A Computational Approach to Understanding Maladaptive Cognition in Depression

Lead Research Organisation: Royal Holloway, University of London
Department Name: Psychology



Depression is the single leading cause of disability worldwide and a major public health problem. Even with the best treatments, around 30% of patients remain unwell, demonstrating the importance of improving our understanding of depression. Decades of research in clinical psychology suggests that vulnerability to depression is associated with negative cognitive styles, such as attributing negative events to stable and global causes, often blaming oneself, and maladaptive metacognitive beliefs (about one's own cognitive processes), such as low self-confidence. These biases are a focus of psychological therapies such as cognitive behavioural therapy (CBT), but the assessment of maladaptive depressive cognition is limited by imprecise measurement, relying on introspection and self-report.


This project aims to improve our understanding of the maladaptive cognitions driving depressive symptoms. To gain a more mechanistic understanding of the neurocognitive bases of (mal)adaptive cognition, we will leverage computational models of behaviour. This overarching goal will be achieved by conceptualising maladaptive depressive cognition as maladaptive attributions. To test this we will measure:
(a) biases in the attribution of positive and negative events to the self vs. external causes;
(b) biases in the metacognitive evaluations of decision confidence and their potential misattribution to action-outcome learning.
Using cutting-edge analysis methods, across online, clinical, and neuroimaging studies, this project will achieve the following objectives:
1. Clarify the neurocognitive mechanisms underlying adaptive attribution (of external events and of metacognitive signals), in healthy participants.
2. Identify behavioural markers of maladaptive attribution related to depressive symptoms in a non-clinical sample.
3. Test the specificity of markers of maladaptive attribution to depressive symptoms, relative to other common mental health problems.
4. Test the clinical relevance of markers of maladaptive attribution.


Improving our understanding of the mechanisms that drive maladaptive cognition in depression, and underpin attributional processes in healthy participants, will constitute an important scientific contribution to the fields of clinical psychology and cognitive and computational neuroscience. Given the high societal cost of depression, this research is of high societal and clinical relevance. Disseminating our findings to the wider society will demonstrate how a better understanding of basic cognitive processes may translate to understanding everyday behaviour. Presenting our project and findings to people with mental health problems, including service users, will allow receiving their feedback on our experimental designs and findings, and help broaden the perspective for future research. The work will also be regularly disseminated to academic audiences, through publications and conferences, across the fields of psychology, neuroscience, and mental health. Engaging with clinical experts, by organising an interdisciplinary workshop, will help increase our clinical impact, establish novel collaborations, and receive expert feedback. Identifying behavioural and neural markers related to maladaptive cognition in depression offers a unique opportunity to develop novel tools that may subsequently help to refine differential diagnosis and improve treatment selection, as well as provide a foundation for the development of novel psychological interventions.

Planned Impact


Improving our understanding of the mechanisms that drive maladaptive cognition in depression, and underpin attributional processes in healthy participants, will constitute an important contribution to scientific knowledge, as well as being of high societal and clinical relevance. This work will contribute to the fields of clinical psychology and cognitive and computational neuroscience, and more specifically research on decision-making, reinforcement learning, metacognition, and sense of agency. By bringing together these research topics, which have remained largely separate, this work will provide novel insights into the interactions between these important human capacities. Using reinforcement learning tasks will allow us to leverage the large literature and computational models available to investigating the under-researched topic of attributional processes, offering a unique opportunity for translational neuroscience.

This research will directly impact an on-going project of a collaborator, aiming to develop novel clinical interventions. Identifying novel behavioural markers of maladaptive cognition will broaden the toolkit for assessment of individuals, to improve differential diagnosis and treatment allocation. To maximise clinical impact, we will target psychiatry-focused conferences and journals for research dissemination, in addition to psychology and neuroscience outlets, and will organise an international workshop with researchers and clinical experts. This will facilitate establishing new collaborations with clinical experts and promote further testing of the clinical relevance of our work, namely regarding their predictive validity of treatment outcomes.

This project will also develop and validate novel experimental paradigms to investigate attributional processes and their interaction with metacognition. Scripts for the experimental task and analyses, as well as de-identified data (within information governance regulations), will be made freely available to researchers in the field who are interested in adapting and re-using them. This creates novel resources for experiments, as well as for conducting re-analyses, and meta-analyses.


We will present our project and findings to services users, i.e. people who suffer from mental health problems, or people working with, or caring for, those who do. The charity MQ Mental Health holds a yearly Mental Health Science meeting, where scientists present their work to other researchers and service users. UCL's North London Service Users Forum will provide further opportunities to receive feedback on our study designs and findings. Engaging with the potential end-users of future diagnostic tools and clinical interventions will thus broaden the perspective for further research. We will also disseminate our research and findings at local NHS Improving Access to Psychological Treatment services, to receive feedback from clinicians.

Sharing our findings with the general public could additionally have a direct impact on their lives, by helping them better understand potentially maladaptive thinking. Research has shown that raising awareness of the potential for illusions of causality, i.e. erroneous causal attributions, and fostering a more analytic assessment of causal relations, can help reduce illusions and cognitive biases. Highlighting counterfactuals ("what would have happened had I done X instead of Y") may also serve to reduce exaggerated self-attribution. Similarly, research on metacognition has shown that awareness of potentially confusing the sources of metacognitive signals (e.g. difficulty) can reduce misattributions, and hence an unwarranted influence on other cognitive processes. Engaging in open discussions about mental illness may additionally help reduce stigma, motivating people to seek help themselves, or better detect warning signs in others.


10 25 50