Bayesian Workflow for Generative Modeling in Computational Psychiatry

Computational (generative) modelling of behaviour has considerable potential for clinical applications. In order to unlock the potential of generative models, reliable statistical inference is crucial. For this, Bayesian workflow has been suggested which, however, has rarely been applied in Translat...

Full description

Saved in:
Bibliographic Details
Main Authors: Alexander J. Hess, Sandra Iglesias, Laura Köchli, Stephanie Marino, Matthias Müller-Schrader, Lionel Rigoux, Christoph Mathys, Olivia K. Harrison, Jakob Heinzle, Stefan Frässle, Klaas Enno Stephan
Format: Article
Language:English
Published: Ubiquity Press 2025-03-01
Series:Computational Psychiatry
Subjects:
Online Access:https://account.cpsyjournal.org/index.php/up-j-cp/article/view/116
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850197259373772800
author Alexander J. Hess
Sandra Iglesias
Laura Köchli
Stephanie Marino
Matthias Müller-Schrader
Lionel Rigoux
Christoph Mathys
Olivia K. Harrison
Jakob Heinzle
Stefan Frässle
Klaas Enno Stephan
author_facet Alexander J. Hess
Sandra Iglesias
Laura Köchli
Stephanie Marino
Matthias Müller-Schrader
Lionel Rigoux
Christoph Mathys
Olivia K. Harrison
Jakob Heinzle
Stefan Frässle
Klaas Enno Stephan
author_sort Alexander J. Hess
collection DOAJ
description Computational (generative) modelling of behaviour has considerable potential for clinical applications. In order to unlock the potential of generative models, reliable statistical inference is crucial. For this, Bayesian workflow has been suggested which, however, has rarely been applied in Translational Neuromodeling and Computational Psychiatry (TN/CP) so far. Here, we present a worked example of Bayesian workflow in the context of a typical application scenario for TN/CP. This application example uses Hierarchical Gaussian Filter (HGF) models, a family of computational models for hierarchical Bayesian belief updating. When equipped with a suitable response model, HGF models can be fit to behavioural data from cognitive tasks; these data frequently consist of binary responses and are typically univariate. This poses challenges for statistical inference due to the limited information contained in such data. We present a novel set of response models that allow for simultaneous inference from multivariate (here: two) behavioural data types. Using both simulations and empirical data from a speed-incentivised associative reward learning (SPIRL) task, we show that models harnessing information from two different data streams (binary responses and continuous response times) ensure robust inference (specifically, identifiability of parameters and models). Moreover, we find a linear relationship between log-transformed response times in the SPIRL task and participants’ uncertainty about the outcome. Our analysis illustrates the benefits of Bayesian workflow for a typical use case in TN/CP. We argue that adopting Bayesian workflow for generative modelling helps increase the transparency and robustness of results, which in turn is of fundamental importance for the long-term success of TN/CP.
format Article
id doaj-art-168787b3493449ea80e3af6d59d1a452
institution OA Journals
issn 2379-6227
language English
publishDate 2025-03-01
publisher Ubiquity Press
record_format Article
series Computational Psychiatry
spelling doaj-art-168787b3493449ea80e3af6d59d1a4522025-08-20T02:13:14ZengUbiquity PressComputational Psychiatry2379-62272025-03-019176–9976–9910.5334/cpsy.11693Bayesian Workflow for Generative Modeling in Computational PsychiatryAlexander J. Hess0https://orcid.org/0000-0002-2991-7387Sandra Iglesias1https://orcid.org/0000-0002-1778-7239Laura Köchli2https://orcid.org/0009-0001-0038-9592Stephanie Marino3https://orcid.org/0000-0002-6559-9072Matthias Müller-Schrader4https://orcid.org/0000-0001-6672-9230Lionel Rigoux5https://orcid.org/0000-0003-3761-8931Christoph Mathys6https://orcid.org/0000-0003-4079-5453Olivia K. Harrison7https://orcid.org/0000-0003-0897-7142Jakob Heinzle8https://orcid.org/0000-0001-5228-041XStefan Frässle9https://orcid.org/0000-0002-8011-2226Klaas Enno Stephan10https://orcid.org/0000-0002-8594-9092Translational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, ZurichTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, ZurichTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, ZurichTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, ZurichTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, ZurichMax