CMIC Seminar: Machine learning and neuroimaging in psychiatry

Speaker: Janaina Mourao-Miranda
UCL Contact: Dominique Drai (Visitors from outside UCL please email in advance).
Date/Time: 10 May 17, 13:00 - 14:00
Venue: Roberts 106

Abstract

Machine learning techniques have been successfully applied to clinical neuroimaging data leading to a growing body of research focused on diagnosis and prognosis of mental health disorders. However, so far, most of these studies have focused on supervised classification problems, i.e. they summarize the clinical assessment into a single measure (e.g. diagnostic classification) and the output of the models is limited to a probability value and, in most cases, a binary decision (e.g. healthy/patient). Considering that current diagnostic categories in psychiatry fail to align with findings from clinical neuroscience and genetics, this framework fails to capture the underlying biology and fully characterize disease variation. Alternative frameworks, such as unsupervised learning, are therefore needed to study brain diseases whose underlying processes are not yet fully understood and, therefore, might have an unreliable categorical classification. In this talk I will review the machine learning framework commonly applied to neuroimaging in psychiatry and discuss potential alternatives to overcome limitations of this framework.