Identifying spiking neurons for local field potentials

This project aims to advance our understanding of how neural circuits encode motor function and make decisions by developing innovative analytical methods for neural data. Current studies analyze brain activity using two signals: spiking neurons, reflecting individual neuron activity, and local field potentials (LFPs), representing collective signals from multiple neurons. While LFPs offer valuable insights into population dynamics, identifying the specific contributions of individual neurons remains a challenge. To address this, we propose extending the multiscale subspace identification (multiscale SID) algorithm introduced by Ahmadipour et al. (2024). This approach integrates spiking activity and LFPs but currently lacks a strong theoretical foundation.

The project focuses on applying and enhancing this method to analyze neural data recorded from frontoparietal regions of monkeys performing a complex decision-making task with conflicting information. By decomposing LFP signals and pinpointing individual neuron contributions, we aim to create a detailed map of neural networks active during decision-making processes. This work will use advanced signal processing, time-series analysis, and statistical mechanics to bridge the gap between empirical data and theoretical models.

The expected benefits include deeper insights into decision-making at cellular and network levels, fostering collaboration between participating institutions and advancing cognitive neuroscience with tools for experimental validation.

Faculty Supervisor:

Paul Cisek

Student:

Partner:

University of Barcelona

Discipline:

Life Sciences

Sector:

Education

University:

Université de Montréal

Program: