Transcranial Direct Current Stimulation Over Bilateral Temporal Lobes Modulates Hippocampal-Occipital Functional Connectivity and Visual Short-Term Memory Precision

Abstract

Although the medial temporal lobe (MTL) is traditionally considered a region dedicated to long-term memory, recent neuroimaging and intracranial recording evidence suggests that the MTL also contributes to certain aspects of visual short-term memory (VSTM), such as the quality or precision of retained VSTM content. This study aims to further investigate the MTL's role in VSTM precision through the application of transcranial direct current stimulation (tDCS) and functional magnetic resonance imaging (fMRI). Participants underwent 1.5 mA offline tDCS over bilateral temporal lobes using left cathodal and right anodal electrodes, administered for either 20 min (active) or 0.5 min within a 20-min window (sham), in a counterbalanced design. As the electrical current passes through midbrain structures with this bilateral stimulation montage, prior behavioral and modeling evidence suggests that this tDCS protocol can modulate MTL functions. To confirm this and examine its impacts on VSTM, participants completed a VSTM color recall task immediately following tDCS, while undergoing a 20-min fMRI scan and a subsequent 7.5-min resting-state scan, during which they focused on a fixation cross. Behavioral results indicated that this tDCS protocol decreased VSTM precision without significantly affecting overall recall success. Furthermore, psychophysiological interaction analysis revealed that tDCS over the temporal lobe modulated hippocampal-occipital functional connectivity during the VSTM task, despite no main effect on fMRI BOLD activity. Notably, this modulation was also observed during resting-state fMRI 15-20 min post-tDCS, with the magnitude of the effect correlating with participants' behavioral changes in VSTM precision across active and control conditions. Combined, these findings suggest that tDCS over the temporal lobe can modulate the intrinsic functional connectivity between the MTL and visual sensory areas, thereby affecting VSTM precision.

Keywords: fMRI; medial temporal lobe; pattern separation; precision; tDCS; visual short‐term memory.

FIGURE 1

Stimulation protocol, behavioral task, and testing procedure. (A) Participants received a left cathode right anode tDCS over the bilateral temporal lobes with 1.5 mA current. This stimulation protocol, based on a simulated model (Cappiello et al. 2016), can penetrate and lead to current flow in the medial part of the temporal lobe regions. (B) The VSTM color recall task used during fMRI scanning and behavioral follow‐up sessions. (C) The testing procedure and research events in the current study.

FIGURE 2

Behavioral results across experimental conditions. (A) Individual data (gray lines) and overall model fit based on data aggregated across participants (light green lines). Despite similar tails of the recall error distributions across experimental conditions, the width of the central peak (light red lines) is larger under active stimulation relative to sham or behavioral‐only follow‐up sessions. Fitting individual participants' data using the mixture model confirmed these observations, (B) in that recall variability (SD) is largest under the active stimulation condition, (C) whereas the probability of recall success (Pm) is more or less the same across conditions. The model parameters derived from aggregated data are shown alongside the group fit in (A). These values are similiar but not identical to the average parameter values based on inidviudal best‐fit parameters. *p < 0.05; n.s. = not significant.

FIGURE 3

tDCS over the bilateral temporal lobes modulates hippocampal‐neocortical functional connectivity during the VSTM task and in the subsequent resting period. (A) When using the left anterior hippocampus as the seed, active stimulation tends to increase functional coupling in fMRI BOLD signals between the seed and distributed areas in the brain, both during the VSTM task (e.g., left ATL, left prefrontal cortex, bilateral precuneus, and right visual cortex) and during the subsequent resting period 15 min afterward (e.g., bilateral visual cortices). (B) When using the left posterior hippocampus as the seed, the effect of tDCS on functional connectivity is attenuated during the VSTM task. However, the functional connectivity between the seed and bilateral visual cortices remains similar to that observed when using the left anterior hippocampus as the seed.

FIGURE 4

The magnitude of tDCS effect on intrinsic hippocampus‐occipital functional connectivity predicts behavioral changes in VSTM precision in our current study. (A) We identified the intrinsic functional connectivity between the seed region, left aHPC (yellow), and the target region, right visual cortical ROI (pink), based on the correlation of the average residual BOLD signals during resting‐state fMRI. Through repeated‐measures correlation r_rm, we find that the magnitude of tDCS effect on intrinsic hippocampus‐occipital functional connectivity during the resting state predicts overall changes in VSTM precision between active stimulation and control conditions. This holds true whether the right visual cortical ROI is defined by the significant voxels only in resting‐state analysis (B) or by the overlapping voxels identified by the resting‐state analysis and an independent PPI analysis (C). Notably, this effect is not observed for the overall recall likelihood of the VSTM task.