This research studied the connectivity strength changes after tACS at 5-Hz and 2.0 mA for 25 minutes, with a tripod montage covering the frontal, parietal, and temporal regions. The design was proposed as a way to ease anxiety (Lee et al., 2024). Pre- and post-treatment EEGs were collected and analyzed for 24 participants using an imaging method eLORETA, which converted the signals recorded at the scalp to CSD in the brain cortex. Average functional connectivity was thus derived from the CSD time series between the frontal, parietal, and temporal regions. After tACS treatment, the lagged phase synchronization at theta range (i.e., spectrum-specific) significantly increased in the frontal–parietal and the parietal–temporal connections. The lagged coherence was enhanced between the frontal and parietal regions with a P value lower than 0.05, although not surviving Bonferroni correction. Our hypothesis was generally verified but with some caveats, discussed below.
Strong evidence suggested that tACS may influence the timing of neuronal spikes (Krause et al., 2019; Anastassiou et al., 2010; Radman et al., 2007). When tACS is delivered to two distinct brain regions, the firing of neural tissues in these regions is expected to synchronize more closely by the applied frequency. Furthermore, given that the recorded EEGs were not in real-time with the delivered tACS, the significant inter-regional interactions indicated the engagement of a plasticity mechanism. Interestingly, the index derived from lagged phase synchronization was more robust than that of lagged coherence. Our previous study investigating the regional power changes to tACS demonstrated complicated spectral power features incompatible with entrainment theory (Lee and Tramontano, 2024). From a mathematical perspective, the formula of phase synchronization is similar to that of coherence, except for a normalization procedure to discount the influence of the power. Based on the two analytic results (power and connectivity), it was deduced that the inter-regional modulatory effects of tACS were not due to the concurrent entrained oscillation/power at multiple regions. Rather, it was through synchronizing their phase relationship, echoing previous neurophysiological research that recorded and investigated spike timing under tACS (Krause et al., 2019). Compatible with our conjecture, Vossen et al. explored the aftereffects of alpha tACS and concluded that the modulatory effect of tACS was mediated by plasticity rather than entrainment (Vossen et al., 2015). In summary, at the large-scale network level, the inter-regional influence of tACS was mediated by synchronization in phase (crosstalk), not by the concurrent entrainment of powers (regional profile).
It was noticed that the connectivity changes were not significant between the frontal and temporal regions. Again, if the influence of tACS on neural connectivity mainly worked through concurrent entrainment across targeted areas, the interactions between the three explored regions would tighten altogether. We inferred that the differential manifestations originated from the discrepancy in the hardwire underpinnings. It was noted that the superior longitudinal fasciculus (SLF) bridges between the frontal and parietal regions and between the inferior parietal cortex (BA39/40) and the middle temporal cortex (BA 21) (SLF III). The former constitutes the frontoparietal network, and the latter links the two cortical nodes of the default-mode network (Lee and Xue, 2018). No white matter “highway” exists between the dorsolateral prefrontal and middle temporal cortices (note: the inferior frontal cortex and anterior temporal region are connected by uncinate fasciculus). It hints at one of the most paramount plasticity mechanisms, spike-time-dependent plasticity, which requires direct axonal connections to take effect (Levy and Steward, 1983; Markram et al., 2012).
It was observed that applying a particular frequency of tACS can “entrain” or “synchronize” the neural oscillations to match the frequency of the electrical stimulation (Helfrich et al., 2014; Voss et al., 2014), framed as an entrainment theory. However, our recent report demonstrated that narrow band 5-Hz tACS desynchronized neural oscillation (offline measurement and comparison), which affected broad spectra beyond the default frequency of tACS (Lee and Tramontano, 2024). An earlier study by Brignani et al. challenged the idea that tACS effectively modulated brain oscillations (Brignani et al., 2013). Alexander et al. showed that 10-Hz tACS, in fact, reduced alpha power in the frontal region (Alexander et al., 2019). In addition, Lafton et al. applied the intracranial recording and observed no sleep rhythm entrainment to tACS (Lafon et al., 2017). The contradictory findings cannot be resolved by entrainment theory alone but require a broader mechanism to reconcile them. Agreeing with Vossen et al. (Vossen et al., 2015), we regard that neural plasticity could be a better candidate to accommodate the offline (in contrast to real-time) tACS influence on regional powers and inter-regional interactions. Nevertheless, we cannot exclude the possibility that injecting an artificial narrow-band alternating current may interfere with underlying neural synchronization under certain conditions, given our previous analysis and several other reports summarized above (Lee and Tramontano, 2024). It is noteworthy that even if tACS impedes the underlying neural synchronization, it may still be beneficial. In our previous report, power reduction in the right hemisphere due to tACS might reduce emotion reactivity according to emotion lateralization theory and hence, might catalyze the anxiolytic effect (Lee and Tramontano, 2024; Ross, 2021). The merits and demerits of tACS thus could be context-dependent, which requires further research to clarify.