Inhibitory interneuron circuits at cortical and spinal levels are associated with individual differences in corticomuscular coherence during isometric voluntary contraction

Abstract

Corticomuscular coherence (CMC) is an oscillatory synchronization of 15-35 Hz (β-band) between electroencephalogram (EEG) of the sensorimotor cortex and electromyogram of contracting muscles. Although we reported that the magnitude of CMC varies among individuals, the physiological mechanisms underlying this variation are still unclear. Here, we aimed to investigate the associations between CMC and intracortical inhibition (ICI) in the primary motor cortex (M1)/recurrent inhibition (RI) in the spinal cord, which probably affect oscillatory neural activities. Firstly, we quantified ICI from changes in motor-evoked potentials induced by paired-pulse transcranial magnetic stimulation in M1 during tonic isometric voluntary contraction of the first dorsal interosseous. ICI showed a significant, negative correlation with the strength of EEG β-oscillation, but not with the magnitude of CMC across individuals. Next, we quantified RI from changes in H-reflexes induced by paired-pulse electrical nerve stimulation to the posterior tibial nerve during isometric contraction of the soleus muscle. We observed a significant, positive correlation between RI and peak CMC across individuals. These results suggest that the local inhibitory interneuron networks in cortical and spinal levels are associated with the oscillatory activity in corticospinal loop.

Figure 1. Representative examples of EEG/EMG data and motor-evoked potentials (MEPs) for a participant who showed significant CMC (CMC+) and a participant who did not (CMC−).

Raw EEG signals, raw EMG signals, power spectral density functions (PSDs) for EEG and rectified EMG signals, corticomuscular coherence (CMC) spectra during isometric contraction of the first dorsal interosseous (FDI), and MEPs elicited by single-pulse and paired-pulse transcranial magnetic stimulation (TMS) are shown. In the CMC spectra, the estimated significance levels of coherence (SLs, 0.030) are shown as horizontal dotted lines. In MEPs, grey lines show the individual EMG responses of 30 trials, and black lines show averaged waves.

Figure 2

(A) The relationship between intracortical inhibition (ICI) and the peak value of CMC during isometric contraction of FDI across all participants. A horizontal dashed line shows the estimated SL (=0.030). There was no significant correlation between CMC and ICI (p = 0.197). (B) The relationship between ICI and EEG PSD in β-band across all participants. EEG PSD in β-band indicates the ratio of the sum of EEG-PSD within the β-band (15–30 Hz) to that of 3–50 Hz frequency range. There was a significant correlation between them (r = −0.559, p = 0.037). The solid line shows the estimated regression line.

Figure 3. Representative examples of EEG/EMG data and H-reflexes for CMC+ and CMC−.

Raw EEG signals, raw EMG signals, power spectral density functions (PSDs) for EEG and rectified EMG signals, CMC spectra during isometric contraction of the soleus (SOL), and H-reflexes elicited by single (H) and paired-pulse (H′) electrical nerve stimulation are shown. Data for CMC+ and CMC− are shown, respectively. In the CMC spectra, horizontal dashed lines show the estimated SL (=0.091). In H-reflexes, grey lines show individual EMG responses of 20 trials, and black lines show averaged waves.

Figure 4

(A) The relationship between recurrent inhibition (RI) and the peak value of CMC during the isometric contraction of SOL across all participants. A horizontal dashed line shows the estimated SL (=0.091). A significant correlation was observed between RI and CMC (r = 0.663, p = 0.009). The solid line shows the estimated regression line. (B) The relationship between RI and EEG PSD in β-band across all participants. EEG PSD in β-band indicates the ratio of the sum of EEG-PSD within the β-band (15–30 Hz) to that of 3–50 Hz frequency range. There was no significant correlation between them (p = 0.270).

Figure 5. The relationship between peak values of CMC recorded from FDI and SOL.

The horizontal and vertical dashed lines show the estimated SL for FDI (=0.030) and for SOL (=0.091), respectively. There was a significant positive correlation between them (r = 0.882, p = 0.004). The solid line shows the estimated regression line.

Figure 6. Model for the neural system associated with the development of CMC during voluntary isometric contraction.

The present two main findings suggest the roles of ICI and RI for β-oscillation in the corticospinal pathway. At first, the negative feedback loop between pyramidal neurons (PN) and ICI generates β-oscillation. Next, RI of Renshaw cells (RCs) desynchronizes α-motoneurons (αMNs) firing and attenuates the amplitude of β-oscillation derived from the motor cortex.