Cognitive load and prefrontal neuromodulation shape exploration, risk-taking, and information seeking in experience-based decision making

Ankit Singh; Varun Dutt; Akash K Rao

doi:10.1371/journal.pone.0360059

. 2026 Oct 9;21(10):e0360059. doi: 10.1371/journal.pone.0360059

Cognitive load and prefrontal neuromodulation shape exploration, risk-taking, and information seeking in experience-based decision making

Ankit Singh ¹, Varun Dutt ¹, Akash K Rao ^2,^*

Editor: Xianwei Che³

PMCID: PMC13652248 PMID: 42853811

Abstract

Experience-based decision making requires individuals to balance exploration and exploitation under uncertainty, a balance that may be influenced by cognitive demands and prefrontal neuromodulation. We examined the joint effects of trivia-based cognitive load and anodal transcranial direct current stimulation (tDCS) over the left dorsolateral prefrontal cortex (DLPFC) in 60 healthy adults. The primary 2 × 2 between-subjects design crossed Stimulation (anodal tDCS vs sham) with Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control). Within the Trivia-Based Load subgroup, Lower-Load and Higher-Load Trivia were presented across trials for a secondary trial-level analysis. Participants completed 50 decisions involving a safe choice, a probabilistic high-reward choice, and a costly information-seeking choice. Primary analyses showed that Trivia-Based Load slowed responses and reduced risky choice. Anodal tDCS increased risky choice, information seeking, and alternation between choice types, but did not significantly affect reaction time. Stimulation × Cognitive-Load Condition interactions were significant for information seeking and alternation, reflecting different load-related patterns under anodal tDCS and sham, whereas interactions were not significant for reaction time or risky choice. Secondary analyses indicated longer decision times following Higher-Load than Lower-Load Trivia, without corresponding trial-level load effects on the choice outcomes. These findings show that cognitive load and anodal DLPFC tDCS have outcome-specific joint effects on experience-based decision making, including both additive and context-dependent patterns.

1. Introduction

Human decision making in uncertain environments requires a continuous balance between exploring new alternatives and exploiting familiar options that have yielded favorable outcomes in the past [1–3]. This exploration-exploitation trade-off is especially important in decision making from experience, where outcome probabilities are not described explicitly and must be learned through repeated feedback [4–7]. Such decisions depend on the retrieval of previous outcomes, the updating of value representations, and the allocation of attention to recent feedback [6–8].

Cognitive demands and prefrontal control may both influence this adaptive balance. Concurrent cognitive load can limit working-memory and attentional resources, whereas the dorsolateral prefrontal cortex (DLPFC) supports goal maintenance, feedback integration, and flexible adjustment of behavior [2,9–13]. Transcranial direct current stimulation (tDCS) provides an experimental method for examining whether modulation over the DLPFC is associated with changes in decision behavior, although the direction and magnitude of its behavioral effects depend on stimulation parameters, task demands, and individual differences [14–21].

Previous studies have generally examined cognitive load and prefrontal stimulation separately. Consequently, it remains unclear whether concurrent cognitive demand and anodal DLPFC stimulation produce broadly additive effects or whether the effect of stimulation changes when decision making is performed under additional cognitive demand. The present study addressed this question using a repeated decision-from-experience task containing safe, risky, and information-seeking alternatives.

The primary experiment used a 2 × 2 between-subjects design crossing Stimulation (anodal tDCS vs sham) with Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control). Within the Trivia-Based Load condition, Lower-Load and Higher-Load Trivia were presented across trials to examine secondary trial-level cognitive-load effects. The study examined reaction time, risky choice, information seeking, and alternation rate as related but distinct indicators of decision behavior under uncertainty.

The objective was to determine whether anodal tDCS over the left DLPFC and trivia-based cognitive demand were associated with additive or context-dependent differences in experience-based decision making. The following Background section develops the theoretical basis for the behavioral measures and hypotheses while avoiding repetition of the general study rationale presented above.

2. Background

In experience-based decision making, individuals repeatedly sample outcomes and use feedback to decide whether to continue exploiting a familiar option or explore an uncertain alternative [1,2,4,5]. Within Instance-Based Learning Theory and related reinforcement-learning accounts, choice emerges from the retrieval of previous outcome instances and the updating of option values [3,6–8]. When relevant outcomes are readily retrieved and integrated, individuals can adjust their choices flexibly. When concurrent demands interfere with these processes, retrieval may become noisier and value comparisons less stable, which can slow responding and shift behavior toward safer or simpler alternatives.

The DLPFC is relevant to these processes because it contributes to working memory, attentional allocation, goal maintenance, and the integration of recent feedback [2,9,10,13]. Anodal tDCS over the DLPFC has been associated with changes in working memory, risk-related choice, and exploratory behavior, but these effects are not uniform across tasks or stimulation protocols [15–22]. Therefore, the present study does not treat behavioral changes following stimulation as direct evidence of a single neural mechanism. Instead, it examines whether anodal stimulation over the left DLPFC is associated with systematic differences in risky choice, costly information seeking, and switching behavior.

Within the present task, selection of the probabilistic high-reward alternative was treated as an index of risk-taking because participants accepted the possibility of receiving no reward in exchange for a substantially larger potential gain than the guaranteed safe option. Selection of the costly reveal alternative was treated as an index of information seeking because participants paid one point to obtain otherwise unavailable probability information before making their final choice. Alternation rate represented exploratory switching because it measured changes between choice types across successive trials; however, higher alternation could also reflect weaker commitment or unstable value learning. These measures were therefore interpreted as complementary rather than interchangeable indicators of exploration-exploitation behavior.

Resource-limitation and dual-process accounts predict that concurrent cognitive demand should reduce the resources available for evaluating uncertain alternatives, thereby slowing responses and encouraging more conservative choice [11,12,23–28]. In the present design, this prediction was tested primarily by comparing the Trivia-Based Load condition with the No-Trivia Control condition. The secondary within-participant comparison of Lower-Load and Higher-Load Trivia examined whether greater trial-level processing requirements produced additional changes within the loaded condition.

Based on evidence that DLPFC stimulation can alter risk-related and exploratory behavior, we expected participants receiving anodal tDCS to show greater engagement with uncertain and information-rich alternatives than participants receiving sham stimulation. Thus, we expected:

H1. Participants receiving anodal tDCS over the left DLPFC will exhibit greater risky choice, information seeking, and exploratory switching than participants receiving sham stimulation.

Because concurrent trivia requires additional information processing before each decision, we expected it to increase response time and constrain engagement with uncertain alternatives. Thus, we expected:

H2. Participants in the Trivia-Based Load condition will respond more slowly and will show reduced risky and exploratory choice relative to participants in the No-Trivia Control condition. Within the Trivia-Based Load condition, Higher-Load Trivia is expected to produce slower responses than Lower-Load Trivia.

Although cognitive load and anodal DLPFC stimulation were expected to exert opposing directional influences on some outcomes, two theoretical patterns were possible. Under a partially independent account, cognitive load would primarily constrain working-memory and attentional resources, whereas anodal stimulation would modulate prefrontal contributions to exploratory control, producing broadly additive effects. Alternatively, if stimulation strengthens the same processes disrupted by cognitive load, its behavioral effect could differ across Cognitive-Load Conditions, producing a compensatory or context-dependent interaction. Thus, we expected:

H3. Anodal DLPFC stimulation will increase exploratory and risk-related behavior in both the Trivia-Based Load and No-Trivia Control conditions. Under the partially independent account, the magnitude of the stimulation effect is expected to be broadly comparable across Cognitive-Load Conditions, producing an additive pattern rather than systematic moderation by cognitive load.

