COMPUTATIONAL PSYCHOPATHOLOGY RESEARCH GROUP
  • About Us
  • People
  • Publications
  • Announcements
Picture
The Computational Psychopathology Lab is a research group that studies human behaviour and mental health disorders using a combination of experiments and computational modelling. We investigate how individuals learn about the causal structures of their environments, with a focus on how they perceive agency—that is, the control exerted by themselves and others—and how they reason based on these learned causal structures.
Our research includes work with both humans and non-human animals and examines behaviour across developmental stages, from childhood to adulthood. We work closely with researchers from universities, industry professionals, and clinicians to integrate computational tools with clinical expertise. The computational models studied and developed in our lab contribute to fields such as robotics and artificial intelligence by improving our understanding of how learning and decision-making processes work. These models also provide insights into mental health, helping to identify patterns and mechanisms that could inform strategies for diagnosis, prevention, and treatment.

CURRENT RESEARCH PROJECTS

Agency

Our goal here is to understand how individuals learn and infer about their own level of control and that of others in both individual and group contexts. Given the well-established link between increased depression symptoms and a diminished sense of agency, we investigate how agential learning is affected by varying levels of depressive symptoms, including those associated with clinical depression.

To gather data, we employ an adapted free-operant task, where participants freely perform actions, observe the actions of a second agent, and integrate this information with feedback on the outcomes. This self-initiated approach enables us to analyse participants' environmental sampling behaviours and their perceptions of control over outcomes for themselves and the other agent.

Through computational modelling, we aim to illuminate the underlying mechanisms of agential learning. We evaluate a range of models, including associative, statistical, Bayesian, and active inference approaches, to better understand how individuals navigate and interpret their experiences of control.

Relevant articles:
  • https://osf.io/preprints/osf/ckfrt
  • https://link.springer.com/chapter/10.1007/978-3-031-47958-8_8 (or see https://osf.io/preprints/psyarxiv/2gn54 for free access to the pre-print)

Types of Information Informing Contingency Learning

A classic experiment to assess whether individuals accurately perceive contingencies involves a stream of one-second-long trials that participants observe, wherein each trial is has one type of information: "A" is cue and outcome being co-present, "B" is cue being present and outcome being absent, "C" is cue being absent and outcome being present, and "D" is cue and outcome being co-absent. We manipulate the contingencies between the cue and outcome presented to the individuals by varying the proportion of A, B, C, D events presented to participants and participants provide a rating on the relationship between the cue and outcome. We are interested in the role of D events. 

Relevant articles: 
  • https://ora.ox.ac.uk/objects/uuid:ac511c57-e2b6-490a-8cef-7e29a1777b0a/files/s05741t50b
  • https://ora.ox.ac.uk/objects/uuid:354067a0-1768-402a-8a73-45b1fccfe9f4/files/r736664814
​

Logical Reasoning

In this research project, we focus on the early development of logical reasoning in children and whether similar capabilities exist in non-human animals. By examining how abstract concepts and inferences emerge using both experimental methods and computational models, we aim to shed light on the unique aspects of human cognition and explore potential parallels in other intelligent animals. This work could enhance our understanding of cognitive development and offer insights into the broader landscape of intelligence across species and technologies.

Relevant articles: 
  • https://www.researchgate.net/publication/384928560_Reasoning_by_exclusion_in_food-caching_Eurasian_jays_Garrulus_glandarius
  • https://www.pnas.org/doi/10.1073/pnas.2207499119
​

CURRENT COLLABORATIONS

​
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture

PREVIOUS COLLABORATIONS

Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Proudly powered by Weebly
  • About Us
  • People
  • Publications
  • Announcements