Enhancing memory capacity by experimentally slowing theta frequency oscillations using combined EEG-tACS

Abstract

The coupling of gamma oscillation (~ 40+ Hz) amplitude to the phase of ongoing theta (~ 6 Hz) oscillations has been proposed to be directly relevant for memory performance. Current theories suggest that memory capacity scales with number of gamma cycles that can be fitted into the preferred phase of a theta cycle. Following this logic, transcranial alternating current stimulation (tACS) may be used to adjust theta cycles (increasing/decreasing theta frequency) to decrease or increase memory performance during stimulation. Here, we used individualized EEG-informed theta tACS to (1) experimentally "slow down" individual theta frequency (ITF), (2) evaluate cognitive after effects on a battery of memory and learning tasks, and (3) link the cognitive performance changes to tACS-induced effects on theta-band oscillations as measured by post EEG. We found frequency- and task-specific tACS after effects demonstrating a specific enhancement in memory capacity. This tACS-induced cognitive enhancement was specific to the visual memory task performed immediately after tACS offset, and specific to the ITF-1 Hz (slowing) stimulation condition and thus following a protocol specifically designed to slow down theta frequency to enhance memory capacity. Follow-up correlation analyses in this group linked the enhanced memory performance to increased left frontal-parietal theta-band connectivity. Interestingly, resting-state theta power immediately after tACS offset revealed a theta power increase not for the ITF-1 Hz group, but only for the ITF group where the tACS frequency was 'optimal' for entrainment. These results suggest that while individually calibrated tACS at peak frequency maximally modulates resting-state oscillatory power, tACS stimulation slightly below this optimal peak theta frequency is better suited to enhance memory capacity performance. Importantly, our results further suggest that such cognitive enhancement effects can last beyond the period of stimulation and are linked to increased network connectivity, opening the door towards more clinical and applied relevance of using tACS in cognitive rehabilitation and/or neurocognitive enhancement.

Figure 1

The significant behavioral results. (a) The mean values of the VM scores at pre- and post-tACS across the groups. The time*group interaction was significant (p = 0.016). The ITF-1 group had an increased VM score after the tACS. (b) The mean values of the Learning scores at pre- and post-tACS across the groups. The time*group interaction was significant (p = 0.011). ITF-1 group had an increased learning score after the tACS. See Supplementary Fig. 1 for additional representation. VM, visual memory; tACS, transcranial alternating current stimulation; ITF, individual theta frequency. The vertical bars denote 0.95 confidence intervals. Dots represent the observed scores. Asterisks indicate line reflects a significant change over time.

Figure 2

The significant results of the after effect of the tACS on EEG data. (a) The time*group interaction was significant (p = 0.016): ITF group had increased resting theta power after tACS while there were no tACS effects on the other groups. See Supplementary Fig. 2 for additional representation. (b) The time*group interaction was significant (p = 0.01) for VM task. The tACS groups had decreased event-related theta power after tACS while no such difference was observed in the sham group. See Supplementary Fig. 3 for additional representation. (c) The grand average figures of event-related power analysis (4–7 Hz) in time–frequency domain in response to items in the VM task. The left frontocentral area was presented for each group in the figure. tACS groups had decreased event-related theta power after theta-frequency tACS (for ITF-1 p = 0.045, for ITF P = 0.011), not observed in the sham group (p = 0.145). The X-axis represents time, and the Y-axis represents frequency; the point at which the stimulus arrives is marked as a zero point on the X-axis. tACS, transcranial alternating current stimulation; ITF, individual theta frequency. The vertical bars denote 0.95 confidence intervals. Dots represent the observed scores. Asterisks indicate line reflects a significant change over time.

Figure 3

The scatter plots of the significant results in the EEG-behavior correlation analyses. (a) The subjects in the ITF-1 group with increased left frontal-parietal resting theta coherence after the tACS had the higher scores in the VM task (r = 0.544, p = 0.036). (b) The subjects in the ITF-1 group with decreased left frontal-parietal event-related power-based theta connectivity after the tACS had the higher learning scores (r = − 0.614, p = 0.011). (c) The subjects in the ITF-1 group with decreased left frontal-parietal PLV after the tACS had the higher learning scores (r = − 0.539, p = 0.031). Other groups' scatter plots were presented in the Supplementary Fig. 4. Coh., coherence; Diff., difference score; PLV, phase locking value; VM, visual memory; ITF, individual theta frequency. The shaded area denotes the standard error. Dots represent the observed scores.

Figure 4

Experimental design and task procedure. (a) Design of the experiment. The “part” states each experimental phase. The horizontal dark blue arrow shows the time-course and application order of the parts with durations of them in minutes. The vertical grey arrows show the application order of the numbered measurements stated in the part. (b) The representations of the applied visual* (on the left) and auditory (on the right) memory tasks during the EEG recordings. * Given schematic images in Fig. 4b are not actual exemplars from the Boston naming test set; drawn by TA as an example. tACS, transcranial alternating current stimulation; ITF, individual theta frequency; min, minute; sec., second.