PRESCIENT / AMP SCZ
The PRESCIENT project, part of the Accelerating Medicines Partnership Schizophrenia (AMP SCZ) program, focuses on predicting clinical outcomes in individuals at clinical high risk (CHR) for psychosis.
Research
Bridging clinic and computation
Focus
Understanding mental health requires more than looking at symptoms in isolation. It means tracing the dynamic interplay between biology, psychology, the social environment and the digital world. Our lab combines clinical expertise with computational innovation to investigate how environmental exposures, individual vulnerabilities and social interactions shape trajectories of mental health.
We aim to:
- identify risk and protective factors across biology, behaviour and environment (including the exposome)
- develop AI-driven tools for early detection and personalised prevention
- translate digital and clinical innovations into effective interventions in psychiatry
Alongside our research, we run an outpatient service for early recognition and intervention, the FETZ, so that our insights are grounded in clinical practice and benefit patients directly.
Methods
Our team brings together methodologically oriented clinicians and clinically oriented methodologists. This dual perspective lets us move from theory to practice. We apply and advance methods such as:
- Artificial intelligence and machine learning to extract meaningful patterns from complex behavioural, clinical and biological data
- Generative and network models to simulate interactions between symptoms, risk factors and environments
- Digital markers and computational psychiatry approaches to personalise diagnosis and monitoring
- Exposome research to capture the cumulative impact of environmental influences
- Meta-analysis and evidence synthesis to address critical clinical questions with the broadest possible evidence base
Aims
By integrating AI-driven analytics, digital psychiatry and environmental perspectives, we aim to build models that predict individual risk and treatment response with greater accuracy. Ultimately, our goal is to personalise behavioural, pharmacological and psychotherapeutic interventions in psychiatry, bridging clinic and computation to deliver better mental health outcomes in today’s human–digital world.
Research groups
AG Dynamics of Risk and Resilience
Led by Dr. Jessica Hartmann
Mental health is not a static state but a dynamic process. The group studies how mental illness develops across early stages and transdiagnostic trajectories, combining clinical longitudinal research with digital methods such as ecological momentary assessment and smartphone sensing.
Prediction & precision psychiatry
Machine-learning models that forecast individual outcomes in people at clinical high risk.
CARE Network
The CARE study, part of the CARE Network (Computer-assisted Risk Evaluation), is an innovative clinical research project aimed at early detection and prevention of psychosis.
Individualised prediction in clinical high risk
A focus of our lab is the identification of individuals that are at-risk to develop a psychiatric disorder.
Personalized cognitive training
Cognitive symptoms are a central feature of most psychiatric diseases including psychotic and affective disorders. Recent results, including work from our lab (Kambeitz-Ilankovic et al., 2019), demonstrated that cognitive training can effectively improve cognitive performance in patients with psychosis.
Language & digital markers
Speech, language and smartphone data as scalable markers of mental health.
LAMBDA study
The LAMBDA study (Language Markers and Brain Dysfunction in Early Psychosis) investigates the link between language and mental health, specifically focusing on psychosis.
PhenoNetz: transdiagnostic phenotyping with experience sampling
Through our involvement in the Early Recognition and Intervention Service for Mental Disorders (FETZ) of the University Hospital Cologne, we noticed that the currently established prevention approach in psychiatry, which targets diagnosis-specific syndromes, does not cover the majority of…
Computational modeling & generative agents
Brain-network simulations, symptom networks, digital twins and LLM-based agents.
Computational modeling of the functional connectome
This project aims at demonstrating the applicability of biophysical brain network models to neuropsychiatry. We will investigate the mechanisms underlying aberrant whole-brain functional connectivity observed in patients with SCZ using a computational neuro-platform named "The Virtual Brain" (TVB).
Symptom networks in psychotic and affective disorders
Psychiatric disorders are mainly defined by the occurrence of clinical symptoms such paranoia, social isolation or anxiety.
Digital Twins and Generative Agents in Mental Health
We develop computational "digital twins" of patients and generative AI agents to simulate mental health trajectories.
Clinical interventions
Trials and cohorts that turn insights into better treatment for young people.
PsyLetics
PsyLetics is an innovative research project, investigating the effects of high-intensity training (HIT) in patients with schizophrenia.
Longitudinal course of cannabis-induced psychosis (CIP)
Cannabis has been identified as one of the major risk factors for developing psychosis. Recent studies show that up to 50 % of all patients with cannabis-induced psychosis (CIP) continue to develop a permanent form of a psychotic disorder.
Looking for software? See our code & tools.