AG Precision Mental Health, Pavol Mikolas, PhD

Pavol Mikolas

Medizinische Leitung und Ausrichtung Forschungsaktivitäten

(Englisch: Medical lead and alignment of research activities)

Janek Haschke

Organisatorische Leitung und Koordination sowie Aufbau und Betrieb KI Infrastruktur

(Englisch: Lead AI infrastructure and operations)

Wir integrieren multimodale Datensätze und KI-gestützte Methoden, um (a) die Identifikation von Personen mit Risiko für eine psychische Störung, (b) Vorhersage ihres klinischen Verlaufs und (c) Unterstützung bei Therapieentscheidungen zu verbessern. Unser Schwerpunkt liegt insbesondere auf dem Risiko für affektive und psychotische Erkrankungen. Wir wenden maschinelles Lernen auf groß angelegten, multizentrischen Datensätzen wie der Early-BipoLife-Studie an. Darüber hinaus nutzen wir große Sprachmodelle zur Analyse medizinischer Dokumentationen und Sprachdaten, um psychiatrische Symptome, Notfälle und die Früherkennung von psychischen Erkrankungen zu verbessern. Unser Ziel ist es nicht nur, neue prädiktive Modelle zu entwickeln, sondern diese auch in die klinische Praxis zu überführen – beispielsweise in Form eines Clinical Decision Support Systems (CDS).

• Validierung und Optimierung von Risikoklassifikatoren für Psychose und Manie bei Jugendlichen und jungen Erwachsenen unter Verwendung hochheterogener, multinationaler Datensätze und KI-gestützter Methoden.
• KI-gestützte Methoden zur Erfassung und Analyse psychiatrischer Symptome aus medizinischer Dokumentation und Sprache.
• KI-gestützte Identifikation von Lithium-Respondern: Automatisierte Analyse der Alda-Skala aus elektronischen Krankenakten.
• Verbesserung der Vorhersage von Behandlungsergebnissen bei therapieresistenter Depression mittels Methoden der Künstlichen Intelligenz

Wiest IC*, Verhees FG*, Ferber D, Zhu J, Bauer M, Lewitzka U, … Mikolas P.*, Kather JN* et al. Detection of suicidality from medical text using privacy-preserving large language models. The British Journal of Psychiatry 2024; 225: 532–7.

Mikolas P, Marxen M, Riedel P, Bröckel K, Martini J, Huth F, et al. Prediction of estimated risk for bipolar disorder using machine learning and structural MRI features. Psychol Med 2023; : 1–11.

Huth F, Tozzi L, Marxen M, Riedel P, Bröckel K, Martini J,… Mikolas P et al. Machine Learning Prediction of Estimated Risk for Bipolar Disorders Using Hippocampal Subfield and Amygdala Nuclei Volumes. Brain Sciences 2023; 13: 870.

Mikolas P, Vahid A, Bernardoni F, Süß M, Martini J, Beste C, et al. Training a machine learning classifier to identify ADHD based on real-world clinical data from medical records. Sci Rep 2022; 12: 12934.

Mikolas P, Bröckel K, Vogelbacher C, Müller DK, Marxen M, Berndt C, et al. Individuals at increased risk for development of bipolar disorder display structural alterations similar to people with manifest disease. Transl Psychiatry 2021; 11: 485.

Mikolas P*, Tozzi L*, Doolin K, Farrell C, O’Keane V, Frodl T. Effects of early life adversity and FKBP5 genotype on hippocampal subfields volume in major depression. Journal of Affective Disorders 2019; 252: 152–9.

Mikolas P, Melicher T, Skoch A, Matejka M, Slovakova A, Bakstein E, et al. Connectivity of the anterior insula differentiates participants with first-episode schizophrenia spectrum disorders from controls: a machine-learning study. Psychol Med 2016; 46: 2695–704.

