Large-scale changes in cortical dynamics triggered by repetitive somatosensory electrical stimulation

Abstract

Background: Repetitive somatosensory electrical stimulation (SES) of forelimb peripheral nerves is a promising therapy; studies have shown that SES can improve motor function in stroke subjects with chronic deficits. However, little is known about how SES can directly modulate neural dynamics. Past studies using SES have primarily used noninvasive methods in human subjects. Here we used electrophysiological recordings from the rodent primary motor cortex (M1) to assess how SES affects neural dynamics at the level of single neurons as well as at the level of mesoscale dynamics.

Methods: We performed acute extracellular recordings in 7 intact adult Long Evans rats under ketamine-xylazine anesthesia while they received transcutaneous SES. We recorded single unit spiking and local field potentials (LFP) in the M1 contralateral to the stimulated arm. We then compared neural firing rate, spike-field coherence (SFC), and power spectral density (PSD) before and after stimulation.

Results: Following SES, the firing rate of a majority of neurons changed significantly from their respective baseline values. There was, however, a diversity of responses; some neurons increased while others decreased their firing rates. Interestingly, SFC, a measure of how a neuron's firing is coupled to mesoscale oscillatory dynamics, increased specifically in the δ-band, also known as the low frequency band (0.3- 4 Hz). This increase appeared to be driven by a change in the phase-locking of broad-spiking, putative pyramidal neurons. These changes in the low frequency range occurred without a significant change in the overall PSD.

Conclusions: Repetitive SES significantly and persistently altered the local cortical dynamics of M1 neurons, changing both firing rates as well as the SFC magnitude in the δ-band. Thus, SES altered the neural firing and coupling to ongoing mesoscale dynamics. Our study provides evidence that SES can directly modulate cortical dynamics.

Keywords: Low frequency oscillations; Motor cortex; Peripheral nerve; Somatosensory electrical stimulation (SES); Spiking dynamics.

Fig. 1

Schematic of the Experiment. a, Somatosensory electrical stimulation was applied directly to the distal forelimb while neural activity was recorded under anesthesia. b, Schematic of the stimulation paradigm. c, Averaged evoked potential in the local field potential during SES

Fig. 2

Changes in Firing Rate after SES. a, Violin plot of the firing rate changes for all neurons. The red cross represents the mean (1.7918); green triangle is median (− 0.2338). b, Example of either a significant decrease (p < 0.05; top) or increase (p < 0.05; bottom) in firing rate after SES. Also shown are tetrode waveforms and the interspike intervals. The dotted lines represent the mean during the pre-stimulation period. c, Percentage of neurons which significantly increased, decreased, or had no change for one animal (top) and for all animals (n = 7; bottom)

Fig. 3

Changes in Spike Field Coherence (SFC) after SES. a, Schematic depicting neural spikes relative to LFP recordings from M1. b, Schematic of the relation of spiking to LFP for variations in the SFC. c, Comparison of the averaged SFC across each frequency band (see Methods) for all units before and after SES. (*p < 0.001). Error bars represent the standard error of the mean or SEM. d, Percentage of neurons which significantly increased, decreased, or had no change for all animals (n = 7). e, Violin plot of the SFC fold change relative to baseline for all neurons. A value of 1 represents a doubling of the SFC. f, Example single neuron and all neuron SFC plot for one animal. The grey box highlights 0.3–4 Hz band. Error bars are SEM. g, Mean SFC plot for all animal including all neurons (n = 214, *p < 0.001). Follows convention from f

Fig. 4

Comparison of Broad and Narrow-Width Spiking Units. a, Example animal’s distribution of neurons classified by spike widths (n = 46). The color coding is based on k-means clustering. b, Differences in the spiking activity and SFC for narrow-width (left blue column) and broad-width (right red column) (*p < 0.001)

Fig. 5

LFP Power Before and After SES. Shows the power spectrum of the LFP prior to and after SES. There was no significant relationship observed

Fig. 6

Comparison of Changes in Firing Rate versus SFC. Plot shows correlation of single neuron changes in firing rate versus the corresponding SFC change. There was not a significant relationship between the two (r = 0.13, p > 0.05). Line was generated using linear regression