🎓 Background¶
I am a postdoc scholar at the Baillet lab at McGill University, housed at the Montreal Neurological Institute (the Neuro). My research focuses on the electrophysiology of neurological disorders, with a particular interest in corticothalamic and hippocampal oscillations. I am interested to understand the biological sources of order in the temporal structure of brain activity and they impact cognition, perception, and movement. I completed my doctoral training with Prof. Igor Timofeev at Universite Laval, studying the electrophysiology of sleep-dependent memory consolidation. My master’s studies were completed at the University of Toronto with Prof. William D Hutchison, where I worked on movement disorders, studying recordings obtained during surgery for deep brain stimulation.
📝 Project¶
Leveraging multimodal neuroimaging to forecast risk and resilience in Alzheimer’s Disease¶
The overarching aim of my postdoctoral project is to identify electrophysiological biomarkers of Abeta/tau burden in preclinical Alzheimer’s disease (AD). I will leverage EEG/MEG recordings from the PREVENT-AD dataset in combination with extensive longitudinal data that include structural MRI, Ab/tau PET, cerebrospinal fluid (CSF) biochemistry and extensive clinical evaluations of neurosensory and cognitive functioning. The first specific aim is to develop a computational toolbox for integrating MEG/EEG with MRI, PET and clinical data for staging, understanding, and treating AD. The AD Toolbox will build on existing computational tools within Brainstorm, an open-source software environment developed and maintained by my host lab (Prof.Baillet) and collaborators. Brainstorm is a widely used research tool, with 2,300 published articles reporting analyses performed with Brainstorm, >36,000 user accounts, 18,000 downloads/month, and 2,500 researchers trained in Brainstorm workshops. The AD Toolbox will add functionalities specific to the study of AD pathophysiology and will feature two modules, organized along complementary analytical strategies: regional analysis – focused on cortical areas associated with early-stage AD and whole-brain analysis – focused on large-scale connectivity measures. AD Toolbox will be designed and distributed as an open-source, user-friendly resource, intended to facilitate future use of MEG/EEG data in the study of AD pathophysiology. The second specific aim is to investigate global changes in brain electrophysiology arising from local Abeta/Tau accumulation, making use of the added Brainstorm functionalities to co-register Aβ/tau maps with the topology of large-scale cortical activity. Using the co-registered data, we will train models to forecast voxel-wise estimates of tau and Abeta from MEG, using either hierarchical attention networks (HANs) or convolutional neural networks (CNNs).