The Influence of Stimulation Timing on the Modulation of Working Memory

Affiliations:
Institute of Brain and Psychological Sciences, Sichuan Normal University, Chengdu 610066, China
Center for Cognitive and Psychoneurobiological Studies, University of Liège, Belgium
University Medical Center Göttingen, Georg-August University, Göttingen, Germany
Research Center for Brain and Cognitive Neuroscience, Liaoning Normal University, Dalian, China
Psychological Research and Counseling Center, Southwest Jiaotong University, Chengdu 611756, China

Ruiqiao Guo and Wenrui Li contributed equally to this work.

Abstract

Although transcranial Alternating Current Stimulation (tACS) can enhance working memory (WM) performance by modulating cognitive resources, existing research has reported inconsistent effects. This study identifies a critical moderating variable for these regulatory effects. Through two rigorously designed experiments, we demonstrate that the regulatory impact of tACS is influenced by the temporal alignment between stimulation and task performance. Compared to sham stimulation, theta-tACS significantly increases memory capacity while maintaining precision when applied during the task. However, once a quantity-precision trade-off strategy has been established in the task, this strategy persists under the influence of psychological inertia; specifically, compared to sham stimulation, tACS selectively improves memory precision without altering the quantity of items remembered. Furthermore, it was discovered that the influence on trade-off strategies does not transfer; that is, the strategy formed by a participant in the initial task is disrupted when encountering a new task. Faced with a novel scenario, individuals develop new strategies based on specific task characteristics and their total available resources, demonstrating the flexible reallocation of cognitive resources. Consequently, we conclude that the regulatory effect depends on the timing of the stimulus application: if applied prior to practice, it primarily affects memory quantity; if applied afterward, it mainly influences memory precision. However, the regulatory effects already established do not transfer to new tasks.

Keywords: theta-tACS, working memory, accuracy-quantity trade-off, psychological inertia

Introduction

Visual working memory (VWM) is a core component of cognitive function, responsible for the brief maintenance, manipulation, and storage of visual information. Research indicates that individuals with high visual working memory capacity exhibit significant cognitive advantages compared to those with low capacity. These differences are often reflected in the efficiency of information processing and the ability to manage complex cognitive tasks. Significant individual differences exist in VWM capacity; individuals with higher capacity demonstrate superior performance in tasks involving machine learning \cite{Zhang2019}, language understanding \cite{Ding2023}, and the suppression of irrelevant information \cite{Vogel2005}. Conversely, low VWM capacity can limit an individual's cognitive potential. Consequently, researchers are dedicated to exploring non-invasive methods to effectively enhance VWM capacity and improve overall cognitive function.

Transcranial Alternating Current Stimulation (tACS) at the theta frequency has been shown to significantly improve working memory performance by modulating neural oscillation synchrony in parietal regions \cite{Biel2022}. The underlying mechanism is primarily driven by theta-gamma coupling \cite{JensenLisman1998, 2005}, where a single gamma cycle corresponds to a specific memory item \cite{Axmacher2010}. By delivering weak electrical currents at specific frequencies to synchronize brain oscillations, tACS can influence cognitive performance \cite{Antal2008, Nitsche2008, Riecke2015}. Specifically, theta-tACS can reduce an individual's theta frequency, thereby accommodating more gamma cycles within a single theta cycle to expand working memory capacity.

Current evidence suggests that theta-gamma coupling is primarily involved in the encoding stage of working memory, whereas the maintenance stage relies more heavily on alpha oscillations. Research indicates that alpha oscillations protect working memory representations by inhibiting activity in irrelevant brain regions \cite{JensenMazaheri2010}, while beta oscillations are responsible for updating and maintaining memory content \cite{SpitzerBlankenburg2011}. This functional dissociation suggests that the regulatory effects of theta-tACS may be stage-specific. Specifically, it may enhance information input efficiency during the encoding phase of a trial by modulating theta-gamma coupling, while having a limited impact on the maintenance phase. The potential for enhancing working memory performance has gained significant recognition, yet results across different studies remain inconsistent. For instance, Zhang et al. found that $\theta$ stimulation can effectively improve visuospatial working memory performance \cite{Zhang2022}, while others did not observe significant enhancement, suggesting that outcomes are influenced by multiple factors such as frequency, target localization, and timing.

The timing of stimulation may serve as an important latent factor. Studies applying stimulation prior to task performance generally find that in-phase stimulation has a significant enhancing effect on working memory \cite{Biel2022, Jones2019}. However, studies applying stimulation after task initiation show a different trend; when participants have already begun the task prior to stimulation, no significant impact on working memory capacity is observed. This discrepancy may be related to psychological inertia—the tendency for individuals to maintain automated behavioral patterns formed in repetitive situations. While psychological inertia maintains behavioral coherence, it can lead individuals to ignore optimal choices and increase cognitive rigidity when it conflicts with rational decision-making.

During the early stages of a task, theta-tACS may break existing neural oscillation patterns through phase resetting, lengthening the theta cycle to accommodate more gamma activity and helping to establish an optimal memory strategy. Conversely, after task practice, frontoparietal activation decreases as memory strategies become automated. At this stage, applying stimulation no longer adjusts the theta cycle to adapt to the task; instead, it enhances the synchronization and activation intensity of gamma neurons through theta-gamma coupling \cite{Lega2016}, thereby improving the quality (precision) of memory representations. The present study systematically investigates how stimulation timing modulates the effects of tACS on working memory through the lens of psychological inertia, using the recall report paradigm to distinguish between memory quantity and precision.

Research Objectives and Hypotheses

This study investigates whether trade-off strategies developed during the practice phase generate psychological inertia and validates the transferability of these strategies across different tasks.

