Integrating biology and behaviour for precision stratification of mental disorders

In many areas of medicine -for example oncology- quantitative markers of
biology have transformed diagnosis and treatment allocation, allowing for
targeted and personalized interventions in the spirit of *precision medicine*
ultimately leading to better outcomes for patients and a better understanding
of disease entities. This is not the case in psychiatry, where disorders are
still diagnosed on the basis of symptoms and biomarkers to assist treatment
allocation remain to be developed. Finding markers to predict and stratify
mental disorders is a formidable goal that requires understanding their impact
across multiple levels from neurobiology to behaviour. Advances in neuroimaging
provide unprecedented ability to measure neurobiology non-invasively across
increasingly large cohorts. Simultaneously, the near ubiquity of smartphone
technology in modern life makes it possible to measure behaviour
quantitatively, unobtrusively and in real-world situations using digital
phenotyping. However, the potential of these techniques for improving clinical
decision-making is not being realized due to a lack of analysis tools to chart
variation across individuals, to integrate complimentary information from
different data modalities and to predict trajectories of functioning at the
level of the individual.

The primary objectives of this study are to: (i) derive quantitative measures
of behaviour on the basis of smartphone monitoring in a healthy cohort and a
cross-diagnostic cohort of patients with mental disorders (e.g. depression,
anxiety and substance use disorders); (ii) establish the validity of these as
measures of mental state by performing quantitative comparisons of individuals
with mood disorders against healthy controls (i.e. in a case-control sense and
via continuous associations with symptom levels) and using these measures to
predict subjective mood scores acquired on a moment-to-moment basis using
ecological momentary assessment; (iii) combine these measures with multi-modal
neuroimaging data to predict measures of mental health across the spectrum of
functioning. For this we will use normative models (*brain growth charts*) to
bind different samples into a common reference space. Throughout this project
behaviour will be conceptualised and quantified in terms of functional domains
(e.g. negative valence, arousal/inhibition) approximately corresponding to
those outlined in the Research Domain Criteria framework.

Observational cross-sectional study

Burden: All subjects will visit the study centre for an intake visit, where
they will install the necessary software on their smartphone and be instructed
on the use of the ecological momentary assessments (approximately 1 hour).
Participants in the clinical group will also visit the study centre a second
time, during which a standardized neuroimaging protocol will be acquired (total
1.5 hours). The use of patients in this study is necessary because the goal is
to develop quantitative and ecological measures of behaviour that can be used
to predict functioning in mental disorders.
Risks: Minimal risk
Benefits: While there will be no direct benefit to the participants themselves,
the study may contribute knowledge that may be helpful in the future for
patients with the same mental disorders