Adapting Human-Computer Interfaces to Working Memory Limitations Using MATCHS
Résumé
We introduce MATCHS (Memory Adaptation Through Cognitive Handling Simulation), a new feature for
human-computer interaction (HCI) that enables tasks to be more efficiently performed when they heavily depend on a user’s
working memory (WM) capacity. WM is the part of human cognition responsible for short-term information storage; it is
key to the proper completion of one’s current task(s). Known to have very limited capacity and a fast decaying time, in
particular with age or when Alzheimer-like diseases set in, WM is one of the strongest factors that explain individual differences in cognitive abilities, thus motivating the introduction of WMspecific parameters in personalized user interface design. Our MATCHS framework for HCI design builds upon a lowlevel model of the user’s WM to predict how much, and for
how long, an information can be expected to be stored there. Thus, the task at hand can, accordingly, be dynamically adjusted. We implemented our approach in a new memory game named Match2s, and tested our framework, including with a small set of
users. Our results suggest that MATCHS is able to properly assess someone’s current WM characteristics and track its evolution
as users’ motivation or focus evolve over time, providing useful information to adapt simple interaction scenarii such as the one in
Match2s. Future work will include more testing and applications, in particular with users for whom such interfaces could be a
major asset, i.e., patients suffering from dementia.