Search

Tasks are increasingly becoming cognitively based instead of physically based, and managing multiple tasks at once is becoming commonplace. Tools that help people to manage their cognitive activity in their lives could therefore be highly valuable. Unlike physical activity trackers, however, it is not yet understood how cognitive activity could be tracked in daily life in order to provide people with meaningful reflections and useful goals, known as personal informatics. Mental workload is a promising concept in this respect, due to its performance defining qualities and evergrowing relevance. Thus, this thesis investigates mental workload tracking in everyday life. Mental workload has typically been investigated from an isolated taskbased, 'work'load perspective, predominantly in safety-critical environments, meaning that our understanding of how mental workload functions in daily life is limited. By adopting a novel longitudinal, holistic, and personcentred perspective, the research presented in this thesis aimed to improve understanding of 1) physiological mental workload measurements in realworld environments, 2) how mental workload could be useful as a form of personal informatics and mental workload as a concept itself, 3) how mental workload data can be meaningfully communicated to users, and 4) ethical considerations for tracking devices. Two empirical studies were conducted in relation to this. Firstly, a naturalistic laboratory study used brain imaging methods to physiologically measure mental workload levels for general work tasks. Officeworker participants completed personalised reading and writing tasks at different levels of difficulty. Verbal interruptions were incorporated and coffee drinking was largely unrestricted. Results found that the measure was sensitive to reading tasks but not writing tasks, which helped to identify challenges for real-world mental workload tracking in terms of maximising the sensitivity of the measures. Interruptions were also found to affect mental workload levels, and these findings were interpreted using mental workload models. Study 2 had a quantitative and qualitative phase. It explored mental workload as as a concept and as a form of personal informatics. The quantitative phase involved participants subjectively tracking their mental workload levels at regular intervals; questionnaires related to wellbeing were also completed each evening. The qualitative phase interviewed the same participants in depth about their experiences and perceptions of mental workload and mental workload tracking devices. From this research, an apparent Mental Workload Cycle was developed, where participants aimed to fluctuate between low, medium, and high mental workload levels. This was because each level serves a purpose but sustaining any level for too long results in negative consequences. Factors were identified that could disrupt the Cycle, and indeed our quantitative data indicated that actual behaviour often did not align with qualitative preferences. Qualitative insights also investigated the design of mental workload tracking technology in terms of design and ethical considerations. Design considerations related to metaphors, colours, shapes, and descriptors. Ethical concerns related to data privacy, validity, misinterpretation, and personal identity. The important questions for cognitive activity tracking and understanding mental workload in everyday life are human-computer interaction ones. These relate to, for example, what useful data consumer neurotechnology could be used to track, what goals we could set for healthy lives, and how personal cognitive informatics will relate to the pervasive way we use physical activity tracking. Towards understanding this future, this thesis makes four contributions. First, we identify challenges for physiologically measuring mental workload in uncontrolled environments. Second, we develop the Mental Workload Cycle, a model that progresses understanding of 1) how mental workload can be used for personal informatics, and 2) mental workload as a concept in terms of the factors that contribute of the states of overload and underload. Third, we produced design recommendations for communicating mental workload data. And fourth, we explicated ethical concerns for future consumer neurotechnology. These findings should be used to progress personal informatics and human factors research, and implicate the direction of consumer neurotechnology as it develops towards longitudinal tracking of cognitive activity.