Connecting people through physiosocial technology

Social connectedness is one of the most important predictors of health and well-being. The goal of this dissertation is to investigate technologies that can support social connectedness. Such technologies can build upon the notion that disclosing emotional information has a strong positive influence on social connectedness. As physiological signals are strongly related to emotions, they might provide a solid base for emotion communication technologies. Moreover, physiological signals are largely lacking in unmediated communication, have been used successfully by machines to recognize emotions, and can be measured relatively unobtrusively with wearable sensors. Therefore, this doctoral dissertation examines the following research question: How can we use physiological signals in affective technology to improve social connectedness? First, a series of experiments was conducted to investigate if computer interpretations of physiological signals can be used to automatically communicate emotions and improve social connectedness (Chapters 2 and 3). The results of these experiments showed that computers can be more accurate at recognizing emotions than humans are. Physiological signals turned out to be the most effective information source for machine emotion recognition. One advantage of machine based emotion recognition for communication technology may be the increase in the rate at which emotions can be communicated. As expected, experiments showed that increases in the number of communicated emotions increased feelings of closeness between interacting people. Nonetheless, these effects on feelings of closeness are limited if users attribute the cause of the increases in communicated emotions to the technology and not to their interaction partner. Therefore, I discuss several possibilities to incorporate emotion recognition technologies in applications in such a way that users attribute the communication to their interaction partner. Instead of using machines to interpret physiological signals, the signals can also be represented to a user directly. This way, the interpretation of the signal is left to be done by the user. To explore this, I conducted several studies that employed heartbeat representations as a direct physiological communication signal. These studies showed that people can interpret such signals in terms of emotions (Chapter 4) and that perceiving someone's heartbeat increases feelings of closeness between the perceiver and sender of the signal (Chapter 5). Finally, we used a field study (Chapter 6) to investigate the potential of heartbeat communication mechanisms in practice. This again confirmed that heartbeat can provide an intimate connection to another person, showing the potential for communicating physiological signals directly to improve connectedness. The last part of the dissertation builds upon the notion that empathy has positive influences on social connectedness. Therefore, I developed a framework for empathic computing that employed automated empathy measurement based on physiological signals (Chapter 7). This framework was applied in a system that can train empathy (Chapter 8). The results showed that providing users frequent feedback about their physiological synchronization with others can help them to improve empathy as measured through self-report and physiological synchronization. In turn, this improves understanding of the other and helps people to signal validation and caring, which are types of communication that improve social connectedness. Taking the results presented in this dissertation together, I argue that physiological signals form a promising modality to apply in communication technology (Chapter 9). This dissertation provides a basis for future communication applications that aim to improve social connectedness