Team

Andreea Diaconescu

Andreea Diaconescu PhD, Assistant Professor University of Toronto Personal Page

Dr Andreea Diaconescu is an Independent Scientist at the Krembil Centre for Neuroinformatics at CAMH and Assistant Professor in the Department of Psychiatry at the University of Toronto.

Andreea worked as a Junior Group Leader supported by the Swiss National Foundation (Ambizione grant) at the University in Basel in the Department of Psychiatry. There, she led a project on early detection of psychosis using neurocomputational models of persecutory ideation fit to behaviour and neuroimaging (EEG and fMRI) data. After completing her PhD in Cognitive Neuroscience with Prof. Randy McIntosh at the Rotman Research Institute (University of Toronto), Andreea held a postdoctoral position at the Translational Neuromodeling Unit (University of Zurich and ETH Zurich). Under the supervision of Prof. Klaas Enno Stephan, she developed and validated computational models of social learning and decision-making.

At the Krembil Centre for Neuroinformatics, Andreea focuses on neurocomputational models of suicidal ideation in psychoaffective disorders.

Colleen Charlton

Colleen Charlton MSc, Research Analyst

Colleen Charlton received her BScH in Life Sciences with a specialization in Neuroscience, followed by a Graduate Diploma in Biomedical Informatics at Queen’s University in 2018. She completed her MSc degree in Cognitive Science from the University of Edinburgh in 2020 where she took courses in computational neuroscience and machine learning. There, she completed a project on explainable AI in healthcare which investigated the prediction of brain cancer survival using interpretable machine learning techniques. Colleen is interested in the use of computational modelling and machine learning to identify biomarkers for diagnosis and treatment response in psychiatric disorders.

In the CogneMo group, Colleen is working on two projects: identifying neurocomputational markers of clinical high risk for psychosis from electrophysiological data and modelling the rapid antidepressant and antisuicidal effects of ketamine. She is interested in using computational parameters to make individual treatment response predictions in psychoaffective disorders.

Daniel Hauke

Daniel Hauke, PhD Candidate University of Basel, Computer Science Personal Page

Daniel Hauke obtained a Bachelor of Science in Psychology in 2014, after studying at the Georg-August-University Goettingen, Germany, and the Universidade Federal do Ceará in Fortaleza, Brazil. He received his Master of Science degree in Cognitive and Clinical Neuroscience in 2016 following studies at Maastricht University, the Netherlands, and the Translational Neuromodeling Unit, University of Zurich and ETH Zurich, Switzerland. He is now enrolled as doctoral student in the computer science department of the faculty for natural sciences at Basel University, as well as a PhD student in the doctoral program for Data Analytics and an external PhD student at the Neuroscience Center Zurich (ZNZ).

Currently, Daniel Hauke is working in an interdisciplinary project which attempts to exploit generative models, such as the Hierarchical Gaussian Filter and Dynamic Causal Modelling, to investigate the formation and persistence of persecutory delusions in psychosis. He examines the clinical utility of these models to predict transition to psychosis with machine learning.

Peter Bedford

Peter Bedford HBSc, Research Analyst

Peter Bedford completed in 2018 a HBSc specializing in Physics and majoring in Mathematics at the University of Toronto, where he also worked as a Research Assistant with Dr. Erich Poppitz on computational modelling of vortons. In 2019, Peter briefly worked under Dr. Andrew Dimitrijevic at Sunnybrook Research Institute, where he used wireless mobile EEG equipment for the detection of Auditory Steady State Responses (ASSRs).

Peter is currently working on two projects with the CogneMo group, which both involve applying machine learning models using brain connectivity (specifically, whole-brain effective connectivity) as features. The first project aims to show that subjects’ normal brain connectivities are predictive of the types of alterations to their subjective experience after taking lysergic acid diethylamide (LSD); the other aims to show that, for patients with major depression disorder (MDD), their pre-treatment brain connectivities are predictive of their level of symptom amelioration post-treatment.

Peter is interested in the connection between perceptual ability and cognition—in particular, the question of whether perceptual training methods such as those used for inducing synaesthetic phenomenology can also lead to cognitive benefits like enhanced memory.

Povilas Karvelis

Povilas Karvelis PhD, Postdoctoral Fellow University of Toronto Personal Page

Dr. Povilas Karvelis completed his PhD under the supervision of Dr. Peggy Series at the Institute of Adaptive and Neural Computation at the University of Edinburgh, where he investigated perceptual Bayesian inference in autism and schizophrenia. His preceding degrees include MSc Computational Cognitive Science at the University of Edinburgh and MSci Physics and Astrophysics at the University of Birmingham.

Here in the CogneMo group, Povilas is involved in developing neurocomputational models of suicidality. This work integrates normative computational frameworks of Active Inference and Hierarchical Gaussian Filter, probabilistic learning tasks and fMRI recordings to probe maladaptive inference and action selection on behavioural, cognitive, and neurobiological levels. The ultimate aim of this work is to develop tools for accurately assessing the risk of suicidality that would allow improving suicide prevention interventions.