John Axelsson

There is an urgent need for better tools to improve and facilitate diagnosis of bipolar disorder and to detect and prevent affective episodes at the earliest possible stage. We hypothesize that the longitudinal interaction between three psychobiological factors (sleep patterns, motor activity, and decision-making abilities) contains information to distinguish patients with bipolar disorder and to detect upcoming affective episodes. In this project, we will first optimize an existing decision-making battery for longitudinal testing. Thereafter, we will use a smartphone app, a smartwatch, and artificial intelligence algorithms to characterize the interaction between the psychobiological factors at the individual level for 12 weeks in healthy volunteers, patients with bipolar disorder, and patients with recurrent depressive episodes. We will demonstrate that bipolar patients can be distinguished based on the longitudinal interaction between the psychobiological factors. Finally, we will follow up the patients with bipolar disorder for 1 year and use the artificial intelligence algorithm trained on each patient´s data to predict upcoming changes in affective states at the individual level. This project will allow us to develop an app that can be used in the clinic to improve diagnosis and to predict affective episodes in bipolar disorder. As such, this project will move the field of psychiatry towards precision medicine and improve the care of patients with bipolar disorder.

Insufficient sleep has negative effects on several cognitive abilities, but little is known about the extent to which sleep-deprived people are aware of their own impairments. Such awareness is called metacognition, which includes predictions and evaluations of one's own performance, as well as the resulting behavior. This is especially important in situations where sleep loss is common and mistakes are costly, such as for shift working physicians or in transportation. The aims of this project are therefore to determine the effects of sleep loss on metacognitive ability, its consequences for compensatory behavior, and mechanisms underlying these effects. Through two large experimental studies (n=182 and n=300), using different manipulations of sleep (total sleep deprivation or 4h-sleep restriction) and a variety of metacognitive measures (assessing confidence, accuracy, and regulation), we will elucidate how sleep loss affects key aspects of metacognitive ability. We will also establish involved mechanisms (prior beliefs and processing effort) for such effects, and characterize predictors of individual differences. Sleep loss is common in today's society, and people are often expected to perform optimally despite insufficient sleep, e.g. during night-shifts or on-call work. As metacognitive ability when sleep-deprived may protect against some detrimental outcomes of sleep loss, the findings of this project are important not only for sleep research, but for society at large.