Skip to main content

Please note: this event has passed

Title: From Data to Clinical Tools: Leveraging Machine Learning for Advancing Psychiatric Research and Care

Abstract: Machine learning presents significant opportunities and challenges in the field of mental healthcare, providing avenues for advancements in early identification and personalized preventive treatment of affective and psychotic disorders. My talk will discuss the potential of machine learning algorithms trained on diverse datasets, incorporating clinical, neuroimaging, and genetic information, to detect subtle patterns predictive of mental health disorders. Furthermore, I will exemplify how ML techniques can be used to better understand the bio-psycho-social heterogeneity of affective and psychotic disorders. Finally, I will review how validation could be used to assess the generalization capacity of these models across diverse populations, paving the way for their clinical implementation as decision-support tools for clinicians. The talk will conclude with a discussion of ethical considerations such as privacy, data security, and algorithm transparency, as well as a focus on future clinical research needed to evaluate long-term outcomes and effectiveness of machine learning in existing healthcare systems.

Biography: Professor Nikolaos Koutsouleris serves as Chair of Precision Psychiatry at the IoPPN, King’s College London and Ludwig-Maximilian University Munich (LMU) and Fellow of the Max-Planck-Society at the Max Planck Institute for Psychiatry Munich. He coordinated the EU-FP7 funded project PRONIA (“Personalised Prognostic Tools for Early Psychosis Management”; and heads the Centre for Transitional Youth Mental Health and the Section for Precision Psychiatry at the Department of Psychiatry, LMU. Prof Koutsouleris studied medicine at LMU between 1996 and 2003 as scholar of the German National Academic Foundation. He took his first medical & academic appointment in 2004 at the Department of Psychiatry and Psychotherapy, where he finished his doctorate thesis in 2005. Since 2008, Prof Koutsouleris has advanced the use of multivariate pattern recognition methods for the identification and validation of diagnostic and prognostic prediction models in at-risk and early stages of affective and non-affective psychoses. His work was awarded with several national and international prizes and led so far to 168 peer-reviewed, highly cited papers (h-index: 55). In addition, he strived to make robust machine-learning methods available to researchers in the clinical neurosciences in order to improve the methodological rigour of their use based on the proper use of validation and model sharing approaches. These efforts have the lead to the publication of the open-source NeuroMiner machine learning platform available at

Click here to join the meeting