Model-based signal processing enables bidirectional inferring between local field potential and spikes evoked by...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Signal Processing

Background:

• Local field potentials (LFPs) are widely used to measure neural population activity but are complex to interpret.
• A significant gap exists in understanding the precise relationship between individual neuronal spiking activity and LFP signals.
• This study addresses the need to link spiking activity and LFP, particularly under evoked noxious conditions.

Purpose of the Study:

• To hypothesize and test if LFP signals can be accurately inferred from neuronal spike data.
• To develop and validate a method for extracting C-fiber evoked activity from neural recordings.
• To establish a bidirectional relationship between spike data and LFP signals.

Main Methods:

• Extracellular recordings were conducted in the medullary dorsal horn (MDH) of anesthetized rats.
• A model-based approach was used to isolate C-fiber evoked activity by removing A-fiber contributions.
• Convolution kernel theory and optimization algorithms were applied to infer LFP from spikes and vice versa.

Main Results:

• C-fiber LFP was successfully extracted from all recordings.
• C-fiber spikes were found to precede the LFP and correlate with its derivative.
• Inferred LFP from spikes showed high correlation (r=0.9); spike distribution from LFP showed good correlation (r=0.7).

Conclusions:

• Kernel convolution theory provides a robust method for inferring LFP from spikes.
• The study successfully demonstrated the ability to generate spike distributions from LFP data.
• This work bridges the gap between single-unit activity and population-level LFP signals.