CIRCADIAN COMPUTING: SENSING AND STABILIZING BIOLOGICAL RHYTHMS

Abstract

This dissertation lays the groundwork for Circadian Computing with a novel and broad vision of technologies that support and adapt to our innate biological rhythms. Similar to most terrestrial organisms, human physiology and behavior are shaped by a 24-hour periodicity known as circadian rhythm. Indeed, almost every neurobehavioral process including our sleep, metabolism, cognitive performance, and mood reflects circadian rhythms. These rhythms ensure synchronization across different processes and as such, are crucial for our health and well-being. Persistent circadian disruption increases risk for cancer, obesity, and cardiovascular diseases. It has been associated with occupational accidents and serious loss of productivity in the workplace as well. Recent findings have also started identifying links between circadian disruption and mental illnesses including bipolar disorder and schizophrenia. However, in our modern world, circadian disruption is becoming increasingly widespread. The invention of artificial light fundamentally changed our ancestral sleep and wakeup patterns. Since then, we have gradually moved towards a 24-hour society. The recent development in entertainment and communication technologies has also resulted in an “always-on” ethos. The resulting trend is worrisome. Sleep pathologies are reaching an epidemic level with 70\% of the population suffering from significant circadian disruptions. As a result, recently there has been an increased focus on monitoring and identifying disruptions in circadian rhythms. However, these methods and findings are often limited to controlled lab environments. As a result, they are not adequate for granular monitoring of circadian disruptions in the wild over a longitudinal period of time. As such, there is a need for novel pervasive technologies for tracking, monitoring, and modeling circadian disruptions and its impact in the real world. There is also an opportunity for developing intervention tools for maintaining circadian stability. This dissertation is a leading step towards the broad vision of circadian-aware technologies for sensing, adapting to, and stabilizing our innate biological rhythms. In my PhD work, I have shown the feasibility of bringing a circadian-aware perspective across different application domains. Specifically, I have developed and evaluated methods for unobtrusively assessing circadian disruptions. I have also showed that behavioral and contextual data can be used for modeling and predicting alertness — a circadian process integral to our cognitive performance. I have also developed, deployed, and evaluated a data driven tool focusing on identifying circadian anomalies in patients with bipolar disorder.

With this groundwork in place, I believe that there is an exciting opportunity lying ahead for Circadian Computing. In particular, circadian-aware technologies can potentially reshape a number of application domains including education and learning, optimized scheduling, mental health care, and chronotherapy. I hope this dissertation motivates a circadian perspective in future technology development and contributes to the shared effort of improving our productivity, health, and well-being.