Designing a Cognitive Aid for the Home: A Case-Study Approach
Jessica Paradise, Elizabeth D. Mynatt, Cliff Williams, John Goldthwaite · 2004 · Proceedings of the 6th International ACM SIGACCESS Conference on Computers and Accessibility (Assets 04) · doi:10.1145/1028630.1028656
Summary
This paper describes the design of a computational pacing aid for a woman ("C") with mild traumatic brain injury (TBI) who struggles with chronic tardiness during her morning routine at a residential program. The researchers from Georgia Tech and Drexel University used an in-depth case study methodology involving interviews with C and her caregiving staff, paper-based activity mapping exercises, and direct observation of her morning routine. C's tardiness stems from cognitive impairments including difficulty with time estimation, planning and executive control, susceptibility to distraction, and irritability. She relies on external cues like observing other residents' activities to gauge her progress, but these cues are unreliable. The fieldwork produced two complementary sets of design dimensions: user-centered constraints (perception of control, appearance, burden on user, non-distracting, ease of use, design simplicity, burden on staff) and system-centered capabilities (interaction/user prompting, system input, system intelligence, display of history, system mobility, external cues). The resulting prototype is a touch-panel display mounted on the wall that shows C's morning activities as buttons she can touch to confirm completion. The background color gradually shifts from blue to red as she falls behind schedule, providing ambient awareness of her pacing without intrusive alarms. Two display alternatives were developed: a time-based layout showing activities chronologically and a location-based layout organizing activities by the room where they occur.
Key findings
Paper prototype evaluation with C and two staff members revealed overall positive reactions to the system concept, with everyone agreeing the time-based display had the strongest potential to help. The location-based display received less enthusiasm. Staff expressed a desire to build the system and have C try it. A key design tension emerged around perception of control: the researchers chose to have C manually confirm activities rather than using automatic recognition, because preserving C's sense of control was more important than reducing interaction burden. Staff were divided on whether C would actually feel in control, with one noting that C currently feels she lacks control over most aspects of her life. The ambient color-change prompting mechanism (background shifting from blue to red) received enthusiastic response from staff but also raised concerns about whether it might cause C anxiety. Staff identified a critical missing feature: the ability to modify C's target departure time, since last-minute schedule changes are common. The research methodology itself yielded important insights — pairing interviews with C alongside staff interviews was essential because C's recollections were sometimes jumbled, and staff provided contextual details that reframed problems C described.
Relevance
This research offers a valuable model for accessibility practitioners designing cognitive support technologies, particularly for home environments where the user's attention is on external tasks rather than the technology itself. The dual framework of user-centered constraints and system-centered capabilities provides a reusable structure for balancing what the technology can do against what the user needs and can tolerate. The emphasis on perception of control is especially relevant — for people with cognitive impairments who have already lost significant autonomy, assistive technology must be designed to restore rather than further diminish their agency. The ambient display approach (color shifts rather than alarms) illustrates how non-distracting feedback can support cognition without adding to cognitive load. A key limitation is the single-participant case study design, which limits generalizability, though the authors argue this depth was necessary to uncover the nuanced design requirements that broader studies might miss. The work is also limited to the prototype evaluation stage, with no deployment data on whether the system actually improved C's morning routine timing.
Tags: cognitive accessibility · traumatic brain injury · cognitive aid · smart home · independent living · participatory design · case study · pacing · ambient display