3. Methods

3.1. Participants

Sixty healthy adult volunteers (N = 60; age range = 18–45 years; M = 24.8, SD = 3.4; 36 males and 24 females) were recruited from science, humanities, and engineering backgrounds at the Indian Institute of Technology Mandi. Participants were randomly assigned to one of four experimental groups (n = 15 per group). All participants had normal or corrected-to-normal vision and reported no neurological, psychiatric, or major medical conditions. A safety questionnaire was used to screen for contraindications to tDCS, including a history of seizures, skin irritation at electrode sites, and use of psychoactive medication.

Written informed consent was obtained before participation. Participants were informed that they could withdraw at any time without penalty, and anonymity and data confidentiality were maintained. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committee of IIT Mandi (Approval No. IITM/IEC(H)/2025/VD/P1). Recruitment occurred between 5 April 2025 and 18 July 2025. Testing was conducted in a quiet, sound-attenuated, distraction-free room in the Applied Cognitive Science Laboratory. Participants received fixed monetary compensation independent of task performance.

3.2. Study design

The primary experiment followed a 2 × 2 between-subjects factorial design with Stimulation (anodal tDCS vs sham) and Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control) as between-subject factors. The four groups were anodal tDCS with Trivia-Based Load, anodal tDCS with No-Trivia Control, sham with Trivia-Based Load, and sham with No-Trivia Control. Participants in the Trivia-Based Load condition answered a trivia question before each decision trial, whereas participants in the No-Trivia Control condition proceeded directly to the decision task. Within the Trivia-Based Load condition, Trial-Level Cognitive Load was manipulated within participants by presenting 25 Lower-Load and 25 Higher-Load Trivia questions in randomized order.

Randomization was performed using a computer-generated allocation sequence before data collection. Participants were not informed of their active or sham assignment, and the sham ramp procedure was used to support participant blinding. The experimenter was aware of the stimulation condition during testing. Thus, the study followed a single-blind design. Blinding success was not formally assessed. Active or sham stimulation was completed immediately before the behavioral task.

3.3. Transcranial direct current stimulation

Transcranial direct current stimulation was delivered using the Neuroelectrics Starstim 32 system (Neuroelectrics, Barcelona, Spain), a wireless multichannel stimulator with real-time monitoring of electrode contact quality. Stimulation was applied through Neuroelectrics PiStim Ag/AgCl electrodes containing a 12-mm Ag/AgCl pellet (π cm² contact area) embedded in a saline-gel interface. Electrodes were positioned using a neoprene headcap according to the international 10−10 EEG system. The anodal electrode was placed at F3 to target the left DLPFC, and return electrodes were positioned over the right supraorbital region at Fp2 and AF4. The active and sham protocols are illustrated in Fig 1, and modeled voltage and electric-field distributions are shown in Fig 2.

Fig 1 — (A) Active stimulation: current was ramped to 2 mA, maintained for approximately 20 minutes, and ramped down at the end of stimulation. (B) Sham stimulation: current was ramped to 2 mA and then immediately ramped down to reproduce the initial somatosensory experience without sustained stimulation.

Fig 2 — (A) Montage with the anode at F3 and return electrodes over the right supraorbital region at Fp2 and AF4. (B-E) Modeled electric-field magnitude and voltage distributions in cortical and orthogonal views. Simulations were computed using ROAST [29] with an applied current of 2 mA and a standard head model.

Participants in the active condition received a constant current of up to 2.0 mA for 20–22 minutes, including a 45-s ramp-up and a 30-s ramp-down. Participants were seated and asked to remain still and relaxed during stimulation. Participants in the sham condition received the same electrode setup and a 45-s ramp-up to 2.0 mA followed by immediate ramp-down, after which no sustained current was delivered. No serious adverse effects were reported; transient tingling was reported by some participants. Current-flow simulations were conducted using ROAST [29] with a standard head model and an applied current of 2 mA. The simulations were used to characterize the modeled field distribution of the montage rather than to establish individual target engagement.

3.4. Decision-making task

The decision-making task consisted of 50 trials, each requiring participants to select among three alternatives. In the Trivia-Based Load condition, each trial began with a trivia question and feedback on the trivia response, followed by the decision screen and outcome feedback. In the No-Trivia Control condition, the trial began directly with the decision screen. The task sequence is illustrated in Fig 3.

Fig 3 — (A) Trivia question presented before the decision phase in the Trivia-Based Load condition. (B) Trivia-response feedback. (C) Decision screen containing the safe, risky, and information-seeking choices. (D) Outcome feedback. (E) Example of a participant completing the task.

Safe choice (Option A): a guaranteed outcome between 2.5 and 3.5 points, with values jittered uniformly across trials.
Risky choice (Option B): a probabilistic outcome yielding either 0 points or a larger reward between 16 and 18 points. The probability of receiving the larger reward was approximately 20%, with minor jitter between 18% and 22%.
Information-seeking choice (Option R): a costly option that revealed the probability associated with the risky outcome on that trial. Selecting this option deducted one point before the participant proceeded to choose the safe or risky alternative.

All stimuli were presented using PsychoPy v2023.2.3 on a 24-inch LCD monitor (1920 × 1080 pixels, 60-Hz refresh rate; BenQ, Taiwan) at an approximate viewing distance of 55 cm. The spatial positions and response keys for the safe, risky, and information-seeking choices remained fixed throughout the task to maintain a consistent response mapping. Participants responded using designated keyboard keys, and responses and reaction times were recorded automatically. A brief practice session preceded the experiment.

3.5. Trivia-based cognitive-load manipulation

Participants in the Trivia-Based Load condition answered one general-knowledge question before each decision trial. Questions covered science, geography, culture, mathematics, and history. Each item presented four response options, and trivia accuracy did not affect points or decision outcomes. The questions were used exclusively to impose concurrent cognitive demand.

Trial-Level Cognitive Load was manipulated through two item formats. Lower-Load Trivia consisted primarily of concise factual-recall questions with a single correct response. Higher-Load Trivia included assertion-reason items, multiple-statement items, comparative or relational judgments, and questions requiring integration of more than one information component. Each participant in the Trivia-Based Load condition completed 25 Lower-Load and 25 Higher-Load Trivia trials in randomized order. Participants in the No-Trivia Control condition did not receive trivia questions. Trivia accuracy and reaction time were recorded as indices of task engagement.

3.6. Behavioral outcome measures

Four dependent measures were computed. Risk-taking was defined as the proportion of trials on which the risky choice was selected. Information seeking was defined as the proportion of trials on which the costly information-seeking choice was selected. Alternation rate was defined as the proportion of successive trial transitions involving a switch between the safe choice (Option A) and the risky choice (Option B). Mean decision reaction time was calculated from the presentation of the decision screen to the participant’s response. The measures were analyzed separately because they represent related but distinct components of decision behavior.

3.7. Statistical analysis

3.7.1. Primary between-subjects analyses.

The primary analysis consisted of a separate 2 × 2 factorial ANOVA for each dependent measure, with Stimulation (anodal tDCS vs sham) and Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control) as between-subject factors. These models tested the main effects of Stimulation and Cognitive-Load Condition and their interaction. The reaction-time analysis included all 60 participants. The combined choice-outcome dataset contained 59 complete cases; therefore, the primary factorial analyses of risky choice, information seeking, and alternation rate used N = 59. The primary analyses were interpreted outcome by outcome rather than as evidence for a single latent exploratory construct.

3.7.2. Secondary trial-level mixed analyses.

The secondary analyses included the 30 participants assigned to the Trivia-Based Load condition. For each outcome, a 2 (Trial-Level Cognitive Load: Lower-Load Trivia vs Higher-Load Trivia) × 5 (Trial Block: five consecutive blocks of ten trials) × 2 (Stimulation: anodal tDCS vs sham) mixed-design repeated-measures ANOVA was conducted. Trial-Level Cognitive Load and Trial Block were within-subject factors, and Stimulation was a between-subject factor. These analyses tested whether trivia difficulty altered decision behavior within participants, whether behavior changed across the task, and whether either pattern differed by stimulation condition.