Planck Institute for Metabolism Research, CologneInteracting Minds Centre, Aarhus University, AarhusTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, Zurich, CH; Department of Psychology, University of Otago, DunedinTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, ZurichTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, ZurichTranslational Neuromodeling Unit, Institute for Biomedical Engineering, University of Zurich and ETH Zurich, Zurich, CH; Max Planck Institute for Metabolism Research, CologneComputational (generative) modelling of behaviour has considerable potential for clinical applications. In order to unlock the potential of generative models, reliable statistical inference is crucial. For this, Bayesian workflow has been suggested which, however, has rarely been applied in Translational Neuromodeling and Computational Psychiatry (TN/CP) so far. Here, we present a worked example of Bayesian workflow in the context of a typical application scenario for TN/CP. This application example uses Hierarchical Gaussian Filter (HGF) models, a family of computational models for hierarchical Bayesian belief updating. When equipped with a suitable response model, HGF models can be fit to behavioural data from cognitive tasks; these data frequently consist of binary responses and are typically univariate. This poses challenges for statistical inference due to the limited information contained in such data. We present a novel set of response models that allow for simultaneous inference from multivariate (here: two) behavioural data types. Using both simulations and empirical data from a speed-incentivised associative reward learning (SPIRL) task, we show that models harnessing information from two different data streams (binary responses and continuous response times) ensure robust inference (specifically, identifiability of parameters and models). Moreover, we find a linear relationship between log-transformed response times in the SPIRL task and participants’ uncertainty about the outcome. Our analysis illustrates the benefits of Bayesian workflow for a typical use case in TN/CP. We argue that adopting Bayesian workflow for generative modelling helps increase the transparency and robustness of results, which in turn is of fundamental importance for the long-term success of TN/CP.https://account.cpsyjournal.org/index.php/up-j-cp/article/view/116translational neuromodelingcomputational psychiatrybayesian workflowhierarchical gaussian filter (hgf)multimodal response modelsrobust inference
spellingShingle Alexander J. Hess
Sandra Iglesias
Laura Köchli
Stephanie Marino
Matthias Müller-Schrader
Lionel Rigoux
Christoph Mathys
Olivia K. Harrison
Jakob Heinzle
Stefan Frässle
Klaas Enno Stephan
Bayesian Workflow for Generative Modeling in Computational Psychiatry
Computational Psychiatry
translational neuromodeling
computational psychiatry
bayesian workflow
hierarchical gaussian filter (hgf)
multimodal response models
robust inference
title Bayesian Workflow for Generative Modeling in Computational Psychiatry
title_full Bayesian Workflow for Generative Modeling in Computational Psychiatry
title_fullStr Bayesian Workflow for Generative Modeling in Computational Psychiatry
title_full_unstemmed Bayesian Workflow for Generative Modeling in Computational Psychiatry
title_short Bayesian Workflow for Generative Modeling in Computational Psychiatry
title_sort bayesian workflow for generative modeling in computational psychiatry
topic translational neuromodeling
computational psychiatry
bayesian workflow
hierarchical gaussian filter (hgf)
multimodal response models
robust inference
url https://account.cpsyjournal.org/index.php/up-j-cp/article/view/116
work_keys_str_mv AT alexanderjhess bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT sandraiglesias bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT laurakochli bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT stephaniemarino bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT matthiasmullerschrader bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT lionelrigoux bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT christophmathys bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT oliviakharrison bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT jakobheinzle bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT stefanfrassle bayesianworkflowforgenerativemodelingincomputationalpsychiatry
AT klaasennostephan bayesianworkflowforgenerativemodelingincomputationalpsychiatry