Spiking Laguerre Volterra networks-predicting neuronal activity from local field potentials

Area of Science:

• Computational Neuroscience
• Neurophysiology
• Biomedical Engineering

Background:

• Understanding the relationship between local field potentials (LFP) and neuronal spiking is key to deciphering brain information processing.
• Current methods primarily quantify LFP-spike coupling, with limited success in predicting precise spike timing from LFP variations.

Purpose of the Study:

• To develop novel, interpretable models for predicting spike timing based on LFP dynamics.
• To introduce spiking Laguerre-Volterra network (sLVN) models that offer explainable insights into neural signal relationships.

Main Methods:

• Application of sLVN models to extracellular microelectrode recordings from Parkinson's Disease patients undergoing deep brain stimulation (DBS).
• Analysis of LFP-spike pair predictability to identify distinct neuronal populations.

Main Results:

• Identification of three neuronal populations characterized by unique signal properties and sLVN model features.
• Observed indirect associations between these neuronal clusters and motor score improvements post-DBS.

Conclusions:

• The predictability of spiking activity shows potential as an intraoperative biomarker for optimizing DBS lead placement.
• Further research is warranted to validate spiking activity predictability for improved surgical outcomes in Parkinson's Disease.