International Max Planck Research School on Computational Methods in Psychiatry and Ageing Research (COMP2PSYCH)

 

The International Max Planck Research School (IMPRS) COMP2PSYCH serves as the graduate program of the Max Planck UCL Centre for Computational Psychiatry and Ageing Research. In London, the students are based at University College London (UCL); in Berlin, they are based at the MPI for Human Development and Humboldt-Universität zu Berlin.

Co-Speakers: Ulman Lindenberger, MPI for Human Development; Raymond J. Dolan, UCL
Coordinators: Silke Schäfer, MPI for Human Development; Dominique Drai, UCL

COMP2PSYCH teaches and trains concepts and methods from computer science and statistics in relation to substantive research questions in psychiatry and lifespan psychology. It was established by the Max Planck UCL Centre for Computational Psychiatry and Ageing Research in early 2016, was positively evaluated in 2020, and extended until September 2028. The IMPRS COMP2PSYCH is an IMPRS Type B according to the Max Planck Society's categorization, i.e., it is a smaller, project-based, time-limited undertaking for innovative research topics that will end after the (current) second funding period.


Table 1. COMP2PSYCH Faculty
Andreas M. BrandmaierMPI for Human Development
Martin J. DahlMPI for Human Development
Christian DoellerMPI for Human Cognitive and Brain Sciences
Ray DolanUniversity College London
Stephen FlemingUniversity College London
Tobias HauserUniversity College London
Ralph HertwigMPI for Human Development
Quentin HuysUniversity College London
Simone KühnMPI for Human Development
Ulman LindenbergerMPI for Human Development
Aaron PeikertMPI for Human Development
Tali SharotUniversity College London
Stephanie ThevesMPI for Human Development
Arno VillringerMPI for Human Cognitive and Brain Sciences
Markus Werkle-BergnerMPI for Human Development
G. Elliott WimmerUniversity College London

 

COMP2PSYCH gives predoctoral students the opportunity to learn, apply, and develop computational and statistical methods that foster our understanding of individual development from childhood to old age, with an emphasis on mental illness and healthy cognitive aging. It combines and integrates training in computer science, applied mathematics and statistics, psychology, and psychiatry with the goal of enabling the students to adapt and develop computational and statistical tools according to their data-analytic and scientific needs in the fields of psychiatry and lifespan psychology. Teaching fundamental concepts in cognitive science, lifespan psychology, and psychiatry in relation to computational and statistical theories and methods creates a superior degree of synergy between fields. In addition, the students benefit from the excellent research opportunities offered at both the Berlin and London sites. Predoctoral students have been recruited from one of the disciplines relevant to the school's program, for example cognitive neuroscience, psychology, psychiatry, health science, or statistics.

During the reporting period, 11 doctoral students were active in the IMPRS COMP2PSYCH in Berlin (see Table 2). Nine worked at the MPI for Human Development: five in the Center for Lifespan Psychology (LIP), two at the Center for Environmental Neuroscience (CEN), one at the Max Planck Research Group NeuroCode (Neural and Computational Basis of Learning, Memory and Decision Making), and one at the ERC Research Group AMD (Adaptive Memory and Decision-making). In addition, two students worked at the Department of Psychology, Humboldt-Universität zu Berlin. Five successfully completed their dissertations and joined the growing group of COMP2PSYCH alumni. During the reporting period, 12 students worked on their dissertation projects in London (see https://www.imprs-comp2psych.mpg.de).

 

Table 2. Berlin COMP2PSYCH Students' Affiliations and Dissertation Projects (2023–04/2026)
NameInstitutionDissertation Project
Manuel Arnold*HUScore-based approaches to heterogeneity in psychological models
Fabio BauerMPIBAbstraction models of perception and working memory
Moana BeyerMPIBAssociations between the physical environment and brain structure and function in different cohort studies
Moritz KetzerHUPotentials and limitations of new approaches to Bayesian SEM
Agnieszka KuleszaMPIBNeuromodulation, plasticity and cognition: Unlocking their interplay across timescales
Anika Löwe*MPIBInvestigating the computational role of hippocampal replay in planning and value computation
Zoya MoorajMPIBCross-sectional vs. longitudinal approaches to understanding the aging brain
Aaron Peikert*MPIBTowards transparency and Open Science: A principled perspective on computational reproducibility and preregistration
Kira PohlmannMPIBComputer science methods in environmental psychology
Liliana Polanski*MPIBComputational and neural mechanisms of human exploration-exploitation behavior
Alexander Skowron*MPIBNeural variability as a marker of cognitive flexibility during learning and decision-making under uncertainty
* PhD thesis has been defended.
Humboldt-Universität zu Berlin (HU), MPI for Human Development (MPIB).