3.7.3. General statistical considerations.

Analyses were performed in SPSS version 25 (IBM, Armonk, NY, USA). Statistical significance was evaluated at p < .05, and partial eta-squared (ηp²) is reported as the ANOVA effect-size measure. Homogeneity of variance was examined using Levene’s test. Sphericity was evaluated for within-subject effects with more than two levels, and Greenhouse-Geisser-corrected results were used when sphericity was violated. Exact p-values are reported except when p < .001.

4. Results

4.1. Primary factorial analyses: Overview

The primary factorial analyses produced different patterns across the four dependent measures. Cognitive-Load Condition significantly affected reaction time and risky choice. Stimulation significantly affected risky choice, information seeking, and alternation rate. The Stimulation × Cognitive-Load Condition interaction was significant for information seeking and alternation rate, but not for reaction time or risky choice. The findings therefore did not support a uniform additive pattern across all outcomes.

4.2. Reaction time

Participants in the Trivia-Based Load condition responded more slowly (M = 4.35 s, SD = 2.15) than participants in the No-Trivia Control condition (M = 3.27 s, SD = 1.99). The main effect of Cognitive-Load Condition was significant, F(1,56) = 4.219, p = .045, ηp² = .070. The main effect of Stimulation was not significant, F(1,56) = 3.114, p = .083, ηp² = .053, although mean reaction time was numerically shorter under anodal tDCS (M = 3.34 s, SD = 1.73) than under sham (M = 4.28 s, SD = 2.40). The Stimulation × Cognitive-Load Condition interaction was not significant, F(1,56) = 0.264, p = .609, ηp² = .005. Thus, the primary reaction-time result supported the prediction that concurrent trivia increased decision time, without evidence that this effect differed by stimulation condition (Fig 4).

Fig 4 — (A) Mean reaction time in the No-Trivia Control and Trivia-Based Load conditions. (B) Mean reaction time for Lower-Load and Higher-Load Trivia trials in the secondary repeated-measures analysis. Error bars represent 95% confidence intervals.

4.3. Risky choice

Risk-taking was higher under anodal tDCS (M = .591, SD = .070) than under sham stimulation (M = .338, SD = .059), producing a significant main effect of Stimulation, F(1,55) = 323.285, p < .001, ηp² = .855. Risky choice was lower in the Trivia-Based Load condition (M = .428, SD = .142) than in the No-Trivia Control condition (M = .506, SD = .135), F(1,55) = 27.109, p < .001, ηp² = .330. The Stimulation × Cognitive-Load Condition interaction was not significant, F(1,55) = 0.065, p = .800, ηp² = .001. Cell means showed the same directional load-related reduction under sham (No-Trivia M = .377; Trivia-Based M = .301) and anodal tDCS (No-Trivia M = .625; Trivia-Based M = .557). This outcome supported the predicted stimulation-related increase and load-related reduction in risky choice, with an additive rather than moderated pattern (Fig 5).

Fig 5 — (A) Mean proportion of risky choices under Sham and tDCS stimulation. (B) Mean proportion of risky choices in the No-Trivia Control and Trivia-Based Load conditions. Error bars represent 95% confidence intervals.

4.4. Information-seeking choice

Information seeking was higher overall under anodal tDCS (M = .313, SD = .091) than under sham stimulation (M = .261, SD = .087), F(1,55) = 5.916, p = .018, ηp² = .097. The main effect of Cognitive-Load Condition was not significant, F(1,55) = 0.010, p = .921, ηp² < .001. However, the Stimulation × Cognitive-Load Condition interaction was significant, F(1,55) = 14.217, p < .001, ηp² = .205. The cell means showed a crossover pattern: under sham, information seeking was lower in the Trivia-Based Load condition (M = .221) than in the No-Trivia Control condition (M = .303), whereas under anodal tDCS it was higher in the Trivia-Based Load condition (M = .352) than in the No-Trivia Control condition (M = .275). Because simple-effects tests were not included in the supplied output, the cell-specific differences are described directionally rather than interpreted as individually significant. The interaction indicates that the stimulation-related difference in information seeking depended on the concurrent cognitive-load context (Fig 6).

Fig 6 — (A) Mean proportion of information-seeking choices under Sham and tDCS stimulation. (B) Mean proportion of information-seeking choices as a function of Stimulation and Cognitive-Load Condition. Error bars represent 95% confidence intervals.

4.5. Alternation rate

Alternation rate was higher overall under anodal tDCS (M = .187, SD = .081) than under sham stimulation (M = .132, SD = .047), F(1,55) = 10.753, p = .002, ηp² = .164. The main effect of Cognitive-Load Condition was not significant, F(1,55) = 1.059, p = .308, ηp² = .019. The Stimulation × Cognitive-Load Condition interaction was significant, F(1,55) = 6.207, p = .016, ηp² = .101. Under sham, alternation was numerically lower in the Trivia-Based Load condition (M = .120) than in the No-Trivia Control condition (M = .145), whereas under anodal tDCS alternation was numerically higher in the Trivia-Based Load condition (M = .216) than in the No-Trivia Control condition (M = .158). As with information seeking, this crossover pattern indicates context-dependent modulation rather than a uniform additive effect. Levene’s test was significant for alternation rate, p = .002; therefore, the magnitude of this interaction should be interpreted cautiously (Fig 7).

Fig 7 — (A) Mean alternation rate under Sham and tDCS stimulation. (B) Mean alternation rate as a function of Stimulation and Cognitive-Load Condition. (C) Mean alternation rate across the five Trial Blocks in the secondary repeated-measures analysis. Error bars represent 95% confidence intervals.

4.6. Secondary trial-level cognitive load and trial block analyses

Secondary mixed-design analyses were conducted on the 30 participants assigned to the Trivia-Based Load condition to examine the effects of Trial-Level Cognitive Load (Higher-Load vs. Lower-Load Trivia) and Trial Block (five consecutive blocks of ten trials) on decision-making behavior. These analyses were considered secondary because they examined differences in trivia difficulty within the Trivia-Based Load condition rather than the primary comparison between the Trivia-Based Load and No-Trivia Control conditions.

Higher-Load Trivia was associated with significantly longer decision reaction times than Lower-Load Trivia, F(1,28) = 11.631, p = .002, ηp² = .293, indicating that increased trivia difficulty slowed subsequent decision responses. However, the Trial-Level Cognitive Load × Stimulation interaction was not significant, F(1,28) = 0.062, p = .805, ηp² = .002. Likewise, no significant main effect of Trial Block or higher-order interactions involving Trial Block was observed, indicating that the effect of cognitive load on reaction time remained stable across the task. For risky choice, neither the main effect of Trial-Level Cognitive Load, F(1,28) = 0.002, p = .967, ηp² < .001, nor the Trial-Level Cognitive Load × Stimulation interaction, F(1,28) = 0.029, p = .865, ηp² = .001, was significant. The main effect of Trial Block and the remaining interactions involving Trial Block were also not significant, indicating that risky choice remained relatively stable across the five trial blocks irrespective of trivia difficulty. For information-seeking behavior, neither the main effect of Trial-Level Cognitive Load, F(1,28) = 0.089, p = .768, ηp² = .003, nor the Trial-Level Cognitive Load × Stimulation interaction, F(1,28) = 0.544, p = .467, ηp² = .019, reached significance. Likewise, neither the main effect of Trial Block, F(4,112) = 1.169, p = .328, ηp² = .040, nor the remaining interactions involving Trial Block was significant, indicating that information-seeking behavior remained consistent throughout the task. For alternation rate, neither the main effect of Trial-Level Cognitive Load, F(1,28) = 1.112, p = .303, ηp² = .038, nor the Trial-Level Cognitive Load × Stimulation interaction, F(1,28) = 0.500, p = .486, ηp² = .018, was significant. In contrast, a significant main effect of Trial Block was observed, F(4,112) = 2.591, p = .040, ηp² = .085, indicating modest variation in alternation behavior across the five trial blocks. However, no significant interactions involving Trial Block were observed, suggesting that this temporal variation was common across stimulation conditions and was not influenced by trial-level cognitive load.

