Bryan Li

Position

Enrichment Student

Cohort year

2023

Partner Institution

Bio

Bryan is a PhD student in the Biomedical AI CDT program at the University of Edinburgh, under the supervision of Dr. Arno Onken and Dr. Nathalie Rochefort. His main research focuses on modelling neuronal activities recorded from the visual cortex using deep learning methods and tries to make sense of how neural responses reshape over experience. In addition to his interests in computational neuroscience, Bryan is also part of an international collaboration with the University of Barcelona to tackle challenges in machine learning for mental healthcare, specifically in identifying digital biomarkers from wearable data for mood disorders. Prior to joining the University of Edinburgh, Bryan was a Machine Learning Research Engineer at Huawei Noah’s Ark Lab in Toronto where he conducted research in computer vision and human-computer interaction tasks. He received his undergraduate degree in Computer Science from the University of Toronto with distinction.

Research interests

Understanding how the visual system processes information is a fundamental challenge in neuroscience. As the capability of recording large amounts of neural and behavioural data expands, there is increasing interest in modelling neural dynamics during adaptive behaviours in order to understand neuronal encoding of sensory and motor information. A popular approach to identify the underlying computations of the visual system is by building high-performing predictive models that can accurately predict neural responses to naturally occurring stimuli. These models can explain a large part of the stimulus-driven variability and account for the nonlinear response properties of the neural activity, thus allowing computational neuroscientists to generate new hypotheses about biological vision and bridge the gap between biological and computer vision. To that end, Bryan’s research project aims to utilise the latest advancement in deep learning to design and build these biological digital twins and join behavioural and neural data to reveal neural dynamics.