Experiment 1: Psychological Inertia and Resource Allocation

In the Pre-stimulation group, participants receive stimulation at the very beginning of the task. We hypothesize that under active stimulation, participants will exhibit increased memory quantity while maintaining constant memory precision (Hypothesis 1). In the Post-stimulation group, participants receive stimulation only after the practice phase. We expect that the trade-off strategy formed during practice will have developed psychological inertia; thus, memory quantity will remain constant while memory precision improves under active stimulation (Hypothesis 2).

Experiment 2: Strategy Transfer and Phase Resetting

Experiment 2 explores whether $\theta$-tACS can facilitate the formation of new resource allocation strategies when individuals encounter novel tasks. We hypothesize that when individuals face a completely new task, $\theta$-tACS acts through a phase-resetting mechanism to break the cognitive inertia established in the previous task. For the first task, we expect results consistent with the Post-stimulation group in Experiment 1. For the second, novel task, we anticipate participants will form a new trade-off strategy, resulting in increased quantity and constant precision (Hypothesis 3).

1 Experiment 1

1.1.1 Experimental Design

A mixed-design was employed, in which stimulation timing (Pre-stimulation vs. Post-stimulation) served as the between-subjects variable and stimulation type (Active vs. Sham) served as the within-subjects variable.

1.1.2 Participants

Power analysis indicated a required sample size of 128. Undergraduate and graduate students ($N=128$, ages 18–25, $M=21.34$, $SD=1.56$, 64 males) voluntarily participated. Participants were randomly assigned to the two groups. All had normal vision and no history of neurological disorders.

1.1.3 Materials and Procedure

The experiment used a color recall task developed in MATLAB. Memory items were colored squares ($1.1^\circ \times 1.1^\circ$) presented on a gray background. Each trial began with a fixation point and a directional arrow (cued side). The memory array was presented for 200 ms, followed by a 1000 ms retention interval. Participants then selected the target color on a color wheel [FIGURE:1]. The total stimulation duration was 20 minutes.

1.1.5 Transcranial Alternating Current Stimulation (tACS)

Stimulation was delivered via the Starstim 8 system. The stimulation electrode was placed over P4, with the reference electrode over the right supraorbital ridge. Individualized current thresholds were determined ($M = 1.5 \pm 0.3 \text{ mA}$). Active stimulation was compared against a sham condition in a single-blind, counterbalanced design.

1.2 Data Analysis

Data were fitted using the Standard Mixture Model \cite{ZhangLuck2008} via the MemToolbox. This model provides two indices: guess rate ($g$), representing the probability of forgetting (inversely related to memory quantity), and standard deviation ($sd$), representing memory precision.

1.3 Results

1.3.1 Pre-stimulation Group

A $2 \times 2$ repeated-measures ANOVA revealed a significant main effect for stimulation condition on guess rate, $F(1, 17) = 5.64, p = 0.030, \eta_p^2 = 0.25$. The guess rate under active stimulation was significantly lower than sham, indicating increased memory quantity. The main effect on precision was not significant, $F(1, 17) = 0.01, p = 0.930$.

1.3.2 Post-stimulation Group

For the post-stimulation group, the main effect of stimulation on guess rate was not significant, $F(1, 17) = 0.62, p = 0.440$. However, a significant main effect was found for precision, $F(1, 17) = 5.24, p = 0.035, \eta_p^2 = 0.236$, with active stimulation significantly improving memory precision compared to sham.

2 Experiment 2

2.1.1 Experimental Design and Participants

Experiment 2 investigated whether strategies transfer to new tasks. Participants ($N=64$) completed two distinct tasks (color and orientation) in a randomized order. Stimulation was applied after the practice phase of the first task.

2.1.3 Materials and Procedure

The color task followed the same parameters as Experiment 1. The orientation task required participants to recall the orientation of black rectangular lines ($1.5^\circ \times 0.2^\circ$) and adjust a probe to match the target orientation.

2.1.6 Results

For the First Task (Post-practice stimulation), paired-sample t-tests showed that tACS significantly improved precision ($p < 0.05$) but had no effect on memory quantity ($p > 0.05$). For the Second Task (Novel task), tACS significantly increased memory quantity (lower guess rate, $p < 0.05$) but had no effect on precision ($p > 0.05$). This confirms that when the task context changes, individuals break the previous inertia and form a new strategy.

3 General Discussion

Our findings indicate that the timing of electrical stimulation relative to task practice determines the nature of the cognitive enhancement. When stimulation is applied prior to practice, it facilitates an increase in memory quantity. When applied after practice, it enhances memory precision. This suggests that initial practice establishes a "stable strategy" protected by psychological inertia. In the post-practice condition, the brain may be in a stable oscillatory mode, making it resistant to changes in capacity but allowing for the enhancement of representation quality via theta-gamma coupling \cite{Lega2016}.

Furthermore, Experiment 2 demonstrates cognitive flexibility. Participants do not carry over strategies from the first task to a second, different task. Instead, the change in context disrupts psychological inertia, allowing the nervous system to re-evaluate resources and form a new trade-off strategy. This has significant implications for clinical applications, suggesting that the "temporal window" of intervention is critical for achieving specific therapeutic goals, such as improving memory precision in patients with neurodegenerative diseases.

4 Conclusion

The regulatory effects of theta-tACS on working memory are modulated by stimulation timing. Pre-practice stimulation primarily enhances memory quantity, while post-practice stimulation primarily enhances memory precision. These established strategies are task-specific and do not transfer to novel tasks, highlighting the brain's ability to flexibly reallocate resources in response to changing environmental demands.