Overall, these secondary analyses indicate that Trial-Level Cognitive Load primarily influenced decision reaction time, whereas risky choice, information-seeking behavior, and alternation rate were largely unaffected. Apart from modest temporal variation in alternation behavior across trial blocks, no evidence was found that Trial-Level Cognitive Load systematically modified the behavioral effects of stimulation.

5. Discussion

The present study examined how anodal tDCS over the left DLPFC and trivia-based cognitive demand jointly influenced experience-based decision making. The revised analysis distinguishes the primary between-subject contrast between Trivia-Based Load and No-Trivia Control from the secondary within-participant contrast between Higher-Load and Lower-Load Trivia. This distinction revealed an outcome-specific pattern. Concurrent trivia slowed decisions and reduced risky choice, while anodal tDCS was associated with greater risky choice, information seeking, and alternation. The two manipulations produced additive patterns for reaction time and risky choice, but significant interactions for information seeking and alternation. The findings therefore do not support a single conclusion that cognitive load and stimulation operated independently across all measures.

The reaction-time result indicates that performing a trivia task before each decision imposed measurable processing demands. Participants in the Trivia-Based Load condition required more time to respond than participants who proceeded directly to the decision screen. This pattern is consistent with Cognitive Load Theory and dual-process accounts, which predict that concurrent processing demands reduce the immediate availability of working-memory and attentional resources [11,12,23,25–28]. The secondary trial-level pattern, in which Higher-Load Trivia produced longer decision times than Lower-Load Trivia, further suggests that item difficulty affected processing efficiency even within the trivia condition. By contrast, the absence of reliable trial-level difficulty effects on the choice outcomes suggests that the presence of a concurrent trivia task and the difficulty of individual trivia items should not be treated as equivalent manipulations.

The secondary Trial-Level Cognitive Load analyses further clarified this distinction. Within the Trivia-Based Load condition, Higher-Load Trivia was associated with slower subsequent decision responses, whereas risky choice, information seeking, and alternation rate did not differ significantly between Higher-Load and Lower-Load Trivia trials. A modest main effect of Trial Block was observed for alternation rate, although this temporal variation did not interact with Trial-Level Cognitive Load or stimulation. These findings suggest that variation in trivia difficulty was expressed primarily in response speed rather than in the distribution of decision choices across the task.

Risky choice showed the clearest additive pattern. Trivia-Based Load reduced selection of the probabilistic high-reward alternative under both stimulation conditions, while anodal tDCS increased risky choice under both Cognitive-Load Conditions. The load-related reduction is consistent with more conservative responding when attentional resources are constrained [24]. The stimulation-related difference is consistent with previous reports that prefrontal neuromodulation can alter risk-related decisions [17,22,30]. Nevertheless, the large stimulation effect should be interpreted as a behavioral association produced by the present montage and task, not as direct evidence that a specific prefrontal computation or cortical-excitability mechanism caused the choice pattern. Information seeking and alternation produced a different result. In both outcomes, the effect of Trivia-Based Load changed direction across stimulation conditions: the trivia condition was associated with lower information seeking and alternation under sham but higher values under anodal tDCS. Information seeking required participants to pay a cost to reduce uncertainty, whereas alternation captured changes between choice types across successive trials. The parallel crossover patterns suggest that anodal stimulation altered how concurrent cognitive demand was expressed in exploratory behavior. However, information seeking may reflect strategic uncertainty reduction rather than curiosity-driven exploration alone [31,32], and alternation may reflect either flexible exploration or instability in choice preferences. The interaction for alternation should also be interpreted cautiously because homogeneity of variance was violated. These results refine the theoretical alternatives presented in the Background. Reaction time and risky choice were compatible with an additive account, because load-related differences were similar across stimulation conditions. In contrast, information seeking and alternation were compatible with a context-dependent account, because stimulation moderated the pattern associated with the Trivia-Based Load condition. A non-significant interaction for two outcomes cannot establish mechanistic independence, just as a significant interaction does not identify the neural process that generated it. The present behavioral data therefore support a mixed account in which cognitive load demands and anodal DLPFC tDCS can have both additive and interactive consequences depending on the component of decision behavior being measured.

Several limitations should be considered. First, the dependent measures are behavioral proxies rather than process-pure indicators. Risky choice, costly information seeking, and alternation capture different aspects of exploration and should not be combined into a single construct without additional modeling. Second, the positions and response keys for the three choices were fixed, which may have introduced position or key-response biases. Third, participant blinding was not formally evaluated, and the experimenter was aware of the stimulation condition. Fourth, the sample was drawn primarily from a university population, and a fixed stimulation dose was applied across individuals. Variation in individual anatomy, baseline state, and current distribution may therefore have contributed to between-participant differences [29,33,34]. Finally, cognitive load was induced through semantic trivia, and future work should test whether the pattern generalizes to visuospatial, working-memory, or multitasking demands. Future studies could integrate electroencephalography (EEG) or functional near-infrared spectroscopy (fNIRS) to examine the neural correlates underlying cognitive load and stimulation-related differences during experience-based decision making [35,36].

6. Conclusion

The present study combined anodal tDCS over the left DLPFC with a trivia-based cognitive-load manipulation in a repeated three-option decision-from-experience task. Concurrent trivia slowed responses and reduced risky choice, while anodal tDCS was associated with greater risky choice, information seeking, and exploratory switching. The joint pattern was not uniform: reaction time and risky choice showed no significant Stimulation × Cognitive-Load Condition interaction, whereas information seeking and alternation showed significant crossover interactions. The findings are therefore most consistent with outcome-specific modulation, including both additive and context-dependent effects. They provide behavioral evidence that cognitive demand and prefrontal stimulation jointly shape decision making under uncertainty, while leaving the underlying neurophysiological mechanisms to be established by future work.

Acknowledgments

We thank the Centre for Human-Computer Interaction and the Applied Cognitive Science Laboratory at the Indian Institute of Technology Mandi for providing the laboratory facilities, stimulation equipment, and computational resources required for this study. We also thank our laboratory colleagues for their discussions and feedback on the manuscript.

Data Availability

The anonymized behavioral data underlying the findings reported in this study are publicly available in the following Figshare repository: https://doi.org/10.6084/m9.figshare.33317310.

Funding Statement

The author(s) received no specific funding for this work.

