Validation and Refinement of a Laminar Neural Mass Model Using in vivo Mice Data

Abstract

Gamma oscillations (30-80 Hz) play a crucial role in cognitive functions and are associated with neurological disorders, including Alzheimer’s disease. Non-invasive brain stimulation techniques, such as 40 Hz transcranial alternating current stimulation (tACS), offer potential in modulating these oscillations and impact cognitive functions. The complexity of the brain, however, necessitates the use of advanced models for effective understanding and the development of therapies. This study aims to validate a framework combining Neural Mass Models (NMMs) with volume conduction physics that takes into account the brain’s physical properties and the distribution of synapses across cortical layers. The validation involves predicting a synaptic distribution across various neuronal groups and employing a Genetic Algorithm (GA) to iteratively refine the model to match experimental data. Key findings include the ability of the NMM to achieve greater similarity with experimental results by varying stochastic noise and the dominance of gamma and alpha oscillations in experimental data aligning well with model predictions. The GA also shows robustness in fitting the model to experimental data, and the predicted synaptic distribution is evaluated against existing literature for physiological accuracy. Despite limitations, our enhanced NMM provides valuable insights into cortical layer interactions, contributing to the understanding of human brain function and the development of treatments for neurological disorders.