Sabbatical Eva Ceulemans: Online detection of depressive episodes through statistical proces control

Depression has a major impact on all aspects of life and must therefore be done early
may be detected to prevent worse. Although more and more researchers use
mobile phones collect information about fluctuations in emotions and sleep, and demonstrate this
these changes are predictive of depression, there are currently no good ones
statistical methods are available to monitor this information online and prospectively
prevention of depression. I currently have a new line of research in which we
want to develop such methods. We will build on the family of
statistical process control (SPC) techniques. SPC is very commonly used in the industry to
monitoring production processes. To be able to use SPC to impending
However, to detect depressive episodes as early as possible, we need some
problem solving: First, it is unclear what combinations of emotional
dynamics (eg mean, variability, (auto) correlation) optimally differentiate between
depressed and healthy individuals. Second, current SPC techniques are incapable of
dynamics that cover different domains and that on different time scales
measured. Third, it is not known how we identify individuals without a baseline
can start monitoring data immediately with SPC. During this Sabbath, I would like to
read further in these topics and try out ideas to bring this line of research up to cruising speed.