References

1.Daw ND, O’Doherty JP, Dayan P, Seymour B, Dolan RJ. Cortical substrates for exploratory decisions in humans. Nature. 2006;441(7095):876–9. doi: 10.1038/nature04766 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Cohen JD, McClure SM, Yu AJ. Should I stay or should I go? How the human brain manages the trade-off between exploitation and exploration. Philos Trans R Soc Lond B Biol Sci. 2007;362(1481):933–42. doi: 10.1098/rstb.2007.2098 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Wilson RC, Geana A, White JM, Ludvig EA, Cohen JD. Humans use directed and random exploration to solve the explore-exploit dilemma. J Exp Psychol Gen. 2014;143(6):2074–81. doi: 10.1037/a0038199 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Hertwig R, Erev I. The description-experience gap in risky choice. Trends Cogn Sci. 2009;13(12):517–23. doi: 10.1016/j.tics.2009.09.004 [DOI] [PubMed] [Google Scholar]
5.Rakow T, Newell BR. Degrees of uncertainty: An overview and framework for future research on experience-based choice. J Behav Decis Mak. 2010;23(1):1–14. doi: 10.1002/bdm.681 [DOI] [Google Scholar]
6.Dutt V, Gonzalez C. The role of inertia in modeling decisions from experience with instance-based learning. Front Psychol. 2012;3:177. doi: 10.3389/fpsyg.2012.00177 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Gonzalez C, Dutt V. Instance-based learning: Integrating sampling and repeated decisions from experience. Psychol Rev. 2011;118(4):523–51. doi: 10.1037/a0024558 [DOI] [PubMed] [Google Scholar]
8.Gonzalez C, Lerch JF, Lebiere C. Instance-based learning in dynamic decision making. Cogn Sci. 2003;27(4):591–635. doi: 10.1207/s15516709cog2704_2 [DOI] [Google Scholar]
9.Miller EK, Cohen JD. An integrative theory of prefrontal cortex function. Annu Rev Neurosci. 2001;24:167–202. doi: 10.1146/annurev.neuro.24.1.167 [DOI] [PubMed] [Google Scholar]
10.Koechlin E. Prefrontal executive function and adaptive behavior in complex environments. Curr Opin Neurobiol. 2016;37:1–6. doi: 10.1016/j.conb.2015.11.004 [DOI] [PubMed] [Google Scholar]
11.Sweller J. Cognitive load during problem solving: Effects on learning. Cogn Sci. 1988;12(2):257–85. doi: 10.1207/s15516709cog1202_4 [DOI] [Google Scholar]
12.Kahneman D. Thinking, fast and slow. New York: Farrar, Straus and Giroux; 2011. [Google Scholar]
13.Badre D, Nee DE. Frontal cortex and the hierarchical control of behavior. Trends Cogn Sci. 2018;22(2):170–88. doi: 10.1016/j.tics.2017.11.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Nitsche MA, Paulus W. Sustained excitability elevations induced by transcranial DC motor cortex stimulation in humans. Neurology. 2001;57(10):1899–901. doi: 10.1212/wnl.57.10.1899 [DOI] [PubMed] [Google Scholar]
15.Fregni F, Boggio PS, Nitsche M, Bermpohl F, Antal A, Feredoes E, et al. Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory. Exp Brain Res. 2005;166(1):23–30. doi: 10.1007/s00221-005-2334-6 [DOI] [PubMed] [Google Scholar]
16.Brunoni AR, Vanderhasselt M-A. Working memory improvement with non-invasive brain stimulation of the dorsolateral prefrontal cortex: A systematic review and meta-analysis. Brain Cogn. 2014;86:1–9. doi: 10.1016/j.bandc.2014.01.008 [DOI] [PubMed] [Google Scholar]
17.Fecteau S, Knoch D, Fregni F, Sultani N, Boggio P, Pascual-Leone A. Diminishing risk-taking behavior by modulating activity in the prefrontal cortex: A direct current stimulation study. J Neurosci. 2007;27(46):12500–5. doi: 10.1523/JNEUROSCI.3283-07.2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Nitsche MA, Cohen LG, Wassermann EM, Priori A, Lang N, Antal A, et al. Transcranial direct current stimulation: State of the art 2008. Brain Stimul. 2008;1(3):206–23. doi: 10.1016/j.brs.2008.06.004 [DOI] [PubMed] [Google Scholar]
19.Polanía R, Nitsche MA, Ruff CC. Studying and modifying brain function with non-invasive brain stimulation. Nat Neurosci. 2018;21(2):174–87. doi: 10.1038/s41593-017-0054-4 [DOI] [PubMed] [Google Scholar]
20.Dedoncker J, Brunoni AR, Baeken C, Vanderhasselt M-A. A Systematic review and meta-analysis of the effects of transcranial direct current stimulation (tDCS) over the dorsolateral prefrontal cortex in healthy and neuropsychiatric samples: Influence of stimulation parameters. Brain Stimul. 2016;9(4):501–17. doi: 10.1016/j.brs.2016.04.006 [DOI] [PubMed] [Google Scholar]
21.Farhat LC, Carvalho AF, Solmi M, Brunoni AR. Evidence-based umbrella review of cognitive effects of prefrontal tDCS. Soc Cogn Affect Neurosci. 2022;17(1):43–60. doi: 10.1093/scan/nsaa084 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Boggio PS, Campanhã C, Valasek CA, Fecteau S, Pascual-Leone A, Fregni F. Modulation of decision-making in a gambling task in older adults with transcranial direct current stimulation. Eur J Neurosci. 2010;31(3):593–7. doi: 10.1111/j.1460-9568.2010.07080.x [DOI] [PubMed] [Google Scholar]
23.Sweller J. Cognitive load theory. Psychol Learn Motiv. 2011;55:37–76. doi: 10.1016/B978-0-12-387691-1.00002-8 [DOI] [Google Scholar]
24.Whitney P, Rinehart CA, Hinson JM. Framing effects under cognitive load: The role of working memory in risky decisions. Psychon Bull Rev. 2008;15(6):1179–84. doi: 10.3758/PBR.15.6.1179 [DOI] [PubMed] [Google Scholar]
25.Evans JSBT, Stanovich KE. Dual-process theories of higher cognition: Advancing the debate. Perspect Psychol Sci. 2013;8(3):223–41. doi: 10.1177/1745691612460685 [DOI] [PubMed] [Google Scholar]
26.Payne JW, Bettman JR, Johnson EJ. The adaptive decision maker. Cambridge: Cambridge University Press; 1993. [Google Scholar]
27.Otto AR, Skatova A, Madlon-Kay S, Daw ND. Cognitive control predicts use of model-based reinforcement learning. J Cogn Neurosci. 2015;27(2):319–33. doi: 10.1162/jocn_a_00709 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.van Steenbergen H, Band GPH, Hommel B. Does conflict help or hurt cognitive control? Initial evidence for an inverted U-shape relationship between perceived task difficulty and conflict adaptation. Front Psychol. 2015;6:974. doi: 10.3389/fpsyg.2015.00974 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Huang Y, Datta A, Bikson M, Parra LC. Realistic volumetric-approach to simulate transcranial electric stimulation-ROAST-a fully automated open-source pipeline. J Neural Eng. 2019;16(5):056006. doi: 10.1088/1741-2552/ab208d [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Knoch D, Gianotti LRR, Pascual-Leone A, Treyer V, Regard M, Hohmann M, et al. Disruption of right prefrontal cortex by low-frequency repetitive transcranial magnetic stimulation induces risk-taking behavior. J Neurosci. 2006;26(24):6469–72. doi: 10.1523/JNEUROSCI.0804-06.2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Gruber MJ, Gelman BD, Ranganath C. States of curiosity modulate hippocampus-dependent learning. Neuron. 2014;84(2):486–96. doi: 10.1016/j.neuron.2014.08.060 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Murayama K, FitzGibbon L, Sakaki M. Process account of curiosity and interest: A reward-learning perspective. Educ Psychol Rev. 2019;31:875–95. doi: 10.1007/s10648-019-09499-9 [DOI] [Google Scholar]
33.Saturnino GB, Puonti O, Nielsen JD, Antonenko D, Madsen KH, Thielscher A. SimNIBS 2.1: A comprehensive pipeline for individualized electric field modelling for transcranial brain stimulation. In: Makarov S, Horner M, Noetscher G. Brain and human body modeling: Computational human modeling at EMBC 2018. Cham: Springer. 2019. p. 3–25. doi: 10.1007/978-3-030-21293-3_1 [DOI] [PubMed] [Google Scholar]
34.Laakso I, Tanaka S, Koyama S, De Santis V, Hirata A. Inter-subject variability in electric fields of motor cortical tDCS. Brain Stimul. 2015;8(5):906–13. doi: 10.1016/j.brs.2015.05.002 [DOI] [PubMed] [Google Scholar]
35.Herff C, Heger D, Fortmann O, Hennrich J, Putze F, Schultz T. Mental workload during n-back task-quantified in the prefrontal cortex using fNIRS. Front Hum Neurosci. 2014;7:935. doi: 10.3389/fnhum.2013.00935 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Sood M, Besson P, Muthalib M, Jindal U, Perrey S, Dutta A, et al. NIRS-EEG joint imaging during transcranial direct current stimulation: Online parameter estimation with an autoregressive model. J Neurosci Methods. 2016;274:71–80. doi: 10.1016/j.jneumeth.2016.09.008 [DOI] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0360059.r001

Decision Letter 0

Xianwei Che

6 Jul 2026

PONE-D-26-16597

Cognitive Load and Prefrontal Neuromodulation Independently Shape Exploration, Risk-Taking, and Information Seeking in Experience-Based Decision Making

PLOS One

Dear Dr. Rao,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

==============================

Please submit your revised manuscript by Aug 20 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

As the corresponding author, your ORCID iD is verified in the submission system and will appear in the published article. PLOS supports the use of ORCID, and we encourage all coauthors to register for an ORCID iD and use it as well. Please encourage your coauthors to verify their ORCID iD within the submission system before final acceptance, as unverified ORCID iDs will not appear in the published article. Only the individual author can complete the verification step; PLOS staff cannot verify ORCID iDs on behalf of authors.

