Design of Cognitive Interfaces for Personal Informatics Feedback

Abstract

The emergence of embedded low-cost sensors in mobile devices allows us to capture unprecedented data about human behavior. Hence personal informatics systems are becoming an integrated part of our everyday life: Capturing various aspects from our health, work-life, to economic balance, and utility consumption. All of which are aimed to provide knowledge of oneself, on which we can reflect. Many personal informatics systems are characterized by mainly focusing on collecting and analyzing data, rather than translating the data into meaningful feedback. This dissertation presents challenges related to personal informatics systems, and propose an approach to design cognitive interfaces, which considers both users’ motivations, needs, and goals. In this thesis I propose a new personal informatics framework, the feedback loop, which incorporates lean agile design principles. Including hierarchical modeling of goals, activities, and tasks to create minimal viable products. While considering how micro-interactions based on an understanding of data, couples with user needs and the context they appear in, can contribute to creating cognitive interfaces. Designing cognitive interfaces requires a focus on translating data into meaningful feedback, which the users can reflect on in order to gain insights. Thus I present tools such as personalized baselines and thresholds to enable reflection, while creating personalized goals, scenarios, trade-offs in order to provide actionable feedback, which can help users to adjust their behavior. Although feedback can be provided in many different ways, it basically consists of audio, visual, and haptic components, which combined may reinforce each other to support the underlying interaction. The papers included in this thesis cover selected parts of the feedback loop. For instance, examining emotional responses to pleasant and unpleasant media content from brain activity, reveals the large amount of data and extensive analysis required to apply this to future personal informatics systems. In addition we analyse challenges related to temporal aspects of the feedback loop, when users attempt to self-regulate their brain activity based on a real-time feedback. This leads to identification of underlying audio, visual and haptic feedback components, which combined may support the underlying interaction within personal informatics. And with the emerging availability of sensor packed wearable devices, haptic feedback may become an inherent part of personal informatics systems, which could enhance the interaction based visual feedback.