Directed physiological networks in the human prefrontal cortex at rest and post transcranial photobiomodulation

Abstract

Cerebral infra-slow oscillation (ISO) is a source of vasomotion in endogenic (E; 0.005-0.02 Hz), neurogenic (N; 0.02-0.04 Hz), and myogenic (M; 0.04-0.2 Hz) frequency bands. In this study, we quantified changes in prefrontal concentrations of oxygenated hemoglobin (Δ[HbO]) and redox-state cytochrome c oxidase (Δ[CCO]) as hemodynamic and metabolic activity metrics, and electroencephalogram (EEG) powers as electrophysiological activity, using concurrent measurements of 2-channel broadband near-infrared spectroscopy and EEG on the forehead of 22 healthy participants at rest. After preprocessing, the multi-modality signals were analyzed using generalized partial directed coherence to construct unilateral neurophysiological networks among the three neurophysiological metrics (with simplified symbols of HbO, CCO, and EEG) in each E/N/M frequency band. The links in these networks represent neurovascular, neurometabolic, and metabolicvascular coupling (NVC, NMC, and MVC). The results illustrate that the demand for oxygen by neuronal activity and metabolism (EEG and CCO) drives the hemodynamic supply (HbO) in all E/N/M bands in the resting prefrontal cortex. Furthermore, to investigate the effect of transcranial photobiomodulation (tPBM), we performed a sham-controlled study by delivering an 800-nm laser beam to the left and right prefrontal cortex of the same participants. After performing the same data processing and statistical analysis, we obtained novel and important findings: tPBM delivered on either side of the prefrontal cortex triggered the alteration or reversal of directed network couplings among the three neurophysiological entities (i.e., HbO, CCO, and EEG frequency-specific powers) in the physiological network in the E and N bands, demonstrating that during the post-tPBM period, both metabolism and hemodynamic supply drive electrophysiological activity in directed network coupling of the prefrontal cortex (PFC). Overall, this study revealed that tPBM facilitates significant modulation of the directionality of neurophysiological networks in electrophysiological, metabolic, and hemodynamic activities.

Keywords: Generalized partial directed coherence; Infra-slow oscillation; Neurometabolic coupling; Neurovascular coupling; Transcranial photobiomodulation.

Figure 1

(a) Experiment setup including two channels of bbNIRS on the lateral forehead and EEG cap. (b) The experimental protocol of this study; it consists of 5 visits with 7-min eyes-closed pre-stimulation at rest, followed by 8 min of randomized tPBM or sham stimulation on the left forehead (L800), right forehead (R800), left sham (LS), or right sham (RS) in four of the five visits, and 7-min post-stimulation. (c) Schematic diagram illustrating an EEG and 2-bbNIRS setup; the latter consists of a light source for 2 channels of bbNIRS, two separate spectrometers, and optical connections. The 2-bbNIRS and EEG data were concurrently collected pre- and post-stimulation. (d) Locations of the tPBM/sham stimulation delivered on the left (red circle) or right (blue circle) forehead. The bbNIRS holder and EEG channel Fp1 or Fp2 were removed during the left or right tPBM stimulation.

Figure 2

Data processing flow chart, including five steps for (1) EEG data analysis (blue boxes on the left of the figure), (2) 2-bbNIRS data analysis (orange boxes on the right of the figure), (3) construction of neurophysiological network and quantification of unilateral coupling using generalized partial directed coherence (GPDC) (green boxes), (4) quantification of differences in neurophysiological network between pre- and post-tPBM of respective GPDC values (as marked by the yellow box near the bottom of the figure), and (5) statistical analysis utilizing the Wilcoxon signed-rank test with FDR correction to assess the directionality in the neurophysiological network at rest and sham-controlled effects of tPBM (grey boxes on the bottom of the figure). All five steps were repeated for left and right forehead calculations. The dashed box outlines the repeated data-processing operation for the data collected before (i.e., at rest) and after tPBM/sham, while the dashed-dotted box outlines the repeated operation for each human participant.

Figure 3

Schematic illustration of the process for down-sampling EEG power time series. (a) EEG time series segmented in 1.5-s epochs. (b) power spectral density (PSD) obtained from each epoch. As an example, the area under the curve in the beta band is calculated and used to construct (c) an EEG beta-power time series with a time resolution of 1.5 s, which is matched to that of 2-bbNIRS.

Figure 4

Adjacency matrices and graphical illustration of resting-state neurophysiological networks on the prefrontal cortex obtained from dual-mode 2-bbNIRS and EEG dataset. Three columns represent (a) endogenic, (b) neurogenic, and (c) myogenic components of ISO over left and right PFC. The nodes in each network are HbO, CCO, and EEG beta power. (n = 22; data were averaged over 5 repeated measurements).

Figure 5

Directionality assessment in the neurophysiological network links, grouped by coherence with EEG (a,b) alpha power, (c,d) beta power, and (e,f) gamma power. The top and bottom rows correspond to the network links of the left and right PFC at rest. NVC: neurovascular coupling, NMC: neurometabolic coupling, MVC: metabolicvascular coupling. *p < 0.05, **p < 0.01, ***p < 0.001 obtained from the Wilcoxon signed-rank test after FDR correction (n = 22; data were averaged over 5 repeated measurements). Red circles mark highly significant leading roles of EEG and CCO with respect to HbO in the M band, consistently over both lateral PFC and across all three EEG powers. Blue rectangles make significant leading roles of CCO compared to HbO in the E band, consistently over both PFCs for EEG alpha and beta powers. Three triangles mark significant leading roles of EEG compared to HbO only in the EEG beta power case. Two downward arrows mark significant leading roles of EEG beta and gamma activities compared to CCO in the N band of the left and right PFCs, respectively.

Figure 6

Adjacency matrices and graphical illustration of changes (i.e., ΔGPDC_ss) in unilateral coupling in neurophysiological networks on the prefrontal cortex in response to L800 tPBM, as an example. Note that EEG beta power was considered in this case. Two columns represent (a) endogenic and (b) neurogenic components of ISO over the ipsi and contra PFC. The nodes in the network are HbO, CCO, and EEG beta power. Drawn links are the ΔGPDC_j,i,ss values, for which the Wilcoxon signed-rank test between ΔGPDC_j,i,tPBM and ΔGPDC_j,i,sham had to pass a p-value threshold of less than 0.05 (FDR corrected for multiple comparisons) (n = 22).

Figure 7

Adjacency matrices and graphical illustration of tPBM-induced significant changes in unilateral coupling in neurophysiological networks on the PFC in response to R800 tPBM (left column) and L800 (right column). Drawn links are the ΔGPDC_j,i,ss values (scaled by the color bars), for which the Wilcoxon signed-rank test between ΔGPDC_j,i,tPBM and ΔGPDC_j,i,sham passed p < 0.05 (with FDR correction). Note that labels of “alpha”, “beta”, and “gamma” in three respective rows denote EEG alpha, beta, and gamma powers used for GPDC calculations. The nodes in each network are HbO, CCO, and EEG power at a selected band.