We look forward to receiving your revised manuscript.

Kind regards,

Xianwei Che

Academic Editor

PLOS One

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please note that PLOS One has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, we expect all author-generated code to be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

3. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager.

4. In the online submission form, you indicated that "All relevant data from this study (code and data sheet) will be made available upon request through link. it will be available for public ".

All PLOS journals now require all data underlying the findings described in their manuscript to be freely available to other researchers, either 1. In a public repository, 2. Within the manuscript itself, or 3. Uploaded as supplementary information.

This policy applies to all data except where public deposition would breach compliance with the protocol approved by your research ethics board. If your data cannot be made publicly available for ethical or legal reasons (e.g., public availability would compromise patient privacy), please explain your reasons on resubmission and your exemption request will be escalated for approval.

5. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: N/A

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

Reviewer #2: Yes

**********

Reviewer #1: Manuscript Title: Cognitive Load and Prefrontal Neuromodulation Independently Shape Exploration, Risk-Taking, and Information Seeking in Experience-Based Decision Making.

Summary of the study

This manuscript investigates how cognitive load and anodal transcranial direct current stimulation (tDCS) over the left dorsolateral prefrontal cortex (DLPFC) influence experience-based decision-making under uncertainty. The authors use a decision-making task in which participants repeatedly choose between a safe option, a risky option, and an information-seeking option. The study focuses on risk-taking, information seeking, alternation/switching behavior, and reaction time as behavioral indicators of exploration/exploitation dynamics.

Participants were assigned to stimulation and cognitive-load conditions, and the authors report that high cognitive load slowed responses and altered choice behavior, while anodal tDCS increased risk-taking, information seeking, and switching behavior. The authors report no significant Load × Stimulation interaction and interpret this pattern as evidence that cognitive load and prefrontal stimulation influence decision-making through additive and separable mechanisms.

Overall Assessment

The manuscript addresses an interesting and relevant question concerning how cognitive resource demands and prefrontal neuromodulation shape decision-making under uncertainty. The combination of a cognitive-load manipulation, tDCS over the left DLPFC, and an experience-based decision-making task has potential value, particularly because the authors examine several behavioral outcomes rather than focusing only on risk-taking. The topic is suitable for PLOS ONE, and the study appears to present original empirical work.

However, several important aspects require clarification and revision. The most important issue concerns the description of experimental design and cognitive-load manipulation. Across the manuscript, cognitive load is described both as a between-subjects factor and as a trial-wise or within-participant manipulation, which creates uncertainty about the actual design and whether the statistical analyses are appropriate. This issue should be resolved across the Methods, Results, and Discussion.

A second important concern is that the statistical approach and interpretation of the findings need to be aligned more carefully. The manuscript reports different analytic approaches, including factorial ANOVAs and mixed ANOVAs, but it is not always clear which analysis should be considered primary. Some effects, especially those related to cognitive load, appear to vary across analyses. Therefore, the authors should distinguish between robust findings and findings that are dependent on analysis.

Finally, the interpretation of the tDCS findings and the absence of interactions should be softened. The results may be consistent with additive effects, but a non-significant interaction does not by itself provide strong evidence for independent or separable mechanisms. Similarly, behavioral effects following tDCS should not be interpreted too strongly as direct evidence for specific neural or cognitive mechanisms without additional neural, computational, or mediation evidence.

Overall, the study has promising research questions and potentially interesting findings, but the manuscript requires some clarification of the design, statistical analyses, and interpretation in order the conclusions to be relevant and fully supported by the data.

Recommendations for the authors (major comments)

1. Clarify experimental design and cognitive load manipulation.

The manuscript describes the study as a 2 × 2 between-subjects design with Stimulation and Cognitive Load as between-subject factors. However, cognitive load is also described as trial-wise, within-participant, and in one section as including both low-load and high-load trivia questions, as well as a no-load condition. This creates substantial ambiguity about the actual design.

Recommendation: Please clarify whether cognitive load was manipulated between participants, within participants, or both. The abstract, Methods, Results, and Discussion should use consistent terminology. If participants were assigned to one load condition only, the manuscript should avoid describing load as a within-participant manipulation. If both load levels were presented within participants, the design and analysis should be revised accordingly.

2. Align the statistical analyses with the experimental design

The Results report analyses treating Load both as a between-subjects factor and as a within-subject factor. This makes it difficult to evaluate whether the statistical models are appropriate. It is also unclear which analysis is intended as the primary analysis, and some effects appear significant in one model but not in another.

Recommendation: Please revise the statistical analysis section so that the models match the actual design. The authors should clearly identify the primary analysis for each outcome, report consistent sample sizes/degrees of freedom, and distinguish robust findings from effects that are analysis-dependent. Given the repeated-trial structure and choice outcomes, the authors may also consider mixed-effects models as a more appropriate analytic approach.

3. Clarify the theoretical logic behind the hypotheses, especially the expected absence of an interaction.

The introduction presents cognitive load and anodal tDCS as opposing influences on exploration/exploitation behavior. Cognitive load is expected to reduce risk-taking and exploration, whereas anodal DLPFC stimulation is expected to increase exploratory behavior. However, the rationale for expecting additive effects rather than a Load × Stimulation interaction is not sufficiently developed.

Recommendation: Please clarify why cognitive load and tDCS are hypothesized to operate through partially independent mechanisms. If tDCS could compensate for load-related disruption, then an interaction would also be theoretically plausible. The manuscript should present these alternative possibilities more clearly before stating H3.

4. Clarify how the behavioral measures map onto the theoretical constructs

Recommendation: Please define more clearly what each dependent measure is intended to capture. The authors should explain why Option B indexes risk-taking, why Option R indexes information seeking, and why alternation rate is interpreted as exploratory switching. This would strengthen the connection between the task, hypotheses, and interpretation of the findings.

5. Soften causal claims regarding tDCS and prefrontal excitability.

The manuscript uses strong causal language when describing tDCS and prefrontal excitability, including claims that the study provides a causal or mechanistic account of how prefrontal excitability shapes adaptive behavior. Although tDCS provides an experimental manipulation, its behavioral effects can vary depending on stimulation parameters, task demands, individual differences, and target specificity.

Recommendation: Please soften the causal interpretation of the tDCS findings. Rather than stating that the study directly tests how prefrontal excitability shapes adaptive behavior, the authors could state that it experimentally examines whether anodal tDCS over the left DLPFC modulates decision-making behavior under cognitive load. The manuscript should also acknowledge that behavioral effects following tDCS cannot be attributed to a specific cognitive or neural mechanism without additional neural or computational evidence.