The three-year training program involves seminars, workshops, participation in summer schools, and collaborative supervision of research training. The focus of the Berlin site is on computational and statistical advances in studying individual differences in development from childhood to old age. Methods include structural and functional neuroimaging, computational modeling, and statistical methods for the multivariate analysis of longitudinal changes in brain and behavior. Table 3 shows the talks and workshops offered for COMP2PSYCH students during the reporting period.

 

Table 3. Talks & Workshops for COMP2PSYCH Students (2023–04/2026)
SemesterTopicInstructor(s)
Winter 2022/23Brain network phenotype in neuropsychiatric disorders: Targeting early stageHelen Zhou, University of Singapore
Trait anxiety is associated with hidden state inference during aversive reversal learningOndrej Zika, MPIB
Measuring intelligent use of cognitive resources in online chessEvan Russek, Princeton University
How does anxiety impair computing uncertainty?Payam Piray, University of Southern California
Decomposing the motivation to exert mental effortAmitai Shenhav, University of California, Berkeley
Imaging memory consolidation in wakefulness and sleepMonika Schönauer, Princeton University
Behavioural economics of striatal dopamine releaseNeir Eshel, Stanford University
Neurobiological effects of antipsychotic medication in psychotic depressionNicholas Neufeld, University of Toronto
Reward-predictive state abstractions in human reinforcement learningAlana Jaskir, New York University
Artificial intelligence-based screening of electronic health records to detect individuals at risk of psychosisPaolo Fusar-Poli, Brown University
Summer 2023A computational account of how affective states influence economic decisionsRoeland Heerema, UCL
Irrational decision-making through inhibitory competition within multiple stages of processingKonstantinos Tsetsos, University of Bristol
Neural mechanism of value-based economic decisions: Examples from humans and monkeysLeo Chi Seak, UCL
Towards an explanatory empirical programme in developmental cognitive neuroscienceAnastasia Christakou, University of Reading
Self—other representation in borderline personality disorderGiles Story, UCL
Emotion regulation, emotion beliefs, and adolescent mental healthMatt Somerville, UCL
Tethered rationality: A model of behavior for the real worldVinod Goel, York University
The dynamic nature of procrastinationPeiyuan Zhang, MPIB
Workshop "An introduction to the Julia programming language"Aaron Peikert, MPIB
Two-day scientific presentation trainingSteve Weir, Freie Universität Berlin
Natural language processing enables a paradigm shift in understanding aphasia: Introducing the information restoration paradigmNeguine Rezaii, Harvard Medical School
Winter 2023/24A cellular basis for mapping behavioural structureMohamady El-Gaby, University of Oxford
Human decision-making under computational complexityPeter Boasserts, University of Cambridge
Computational principles of real-world memoryKenneth Norman, Princeton University
Perceptions of artificial intelligenceNadia Said, University of Tübingen
Attenuated flexible behaviour through the lens of reinforcement learning across diagnosesLei Zhang, University of Birmingham
Investigating complex dynamics of inclusion and exclusion in the Cyberball game in Borderline Personality DisorderPaolo Ossola, University of Parma
Mental health across the life courseAvshalom Caspi, King's College London
Towards naturalistic representation learning in health and diseaseAngela Radulescu, Icahn School of Medicine at Mount Sinai
Causality in mind: Learning, reasoning and blamingDavid Lagnado, UCL
Moodeling: A neuro-computational assessment of mood fluctuations and their impact on decision-makingFabien Vinckier, Université de Paris
Human cognition of geometric shapes: A window into the mental representation of abstract conceptsMathias Sable-Meyer, UCL
Computational network models for flexible decision making in health and diseaseNuttida Rungratsameetaweemana, Columbia University
Steering by the torch of chaos and doubtCharley Wu, University of Tübingen
Two-day workshop "Good scientific practice"Roland Bluhm
A recurrent network model of planning predicts hippocampal replay and human behaviourMarcelo Mattar, New York University
Spectral graph theory: An introduction and some applicationsFrancesco Vigano, Imperial College London
Summer 2024Staying afloat on Neurath's boatNeil Bramley, University of Edinburgh
Leveraging brain-based predictive models to understand the neural underpinnings of psychiatric riskElvisha Dhamala, Feinstein Institutes for Medical Research
Addressing two classic debates in cognitive science with AIBrenden Lake, New York University
