Fig. 2: Population coupling and spike-EEG coupling are stronger for narrow-spiking neurons and oscillate in alpha-frequency band.

a–c Percentage-change of spike-triggered population firing rate (mean ± STD) as a function of time-to-spike. The percentage-change is computed relative to the corresponding baseline value at t = −500 ms in PPC, VC, and LP. The spike-triggered population firing rate functions have larger amplitude at spike instant (time lag zero) for narrow-spiking units (red) compared to broad-spiking units (blue) in both cortical regions (two-sample t-test between broad and narrow-spiking units at each time point at two significant levels indicated by horizontal green lines). d–f The power spectrum (mean ± SEM) of the spike-triggered population firing rate in PPC, VC, and LP. A prominent peak in the alpha (11–17 Hz) frequency band is evident for all three regions. The alpha power is higher for narrow-spiking units (red) when compared to broad-spiking units (blue) in both cortical regions (two-sample t-test at each frequency point at two significant levels indicated by horizontal green lines). g–o Alpha synchronization between EEG and spikes, and prominent engagement of narrow-spiking neurons. g–i Spike-triggered EEG (mean ± STD) as a function of the time-to-spike in PPC, VC, and LP. Spike-triggered average EEG shows an oscillatory pattern for all three regions and shows larger amplitudes at peaks and troughs for narrow-spiking (red) when compared to broad-spiking (blue) neurons (two-sample t-test at each time point at two significant levels indicated by horizontal green lines). j–l The histogram of phase preference of spiking units in PPC, VC, and LP as a function of the instantaneous phase of the alpha-band (12–17 Hz) EEG oscillation. Color denotes the counts for narrow- and broad-spiking neurons. Two classes of neurons show similar phase preference in PPC (Broad-spiking neurons: 3.02 ± 0.02 rad, Narrow-spiking neurons: 3.05 ± 0.02 rad), and slightly different phase preference in VC (broad-spiking neurons: 1.91 ± 0.03 rad, narrow-spiking neurons: 1.64 ± 0.04 rad). m–o Phase-locking between single units and EEG measured by PLV (mean ± SEM) as a function of EEG frequency. Synchronization was increased across a broad range from 3 to 32 Hz range with a prominent peak in the alpha (12–17 Hz) frequency range for recorded cortical and thalamic sites. The narrow-spiking cells (red) show significantly higher PLV compared to broad-spiking units (blue) (two-sample t-test at each frequency point with p-values < 0.05 indicated by horizontal line). p–r Population coupling of single neurons and their engagement with globally measured alpha correlate strongly in PPC but not in VC and LP. Significant correlations were tested by a two-sided Pearson correlation test. No multiple comparison adjustment for p-value was performed. p Population coupling of single units versus their engagement by alpha-band EEG is plotted for PPC. There is a strong correlation between population coupling and large-scale alpha synchronization for narrow-spiking (red, n = 403, Pearson correlation test, r = 0.71, 2-sided p = 1.3*e⁻⁶¹, 95% CI [0.66, 0.75]) and broad-spiking units (blue, n = 582, Pearson correlation test, r = 0.86, 2-sided p = 3.46*e⁻¹⁶⁸, 95% CI [0.84, 0.88]) in PPC. Color denotes the neuron type. q Population coupling of single units versus their engagement by alpha-band EEG is plotted for VC. Population correlation and large-scale alpha synchrony do not correlate in VC (narrow-spiking: red, n = 173, Pearson correlation test, r = 0.06, 2-sided p = 0.46, 95% CI [−0.09, 0.21], and broad-spiking: blue, n = 221, Pearson correlation test, r = 0.01, p = 0.88, 95% CI [−0.12, 0.14]). r Population coupling of single units versus their engagement by alpha-band EEG is plotted for LP. This analysis revealed that population coupling and long-range synchrony are weakly correlated in LP (n = 251, Pearson correlation test, r = 0.35, 2-sided p = 3.5*e⁻⁷, 95% CI [0.22, 0.46]). The values on the vertical axis for all figures are the percentage-change in spike-triggered population firing rate at the time of spike generation relative to the corresponding baseline value at t = −500 ms.