6. Figures related recommendation.

The figures should be improved for readability and consistency. In several places, the text and numerical values appear blurry or difficult to read. I recommend increasing the figure resolution and ensuring that all axis labels, legends, numbers, and annotations are clearly visible. For figures with multiple result panels or boxes, it would also be helpful to add short titles or labels explaining what each panel/box shows. In addition, please use consistent font, font size, scaling, and subscale presentation across all figures.

Minor recommendations for the authors (minor comments).

1. The statements on lines 47–49 would benefit from supporting references.

2. Please consider adding references to support the statement on lines 50-51 and softening the causal wording, as the cognitive effects of anodal tDCS can be task- and protocol-dependent.

3. “Studies that used the DLPFC” is awkward. Better: “Studies targeting the DLPFC…” (line 52)

4. The research questions and hypotheses are relevant, but the paragraph presenting them is somewhat difficult to follow. The wording shifts between questions and statements, and the manuscript uses future tense, such as “our research will,” even though the study has already been conducted. This reduces the clarity and professionalism of the introduction.

5. Please clarify the blinding procedure. The manuscript refers to a single-blind design and states that the experimenter was blind to stimulation condition, while the sham procedure is also described as maintaining participant blinding. Please specify exactly who was blinded and whether blinding success was assessed.

6. The sham procedure is described more clearly in the tDCS section, but the wording in the Study Design section is unclear. Please harmonize these descriptions.

7. The term “stanzas” is unusual in this context. Consider replacing it with blocks, rounds, or trial blocks throughout the Results section.

8. Please clarify whether the position of the safe, risky, and information-seeking options was fixed or counterbalanced. If positions were fixed throughout the task, the authors should briefly acknowledge the possibility of position or key-response biases.

9. Some wording in the Background and Discussion should be revised for clarity. For example, “must be learned feedback” should be “must be learned through feedback,” and “systemic modifications in risk-taking behavior” could be revised to “systematic changes in risk-taking behavior.”

10. The conclusion describes the task as a “repeated binary-choice paradigm,” but the task appears to include safe, risky, and information-seeking options. Please revise this wording so that it accurately reflects the task structure.

11. Please avoid definitive wording such as “clearly indicates,” “robust and consistent,” or “demonstrated” when the findings are mixed across analyses. More cautious wording would better match the reported results.

12. The title is clear and informative, but the wording “independently shape” may overstate the conclusion. Since independence is inferred mainly from non-significant interaction effects, a more cautious title would be preferable. Maybe skip “Independently”

13. The abstract should describe the cognitive-load manipulation consistently with the Methods. It currently states that the study used a 2 × 2 between-subjects design, but also says that cognitive load was manipulated “on a trial-wise basis.” This should be clarified to avoid confusion about whether cognitive load was manipulated between participants, within participants, or both.

14. The final sentence of the abstract should be softened. Claims about “prefrontal control dynamics,” “cortical excitability,” and “independent mechanisms” go beyond what can be directly concluded from behavioral outcomes and non-significant interaction effects. A more cautious formulation would be that the findings are “consistent with additive effects of cognitive load and anodal tDCS on decision-making behavior.”

15. Please consider replacing “information searching” with “information seeking” throughout the abstract and title for consistency with the rest of the manuscript.

16. Some reported F-values seem surprisingly large relative to the reported means and standard deviations, especially for risk-taking. For example, the risk-taking main effect of stimulation is reported as F(1,56) = 331.450 with ηp² = .855, despite group means of 0.54 vs. 0.31 and SDs of 0.14 vs 0.06. This may be correct depending on the exact model/data structure, but it should be checked carefully.

Reviewer #2: The prefrontal cortex (PFC) is related to two different neurocomputational frameworks (Resource-limitation vs Instance-based Learning), but the overt behavioral result cannot be dissociated without direct manipulation of cortical activity. In the present paper, the authors explore the effect of transcranial Direct current stimulation (tDCS) on the strategies employed by human participants to resolve a decision making task under two different levels of cognitive load. In addition, they manipulate the excitability of PFC thought the tDCS intervention. With this approach, they aim to provide support for the hypothesis that these two effects are either independent, and the overt behavior is the result of two separate cortical mechanisms, or that they interact at the level of the PFC. The results of the study suggest that two independent mechanisms coexist in the PFC.

Overall, the paper is well written, the experimental design is sound, and the presented evidence is compelling. However, keeping track of the aims, context, and interpretation of the results was challenging due to the way the manuscript is organized. I believe the paper would benefit from a major rewrite and reorganization of ideas to better present the motivation of each analysis, and the how the results support not only a hypothesis but the underlying theory from which the hypothesis arises.

I think the authors can make the case for a very specific and meaningful contribution, but in the current state of the manuscript, such contributions is presented in a way that makes it underwhelming.

I elaborate on some specific point below:

Main comments:

1) Sections "Introduction" and 'Background". It feels that these two sections contain redundant information, very summarized on the introduction, that is then re-exposed and with more elaboration in the background section. Specific hypotheses are presented only in the background section (very useful and key piece of the puzzle). I would suggest to consolidate these two sections to eliminate redundancies and make the exposition of hypothesis more concise.

2) Mixed design analysis (lin3 291). The first issue that came to my mind was that "Cognitive Load" should have been considered a within-subjects factor. After presenting analysis including Load as a between-subjects factor in the model, the authors repeated the analysis now considering (as is appropriate) Load as a within-subjects factor. I do not see why it was necessary to do the analysis with Load as between subjects factor, as the authors do not clearly explain their rationale or motivation to do so. If feels more as if they were throwing any model they could think post hoc of at the data instead of using a model that was chosen a priori.

Minor comments:

3) In the results section, it would be much better to refer to the hypotheses explicitly, rather than writing "H1", etc. For example, lines 323-324.

4) This is also the case for references to "Option B", "Option R", etc. Readability would greatly improve if these instances read explicitly as "Risky Choice" or "Information-seeking choice" in the text and figures.

**********

what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #1: Yes: Olgerta Asko

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

PLoS One. 2026 Oct 9;21(10):e0360059. doi: 10.1371/journal.pone.0360059.r002

Author response to Decision Letter 1

26 Aug 2026

Reviewer #1

General response: We thank the reviewer for recognizing the relevance and originality of the study and for identifying the need to clarify the design, analytical hierarchy, and interpretation. The revision was organized around these three concerns. The title and abstract were softened, the Methods now separate the primary and secondary manipulations, the Results identify the primary factorial analyses before the secondary mixed analyses, and the Discussion now emphasizes outcome-specific additive and context-dependent patterns rather than a uniform claim of independence.

Reviewer #1 - Comment Major 1

Clarify experimental design and cognitive load manipulation. The manuscript describes the study as a 2 × 2 between-subjects design with Stimulation and Cognitive Load as between-subject factors. However, cognitive load is also described as trial-wise, within-participant, and in one section as including both low-load and high-load trivia questions, as well as a no-load condition. This creates substantial ambiguity about the actual design.

Response: We sincerely thank the reviewer for identifying this ambiguity. We revised the manuscript to distinguish two different manipulations that had previously been described using the same term. The primary between-subject factor is now termed Cognitive-Load Condition and compares Trivia-Based Load with No-Trivia Control. Within the Trivia-Based Load condition only, trivia items were classified as Lower-Load and Higher-Load Trivia; this is now termed Trial-Level Cognitive Load and is treated as a secondary within-participant factor. This distinction is used consistently throughout the Abstract, Methods, Results, Discussion, and Conclusion. We believe this revision addresses the reviewer's concern and improves the clarity and scientific rigor of the manuscript.

Revision Made: Abstract; Section 3.2 (Study design); Section 3.5 (Trivia-based cognitive-load manipulation); Section 3.7 (Statistical analysis); Section 4.6; and Discussion.

Revised Text:

“The primary experiment followed a 2 × 2 between-subjects factorial design with Stimulation (anodal tDCS vs sham) and Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control) as between-subject factors. ... Within the Trivia-Based Load condition, Trial-Level Cognitive Load was manipulated within participants by presenting 25 Lower-Load and 25 Higher-Load Trivia questions in randomized order.”

