Evidence-based diagnostics of mental illness. Dr. Sebastian Olbrich from University Hospital Zurich is working on a deep-learning platform for psychiatrists and psychotherapists.
You are an experienced psychiatrist in your late 30s. Have you been doing everything wrong up to now, Sebastian?
I wouldn’t say that, but something has always made me think. In psychiatry, the type of therapy usually depends on the personal and professional background of the treating physician – and not on what can be learned about the individual patient.
How do you want to change that?
There is increasing scientific evidence that certain diseases correlate with typical brain activity. If this is the case, it should be possible to draw conclusions from the encephalogram (EEG) about specific manifestations of a disease, and from there about the best form of therapy. What has been missing so far is a systematic evaluation of patient EEGs and comparison with the success of the chosen treatment. We want to close this gap using our deep- learning algorithms. It would be a step on the way to biomarker-based psychiatric diagnostics.
What stage are you currently at?
The algorithms are available, and the functionalities of the platform have been defined. Starting in the fall, clinics, doctors, and patients will have the possibility to upload EEGs. But we’ll limit ourselves to the manifestations of depression for now.
Why depression specifically?
Because this is the most common of all mental illnesses. Here in Western Europe, about 10 percent of the population suffers from some form of depressive disorder. The economic cost alone of lost working hours is enormous.