Feindel Brain and Mind Seminar Series: High-Dimensional Neural Mass Models with Distributed-Delay Connectome Tensors

Monday, March 4, 2024 13:00to14:00
Montreal Neurological Institute De Grandpre Communications Centre, 3801 rue University, Montreal, QC, H3A 2B4, CA

The Feindel Brain and Mind Seminar Series will advance the vision of Dr. William Feindel (1918–2014), Former Director of the Neuro (1972–1984), to constantly bridge the clinical and research realms. The talks will highlight the latest advances and discoveries in neuropsychology, cognitive neuroscience, and neuroimaging.

Speakers will include scientists from across The Neuro, as well as colleagues and collaborators locally and from around the world. The series is intended to provide a virtual forum for scientists and trainees to continue to foster interdisciplinary exchanges on the mechanisms, diagnosis and treatment of brain and cognitive disorders.

To attend in person, register here

To watch via Vimeo, click here

Anisleidy Gonzalez Mitjans

Post-Doctoral Researcher, Brain Imaging Center, McGill University, The Neuro

Host: justine.clery [at] mcgill.ca (Justine Clery)

Abstract: The Jansen and Rit Neural Mass Model (JR NMM) serves as a concise yet potent framework for comprehending the dynamics within a cortical column and its interactions with the thalamus. While adept at simulating diverse neural processes and applied in the exploration of phenomena related to epileptic seizures and brain-computer interfaces, the existing algorithms encounter challenges in scaling with an increasing number of neural masses. This limitation hampers real-time feedback and impedes the applicability of Neural Mass Models (NMMs) in resolving EEG/MEG inverse problems. To address these issues, this study introduces a novel approach along with a Distributed-delay Neural Mass Model (DD-NMM) Toolbox, grounded in three pivotal aspects: i. Preservation of Network Dynamics: Leveraging the Local Linearization Method (LLM), numerical methods that may disrupt network properties (attractors) are circumvented. ii. Decoupling of Neural Mass Integration: Enhancing the simulation sampling frequency facilitates treating inputs to each neural mass as exogenous. This, in turn, streamlines the symbolic solution of the corresponding equations. iii. Efficient Input Computation: Employing a differential algebraic formulation, a tensor product is utilized between past outputs of all masses and the Connectome Tensor (CT). This innovative approach creates the present input to each NMM, allowing for the modeling of various connectivities and delays, including distributed delays. Through these advancements, this work aims to overcome the scaling challenges faced by current algorithms, paving the way for enhanced real-time feedback and the broader application of NMMs in tackling EEG/MEG inverse problems.

Back to top