Physiology-inspired bifocal fronto-parietal tACS for working memory enhancement

Abstract

Aging populations face significant cognitive challenges, particularly in working memory (WM). Transcranial alternating current stimulation (tACS) offer promising avenues for cognitive enhancement, especially when inspired by brain physiology. This study (NCT04986787) explores the effect of multifocal tACS on WM performance in healthy older adults, focusing on fronto-parietal network modulation. Individualized physiology-inspired tACS applied to the fronto-parietal network was investigated in two blinded cross-over experiments. The first experiment involved monofocal/bifocal theta-tACS to the fronto-parietal network, while in the second experiment cross-frequency theta-gamma interactions between these regions were explored. Participants have done online WM tasks under the stimulation conditions. Network connectivity was assessed via rs-fMRI and multichannel electroencephalography. Prefrontal monofocal theta tACS modestly improved WM accuracy over sham (d = 0.30). Fronto-parietal stimulation enhanced WM task processing speed, with the strongest effects for bifocal in-phase theta tACS (d = 0.41). Cross-frequency stimulations modestly boosted processing speed with or without impairing task accuracy depending on the stimulation protocol. This research adds to the understanding of physiology-inspired brain stimulation for cognitive enhancement in older subjects.

Keywords: Cognition; Electric field modelling; Healthy aging; Multifocal; Neuroimaging; Orchestrated brain stimulation; Systems neuroscience; Working memory; tACS.

Fig. 1

A. Design of the study. Prior to the stimulation sessions, participants underwent baseline multichannel scalp EEG during the task and task-fMRI measurements to personalize the stimulation parameters in terms of individual frequency and target location. Left: For the MUNI cohort, the stimulation protocols aimed to entrain oscillations within the theta range and synchronize fronto-parietal regions. Participants received monofocal stimulation in frontal or parietal areas and bifocal frontal and parietal stimulation in and out of phase. In the MUNI cohort only, resting-state fMRI (rs-fMRI) scans were acquired before and immediately after each stimulation session. The tACS and n-back tasks were conducted in the NIBS laboratory adjacent to the MRI scanner, with subjects easily moved between the two labs with a maximum of 5 min between the examinations. An alternating current was delivered using pairs of concentric rubber electrodes (Outer electrode: Ø 75 mm with a hole Ø 40 mm, inner electrode Ø 20 mm) Right: For the EPFL cohort, the stimulation protocols aimed to affect cross-frequency coupling between theta and gamma oscillations. Individualized gamma and theta cross-frequency tACS stimulation over the prefrontal and parietal cortex was applied. Continuous theta stimulation was applied while gamma stimulation was delivered either continuously or in bursts aligned to the theta peak. There was no before-and-after fMRI. B. Stimulation montage for bifocal stimulation. For monofocal stimulation, the electrode placement was the same, but only one current source was active. Image created with the use of SimNIBS software. Head is derived from a standard template provided. C. Stimulation protocols for inter-areal synchronization in experiment 1 (MUNI) D. Stimulation protocols for inter-areal cross-frequency stimulation in experiment 2 (EPFL).

Fig. 2

A. Schematic diagram of a study timeline. B. Main behavioral outcome – n-back task modified from Ref. [35]; task performed online during each stimulation condition. Participants completed two levels of difficulty during the stimulation. Participants indicate whether the current stimulus matches the one from n steps (2-back/3-back) earlier in the sequence. The task was performed also during fMRI to spatially individualize targets based on each participants activation.

Fig. 3

Normalized accuracy in the n-back tasks for the MUNI (experiment 1) and EPFL (experiment 2) cohorts. *sig. p < 0.05.

Fig. 4

Normalized speed in the n-back tasks for the MUNI (experiment 1) and EPFL (experiment 2) cohorts. *sig. p < 0.05.

Fig. 5

A Visualization of FPCN network seeds used for the rs-fMRI data analysis. raPFC-rdlPFC connection is highlighted by a line. B Connectivity between the raPFC-rdlPFC. Legend: l/raPFC = left/right anterior prefrontal cortex, ACC – anterior cingulate cortex, l/raIPL = left/right anterior intraparietal lobule, l/rdlPFC = left/right dorsolateral prefrontal cortex, l/rINS = left/right insula.

Fig. 6

Visualization of triple interactions among the bifocal stimulation conditions and the magnitude of the EFs induced at both stimulation sites (i.e., prefrontal and parietal cortices) acting on the speed in the 2-back and the 3-back task. The dots represent individual datapoints; the planes reflect the statistical model adjusted to these points. In the in-phase condition, the model describes steeper improvements in speed predicted for individuals exposed to EFs of higher magnitude. In contrast, the model describes speed decreases in individuals receiving out-of-phase stimulation who were predicted to have speed decreases when exposed to EFs of higher magnitude. This detrimental effect is only apparent in the 2-back and not the 3-back task. Note that the vertical axis has the same scale in all four panels. Even though there is no numerical grading along the axes, all plots have the same scale.

Fig. 7

Visualization of triple interactions among the stimulation and the magnitude of the EFs induced at both stimulation sites (i.e., prefrontal and parietal cortices) acting on the accuracy in the 3-back task. The dots represent individual datapoints; the planes reflect the statistical model adjusted to these points. Specifically, when applying tACS within the gamma band to the prefrontal cortex and theta-tACS to the parietal cortex (stimulation condition 1), the rate of improvement in accuracy within a training session became steeper as a function of the magnitude of the induced EF, and this effect was more pronounced when the magnitude was similar at both stimulation sites. In contrast, when applying theta to the prefrontal cortex and gamma tACS to the parietal cortex (stimulation condition 2), the rate of improvement was steepest when the EFs were lateralized, and least pronounced when they were both symmetric and of high intensity. These interactions seem to be more attenuated in both conditions including gamma bursts (i.e., stimulation conditions 3 and 4). Note that even though there is no numerical grading along the axes, all the vertical axes in all plots have the same scale.

Fig. 8

Visualization of the proposed misalignment between the externally induced fast gamma component and the inherent local theta rhythms in different stimulation conditions. Consequently, the theta-gamma coupling might be disproportionately skewed, which in turn would distort specific theta phase-locking present between prefrontal and distant posterior areas.