*shared first/last authors

EKFZ for Digital Health (https://digitalhealth.tu-dresden.de/)

ScaDS.AI Dresden/Leipzig (https://scads.ai)

European College of Neuropsychopharmacology (ECNP) Bipolar Disorders Network (https://www.ecnp.eu/research/ecnp-networks/list-of-ecnp-networks/bipolar-disorders-network)

Institute of Psychiatry, Psychology & Neuroscience (IoPPN), King’s College London, London, UK (https://transcampus.eu)

Max-Planck-Institut für Psychiatrie, München (https://www.psych.mpg.de/)

About

We integrate multimodal datasets and AI-based methods to improve (a) the identification of individuals at risk for mental disorders, (b) the prediction of their clinical course, and (c) support for treatment decisions. Our focus is particularly on the risk of affective and psychotic disorders. We apply machine learning to large-scale, multicentre datasets such as the Early BipoLife study. In addition, we use large language models to analyse medical documentation and speech data to improve psychiatric symptoms, emergencies and the early detection of mental illness. Our goal is not only to develop new predictive models, but also to translate them into clinical practice – for example, in the form of a Clinical Decision Support System (CDS).

Projects

• Validation and optimisation of risk classifiers for psychosis and mania in adolescents and young adults using highly heterogeneous, multinational datasets and AI-based methods.
• AI-supported methods for recording and analysing psychiatric symptoms from medical documentation and language.
• AI-supported identification of lithium responders: automated analysis of the Alda scale from electronic medical records.
• Improvement of the prediction of treatment outcomes for treatment-resistant depression using artificial intelligence methods.

Publications (selection):

Wiest IC*, Verhees FG*, Ferber D, Zhu J, Bauer M, Lewitzka U, … Mikolas P.*, Kather JN* et al. Detection of suicidality from medical text using privacy-preserving large language models. The British Journal of Psychiatry 2024; 225: 532–7.

Mikolas P, Marxen M, Riedel P, Bröckel K, Martini J, Huth F, et al. Prediction of estimated risk for bipolar disorder using machine learning and structural MRI features. Psychol Med 2023; : 1–11.

Huth F, Tozzi L, Marxen M, Riedel P, Bröckel K, Martini J,… Mikolas P et al. Machine Learning Prediction of Estimated Risk for Bipolar Disorders Using Hippocampal Subfield and Amygdala Nuclei Volumes. Brain Sciences 2023; 13: 870.

Mikolas P, Vahid A, Bernardoni F, Süß M, Martini J, Beste C, et al. Training a machine learning classifier to identify ADHD based on real-world clinical data from medical records. Sci Rep 2022; 12: 12934.

Mikolas P, Bröckel K, Vogelbacher C, Müller DK, Marxen M, Berndt C, et al. Individuals at increased risk for development of bipolar disorder display structural alterations similar to people with manifest disease. Transl Psychiatry 2021; 11: 485.

Mikolas P*, Tozzi L*, Doolin K, Farrell C, O’Keane V, Frodl T. Effects of early life adversity and FKBP5 genotype on hippocampal subfields volume in major depression. Journal of Affective Disorders 2019; 252: 152–9.

Mikolas P, Melicher T, Skoch A, Matejka M, Slovakova A, Bakstein E, et al. Connectivity of the anterior insula differentiates participants with first-episode schizophrenia spectrum disorders from controls: a machine-learning study. Psychol Med 2016; 46: 2695–704.

*shared first/last authors

Cooperations (selection):

EKFZ for Digital Health (https://digitalhealth.tu-dresden.de/)

ScaDS.AI Dresden/Leipzig (https://scads.ai)

European College of Neuropsychopharmacology (ECNP) Bipolar Disorders Network (https://www.ecnp.eu/research/ecnp-networks/list-of-ecnp-networks/bipolar-disorders-network)

Institute of Psychiatry, Psychology & Neuroscience (IoPPN), King’s College London, London, UK (https://transcampus.eu)

Max-Planck-Institut für Psychiatrie, München (https://www.psych.mpg.de/)