Neuroscience is a remarkably cross-disciplinary area of research that spans traditional departmental boundaries from physiology to genetics, molecular and cellular biology to mathematics, engineering and artificial intelligence.
The Centre for Integrative Neuroscience Discovery has funded three PhD studentships for candidates undertaking a programme of study across diverse areas of neuroscience, each supervised by two PIs from different departments.
Two of these researchers share highlights of their PhD work on learning in neuropsychiatric diseases and brain-machine interfaces for virtual navigation.
Katharina Zuhlsdorff
Katharina worked with PIs Jeff Dalley in Psychology and Stephen Eglen in Applied Mathematics and Theoretical Physics to investigating reinforcement learning processes in depression and substance use disorder. She explains:
“My research aims to identify biological markers that may predispose to mental health conditions, with a particular focus on cognitive flexibility and impulsivity, which are cognitive functions known to be disrupted in psychiatric disorders. Cognitive flexibility is defined as the ability to adapt to changes in the environment by switching task sets, responses, or strategies, whereas impulsivity is the tendency to act prematurely leading to rash actions without forethought. Specifically, I use computational algorithms to model behavioural data and investigate the underlying neural circuits and their dysfunction via functional magnetic resonance imaging (MRI). Moreover, I develop machine learning algorithms that combine multimodal data, e.g., neuroimaging, cognitive and genetic data, with the aim of building a tool for early prediction of neuropsychiatric disease.
One major strand of my research focuses on reinforcement learning. Reinforcement learning (RL) is the process by which an animal utilises its previous experience in order to improve outcomes of future choices by maximising reward and minimising punishment. My research investigates how RL processes are altered in neuropsychiatric diseases, specifically in Major Depressive Disorder (MDD) and Substance Use Disorder (SUD). Reversal learning tasks are used to study cognitive flexibility, as they require a subject to flexibly adjust their behaviour when stimulus-related contingencies are reversed. RL models can be fitted to reversal learning tasks, allowing a deeper insight into the latent mechanisms underlying behaviour as they involve a trial-by-trial approach. This is in contrast to conventional approaches, which determine behavioural measures based on the average of all trials, failing to take into account behavioural dynamics within a session.”
Katharina defended her PhD thesis in September 2022 and is now the Angharad Dodds John Fellow in Neuropsychiatry and Mental Health at Downing College, and a Research Associate in the Department of Psychology. She works on the application of deep learning algorithms to large-scale, multimodal datasets to identify patterns that help us better understand the aetiology of psychiatric diseases.
Katharina says: “This funding has enabled me to publish and present my research, providing a basis for the future research work that I will conduct. It has enabled me to create a unique skillset that I hope to apply to further questions within the field of neuroscience.”
Ethan Sorrell
Ethan is working with Timothy O’Leary in Engineering and with Chris Harvey and Dan Wilson from the Harvey Lab at Harvard Medical School, on brain-machine interfaces for virtual navigation. He tells us:
“My PhD project centres around a set of Brain-Machine Interface (BMI) experiments we have performed in collaboration with the Harvey Lab at Harvard Medical School. It addresses a few questions. The first is purely practical: Can mice use a simple linear brain-machine interface that decodes directly from images of the brain to successfully navigate a virtual environment. We have shown that this is possible, which opens up opportunities for future neuroscience experiments utilising our BMI to understand the learning, plasticity and function in various brain regions.
My PhD work also addresses questions related to the stability and function of the brain signals we recorded from. Are these brain signals sufficient for driving navigation (which is answered with the success of the BMI as above), and do these signals remain stable when closing the BMI loop and using the BMI for several days. We discovered that the neural signals were remarkably stable when closing the loop, and remained sufficiently stable over 4 days of experiments. This is a great set up for more longitudinal future experiments, to assess the long term effects of BMI use on neural plasticity.
My PhD involves collaboration between engineers and neuroscientists. The experiments we perform require the expertise and experience of neuroscientists to set up and run the physical experiments. My contribution comes through writing the code for the decoder and BMI setup, and performing data analysis on the results and the data collected. I regularly meet with researchers in the Control Group of the Engineering Department, as well as our neuroscience collaborators from the Harvey Lab at Harvard.”
Ethan plans to complete his PhD work in autumn 2023. He says: “This funding has allowed me to complete an interdisciplinary PhD, gaining insight into engineering and neuroscience research. It has helped me make connections across the Atlantic, and given me the opportunity to visit another top university and work with other leading researchers. My current plans for the future are to experience work outside of academia, in industry, possibly with a start-up company with a medical and preferably BMI focus.”
CIND is pleased to have supported Katharina and Ethan’s contributions to interdisciplinary neuroscience research and professional development.