Sleep modulates cortical connectivity and excitability in humans: Direct evidence from neural activity induced by single-pulse electrical stimulation

Abstract

Sleep-induced changes in human brain connectivity/excitability and their physiologic basis remain unclear, especially in the frontal lobe. We investigated sleep-induced connectivity and excitability changes in 11 patients who underwent chronic implantation of subdural electrodes for epilepsy surgery. Single-pulse electrical stimuli were directly injected to a part of the cortices, and cortico-cortical evoked potentials (CCEPs) and CCEP-related high-gamma activities (HGA: 100-200 Hz) were recorded from adjacent and remote cortices as proxies of effective connectivity and induced neuronal activity, respectively. HGA power during the initial CCEP component (N1) correlated with the N1 size itself across all states investigated. The degree of cortical connectivity and excitability changed during sleep depending on sleep stage, approximately showing dichotomy of awake vs. non-rapid eye movement (REM) [NREM] sleep. On the other hand, REM sleep partly had properties of both awake and NREM sleep, placing itself in the intermediate state between them. Compared with the awake state, single-pulse stimulation especially during NREM sleep induced increased connectivity (N1 size) and neuronal excitability (HGA increase at N1), which was immediately followed by intense inhibition (HGA decrease). The HGA decrease was temporally followed by the N2 peak (the second CCEP component), and then by HGA re-increase during sleep across all lobes. This HGA rebound or re-increase of neuronal synchrony was largest in the frontal lobe compared with the other lobes. These properties of sleep-induced changes of the cortex may be related to unconsciousness during sleep and frequent nocturnal seizures in frontal lobe epilepsy.

Keywords: CCEP; consciousness; frontal lobe epilepsy; high-gamma activity; non-REM; sleep.

Figure 1

CCEP measurement and analysis. A: Subdural electrodes of a representative case (Pt. 2L) are non‐linearly co‐registered onto the MNI standard space (ICMB‐152). B: In the daytime, a 1‐Hz SPES (alternating polarity, 4–12 mA) was applied to several pairs of electrodes to map connectivity. C: ECoGs were off‐line averaged time‐locked to the stimulus (analysis window: −300 to +700 ms). In each patient, one or two stimulus sites away from the epileptic focus were selected for the sleep study, based on the presence of discrete CCEP responses in adjacent and/or distant regions (awake: W, Stage I–II: L, Stage III–IV: S, stage REM [rapid eye movement]: R). Only the response sites showing the most discrete N1 (first sharp component) and N2 (second slow component) responses, i.e., the maximum site of the adjacent and remote CCEP fields, were chosen for analysis sites. D: Five indices were used for the analysis. The troughs or positive sharp reflections before and after the N1 peak (31 ms in this representative waveform) were termed P1 (12 ms) and P2 (64 ms), respectively. Each index is highlighted by a rectangle: CCEP sizes (area under the curve) of N1 and N2 (198 ms) (N1_Area and N2_Area); their high‐gamma activities (HGA) counterparts (N1_HG [30 ms], N2_{HG_min} [85 ms], and N2_{HG_max} [480 ms]).

Figure 2

Dynamic modulation of CCEP indices by sleep. A: Stimulus (midpoint of the electrode pair: ○) and analysis (⋄) sites from all patients were coregistered onto the MNI standard space. Those in the right hemisphere were flipped to the left hemisphere for display purposes. Each stimulus site and corresponding analysis sites are shown in the same color. B: Representative CCEP waveform and CCEP‐related HGA across sleep stages (Pt. 2L, recorded from the frontal pole [FP]). C: Modulation of CCEP indices during sleep (N = 45 for W, L, and S, N = 43 for R). Relative values of each index (sleep [L, S, and R] in reference to W) are indicated by box plots (a central line in the box indicates the median of the data and the lower and upper boundary of the box are at the 25% and 75% quantiles of the data). The vertical lines (whiskers) extend to the most extreme data value within 1.5 times the height of the central box. Those outside this range were considered outliers (+). ^†Recorded from the contralateral hemisphere. * Statistically significant at P < 0.05, Wilcoxon signed‐rank test, corrected by false discovery rate (FDR) = 0.05. Note many indices began to show different values once the patients fell into non‐REM sleep. Except for N2_{HG_max}, all indices tended to return to the value close to W when the sleep stage reached R, although CCEP‐related HGA (N1_HG, N2_{HG_min}) remained significantly different.

Figure 3

Partial correlation between CCEP indices during sleep. A: Scattergrams for selected arbitrary pairs are shown across sleep stages (see Supporting Information Fig. 8 for those of all pairs). Unit for N1_Area and N2_Area is μV·s, and that for and N1_HG, N2_{HG_min}, and N2_{HG_max} is 10·dB. *Statistically significant at P < 0.05, corrected by FDR = 0.05. N1_HG has significant partial correlation with N1_Area throughout all stages, and shows a tendency of partial anticorrelation with N2_{HG_min} during sleep. In slow‐wave sleep, N2_Area shows partial anticorrelation with N2_{HG_min}. Regression line is given as reference, but the value of slope is not the same as the value of ρ in each figure because “partial” correlation was calculated, not correlation. B: Potentiated decrease of N2_{HG_min} during sleep. In the scattergrams of N1_HG and N2_{HG_min} (see Fig. 3A lower left), the vectors from the point during the awake state to the point of the other sleep stages were calculated for each analysis site [W →L (N = 45), W →S (N = 45), and W →R (N = 43)] (gray arrows, black arrow denotes the synthetic vector). Note that the direction of the vectors is significantly downward from the regression line of the awake stage in each figure, and that the synthetic vector is the largest and the most downward in W →S and the least in W →R. *Statistically significant at P < 0.05, sign test, corrected by false discovery rate (FDR) = 0.05.

Figure 4

Modulation of intralobar connectivity and excitability during sleep. A: The modulation mode was compared between the frontal lobe (stimulus and analysis sites in the frontal lobe: N = 11) and other lobes ( N = 13). B: Representative CCEP waveform and its HGA counterpart in the frontal (superior frontal gyrus [SFG] of Pt. 8) and parietal (the supramarginal gyrus [SMG] of Pt. 6) lobes. C: Values relative to W were calculated in the sleep stages (L, S, R) for each index, and these values were compared between the Frontal lobe and the non‐Frontal lobes. In each sleep stage, the left column denotes values of the Frontal sites and the right column denotes values of the non‐Frontal sites (N = 11 in L, S, R for Frontal sites; and N =13 in L, S and N =11 in R for non‐Frontal sites). *Statistically significant at P < 0.05, Mann–Whitney U test, corrected by FDR = 0.05. Note the significant difference in N2_{HG_max} (L, S, R) between Frontal and non‐frontal sites. The direction of the change was opposite: increase in frontal sites and decrease in non‐Frontal sites. The other conventions including the box plots are the same as for Figure 2.

Figure 5

Temporal relationship among latencies related with N2–N2 peak, N2_{HG_min}, and N2_{HG_max}. A: The indices for which latencies were analyzed are highlighted by a rectangle. B: Comparison of the latencies (N2 peak vs. N2_{HG_min}, and N2 peak vs. N2_{HG_max}) in each stage. *Statistically significant at P < 0.05, Wilcoxon signed‐rank test, corrected by FDR = 0.05. Note the temporal relationship (N2_{HG_min} →N2 peak →N2_{HG_max}) became consistent once the patients fell asleep.