Functional and dynamic characteristics of neuronal assemblies explored via EEG and graph theory: from healthy aging to neurodegenerative dementiasPaolo Rossini, Università Cattolica del Sacro Cuore
Approaching perceptual choices as internal actionsShervin Safavi, MPI for Biological Cybernetics, Tübingen
Progressive disturbances of vision and space perception: Insights from posterior cortical atrophyKeir Yong, UCL
Dynamic computational phenotyping of human cognitionDaniel Reznik & Roey Schurr, Harvard University
A prospective code for value in the serotonin systemEmerson Harkin, MPI for Biological Cybernetics, Tübingen
Winter 2024/25Memory consolidation during sleep: Mechanisms and representationsBernhard Staresina, University of Oxford
A computational perspective on causal interventions in psychiatryLilian Weber, University of Oxford
Neurocomputational underpinnings of confidence in the context of prior beliefsKobe Desender, KU Leuven
Dynamics of neural representationTimothy O'Leary, University of Cambridge
Learning of successor representations through on-task replay in the visual cortexNicolas Schuck, Universität Hamburg
Cognitive neuroscience of creativityWilliam Orwig, Harvard University
Safe and effective generative clinical AI in mental healthcareMax Rollwage, Limbic
Hippocampal and prefrontal representations underlying flexible decision-makingMona Garvert, Julius-Maximilians-Universität Würzburg
Earworms, memory consolidation, and neural replay for recently heard musicBenjamin Kubit, Princeton University
The neural code underlying flexible choice and compositional reasoningSebastijan Veselic, University of Oxford
Gamma-hydroxybutyrate as a sleep medication in depressionFrancesco Bavato, University of Zurich
On the alleviation of agony: Brain surgery for refractory painHarith Akram, UCL
The "bad bot" apocalypseJon Roozenbeek, King's College London
Parameterising inter-individual diversity in brain activity across health and Parkinson's diseaseJason Da Silva Castanheira, McGill University
Summer 2025Psychopathology through the lens of transdiagnostic decision neuroscienceMichael Hallquist, University of North Carolina at Chapel Hil
Modelling strategic reconfiguration of task set in task-switching paradigmsKevin Lloyd, MPI for Biological Cybernetics, Tübingen
Episodic memory facilitates flexible decision making via access to individual events and their detailsJonathan Nicholas, New York University
Correspondence between reinforcement learning phenotypes and transdiagnostic clinical symptomatology across development Noam Goldway, Kings College London
Metacontrol is guided by task structureWouter Kool, Washington University
Understanding others: How social cognition shapes social behavior and daily social functioning in humans and non-human primatesAnita Tusche, Queen's University, Canada 
Internal uncertainty and abstraction in the development of social influence and social information processing across adolescenceSimon Ciranka, MPIB
A corticostriatal learning model of hallucinationsGuillermo Horga, Columbia University
Environmental inferences drive distinct emotions that shape behaviourEran Eldar, Hebrew University
How metacognition is distorted in anxietySucharit Katyal, Copenhagen University
Memory processes supporting early skill learningLeonardo Cohen, NIH
Winter 2025/26Deciding under uncertainty: Cognitive mechanisms controlling choice variability and information-seekingMarion Rouault, CNRS
Distinct neural mechanisms of observational learningPyungwon Kang, University of Zurich
Meta-decision-making in hierarchical planningTianyuan Teng, MPI for Biological Cybernetics, Tübingen
Using computational approaches to uncover the effects of psychological interventionsIsabel Berwian, UCL
Language models for understanding mental healthShir Lissak, Technion
Exploring high-level cognition through large language models Ariel Goldstein, University of Cambridge
From bowel to brain: The translational journey of an anti-constipation agent into mental healthAngharad de Cates, University of Birmingham
Expanding the value function: Intrinsic rewards and moodBastien Blain, Sorbonne University

The strongest integrative components of the interdisciplinary and international COMP2PSYCH program are the annual meetings in which students and faculty from all COMP2PSYCH institutions participate. Table 4 shows the Academy events that took place during the reporting period.

 

Table 4. COMP2PSYCH Academies 2023–2026
EventVenueDate
Academy 2023Berlin08.–10.11.2023
NSAS Advanced Course in AI and Cognitive NeuroscienceSan Servolo, Venice07.–14.09.2024
Academy 2025London15.–17.11.2025

 

Furthermore, students meet on a regular basis virtually to present their upcoming, in progress, or completed research and receive in-depth feedback by students, postdocs, and faculty.

 

Go to Editor View