interviews); #2 w/ 18 participants (8 PwDs, 10 caregivers) in home environments; studies not methodologically equivalent, but maximally informative; grounded approach to characterise data - to elicit ideas for technological interventions; concepts grouped to from main categories and component sub-categories (axial coding)

• results - study 1 - fig. 1
  • problems in the home: daily activities (dressing, medication, food/drink, washing, toilet), risks (cooker safety, wandering), interpersonal interaction (communication, recognising people);
  • underlying deficits: sequencing (initiating and ordering actions, problem solving), learning (appliances, surroundings), memory/orientation (forgets events, forgets to do things, orientation to time/place, recognising objects)
  • consequences: PwD: physical wellbeing (safety, security, health), control (personal space, achievement, social isolation), caregiver: patient-caregiver relationship: relationship (interaction, frustration), care demands (constant supervision, anxiety)
• results - study 2 - fig. 2
  • problems in the home: daily activities (dressing, medication, food/drink, washing, toilet), domestic (washing-up, locking-up, ironing, cleaning), leisure (TV, loss of interests) (not mentioned in study 1), interpersonal interaction (conversing, telephone, appointments);
  • situated factors: verbal cues (prompts, notes), visual cues (display items, items visible), familiarity (surroundings, appliances, routines);
  • underlying deficits: sequencing (formulating procedures, becomes motionless), memory/orientation (forgets events, forgets to do things, locating items, knowing the time, finding their way)
  • consequences: PwD: physical wellbeing (safety, security, deterioration), control (personal space, stress, filling time - boredom), caregiver: patient-caregiver relationship: relationship (confrontation, frustration), care demands (tiredness (not among professional caregivers), worry)

• discussion
  • high-level characterisation of results useful for engineers and others new to the area
  • extracted parts of the transcript: suggestions for technological interventions

• discussion - dressing
  • no existing products to help one get dressed
  • provide prompts?
  • how to monitor state of dress of PwD? RFID tags?
  • have items been put on appropriately? appropriate choice of clothes for season / occasion?

• discussion - medication
  • tools exist to automatically dispense medication appropriately
  • has medication been swallowed?
  • call-centre operative to call throughout the day?
  • ripe for future research
  • intelligent toilets for monitoring medication compliance?

• discussion - food and drink
  • cooker safety (i.e. leaving on the stove)
  • cooking tasks too complex to manage alone
  • problems in making a hot drink - for oneself and for visitors (self-esteem issue and sense of personal role + status)
  • safety:
    • cooker: lockable gas shut-off for detecting emergencies in kitchen such as extreme temperature? - rely on automated feedback from PwD / caregiver;
    • can pervasive computing techniques be used to infer how an incident arose and choose between alternatives
    • call centre operator?
    • dangerous for use w/ some PwDs
  • prompting + preparing:
    • no commercially available systems for prompting through process of multi-step task such as making a hot drink; clear objective for research
    • RFID utensils? large screen in kitchen cupboards?
    • how to detect completed actions? how to prompt?
  • measuring quantities (i.e. baking)?