Reviewer #1 - Comment Major 2

Align the statistical analyses with the experimental designThe Results report analyses treating Load both as a between-subjects factor and as a within-subject factor. This makes it difficult to evaluate whether the statistical models are appropriate. It is also unclear which analysis is intended as the primary analysis, and some effects appear significant in one model but not in another. Recommendation: Please revise the statistical analysis section so that the models match the actual design. The authors should clearly identify the primary analysis for each outcome, report consistent sample sizes/degrees of freedom, and distinguish robust findings from effects that are analysis dependent. Given the repeated-trial structure and choice outcomes, the authors may also consider mixed-effects models as a more appropriate analytic approach.

Response: We thank the reviewer for this important recommendation. The statistical hierarchy has been rewritten. The primary analyses are separate 2 × 2 factorial ANOVAs testing Stimulation × Cognitive-Load Condition for each outcome. The secondary analyses included the 30 participants assigned to the Trivia-Based Load condition and examined Trial-Level Cognitive Load and Trial Block as within-subject factors, with Stimulation as a between-subject factor. Sample sizes and degrees of freedom are now stated explicitly. We also distinguish findings from the primary and secondary analyses in the Results and Discussion. Because the primary outcomes were participant-level summary measures, we retained the ANOVA framework, while acknowledging trial-level mixed-effects models as an important extension for future work. We believe this revision addresses the reviewer's concern and improves the clarity and scientific rigor of the manuscript.

Revision Made: Section 3.7.1 (Primary between-subjects analyses), Section 3.7.2 (Secondary trial-level mixed analyses), Section 4.1 (overview), Section 4.6, Discussion.

Revised Text:

“The primary analysis consisted of a separate 2 × 2 factorial ANOVA for each dependent measure, with Stimulation (anodal tDCS vs sham) and Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control) as between-subject factors. ... Secondary analyses were restricted to the 30 participants assigned to the Trivia-Based Load condition. For each outcome, a 2 (Trial-Level Cognitive Load: Lower-Load Trivia vs Higher-Load Trivia) × 5 (Trial Block: five consecutive blocks of ten trials) × 2 (Stimulation: anodal tDCS vs sham) mixed-design repeated-measures ANOVA was conducted.”

Reviewer #1 - Comment Major 3

Clarify the theoretical logic behind the hypotheses, especially the expected absence of an interaction. The introduction presents cognitive load and anodal tDCS as opposing influences on exploration/exploitation behavior. Cognitive load is expected to reduce risk-taking and exploration, whereas anodal DLPFC stimulation is expected to increase exploratory behavior. However, the rationale for expecting additive effects rather than a Load × Stimulation interaction is not sufficiently developed. Recommendation: Please clarify why cognitive load and tDCS are hypothesized to operate through partially independent mechanisms. If tDCS could compensate for load-related disruption, then an interaction would also be theoretically plausible. The manuscript should present these alternative possibilities more clearly before stating H3.

Response: We sincerely thank the reviewer. We revised the theoretical framing before H3 to present both plausible patterns rather than treating additivity as the only prediction. The revised Background now contrasts a partially independent account, which predicts broadly additive effects, with a compensatory/context-dependent account, in which stimulation could alter the behavioral consequences of cognitive demand. H3 was correspondingly rewritten to state the additive prediction more cautiously. We believe this revision addresses the reviewer's concern and improves the clarity and scientific rigor of the manuscript.

Revision Made: Background, paragraph immediately preceding H3 and H3.

Revised Text:

“Although cognitive load and anodal DLPFC stimulation were expected to exert opposing directional influences on some outcomes, two theoretical patterns were possible. Under a partially independent account, cognitive load would primarily constrain working-memory and attentional resources, whereas anodal stimulation would modulate prefrontal contributions to exploratory control, producing broadly additive effects. Alternatively, if stimulation strengthens the same processes disrupted by cognitive load, its behavioral effect could differ across Cognitive-Load Conditions, producing a compensatory or context-dependent interaction. ... H3. Anodal DLPFC stimulation will increase exploratory and risk-related behavior in both the Trivia-Based Load and No-Trivia Control conditions. Under the partially independent account, the magnitude of the stimulation effect is expected to be broadly comparable across Cognitive-Load Conditions, producing an additive pattern rather than systematic moderation by cognitive load.”

Reviewer #1 - Comment Major 4

Reviewer comment: Clarify how the behavioral measures map onto the theoretical constructs

The manuscript interprets risky choices, information-seeking choices, and alternation rates as indicators of exploration, curiosity, cognitive flexibility, or exploratory control. However, these measures may reflect partly different processes. For example, information seeking may reflect curiosity, strategic uncertainty reduction, or difficulty committing to a choice, while alternation may reflect exploration but also instability or weak value learning. Recommendation: Please define more clearly what each dependent measure is intended to capture. The authors should explain why Option B indexes risk-taking, why Option R indexes information seeking, and why alternation rate is interpreted as exploratory switching. This would strengthen the connection between the task, hypotheses, and interpretation of the findings.

Response: We thank the reviewer for this helpful recommendation. We now define each behavioral measure operationally and explicitly state that the measures are related but not interchangeable. The Background explains the theoretical interpretation of risky choice, costly information seeking, and alternation, while the Methods provides formal outcome definitions. The Discussion also acknowledges alternative interpretations of information seeking and alternation. We believe this revision addresses the reviewer's concern and improves the clarity and scientific rigor of the manuscript.

Revision Made: Background; Section 3.6 (Behavioral outcome measures); Discussion.

Revised Text:

“Within the present task, selection of the probabilistic high-reward alternative was treated as an index of risk-taking because participants accepted the possibility of receiving no reward in exchange for a substantially larger potential gain than the guaranteed safe option. Selection of the costly reveal alternative was treated as an index of information seeking because participants paid one point to obtain otherwise unavailable probability information before making their final choice. Alternation rate represented exploratory switching because it measured changes between choice types across successive trials; however, higher alternation could also reflect weaker commitment or unstable value learning.”

Reviewer #1 - Comment Major 5

Soften causal claims regarding tDCS and prefrontal excitability. The manuscript uses strong causal language when describing tDCS and prefrontal excitability, including claims that the study provides a causal or mechanistic account of how prefrontal excitability shapes adaptive behavior. Although tDCS provides an experimental manipulation, its behavioral effects can vary depending on stimulation parameters, task demands, individual differences, and target specificity. Recommendation: Please soften the causal interpretation of the tDCS findings. Rather than stating that the study directly tests how prefrontal excitability shapes adaptive behavior, the authors could state that it experimentally examines whether anodal tDCS over the left DLPFC modulates decision-making behavior under cognitive load. The manuscript should also acknowledge that behavioral effects following tDCS cannot be attributed to a specific cognitive or neural mechanism without additional neural or computational evidence.

Response: We sincerely thank the reviewer. We removed statements that treated the behavioral findings as direct evidence of altered cortical excitability or a specific neural mechanism. The Introduction and Background now describe tDCS as an experimental manipulation whose effects are task- and protocol-dependent. The Discussion and Conclusion explicitly state that the behavioral results do not identify a specific neural or computational mechanism, and the limitations note that the study did not include concurrent neural measurements. Future work is proposed using EEG and fNIRS to examine the neural correlates of cognitive-load and stimulation-related differences. We believe this revision addresses the reviewer's concern and improves the clarity and scientific rigor of the manuscript.

Revision Made: Introduction; Background; Section 3.3; Discussion; Conclusion.

Revised Text:

“Transcranial direct current stimulation (tDCS) provides an experimental method for examining whether modulation over the DLPFC is associated with changes in decision behavior, although the direction and magnitude of its behavioral effects depend on stimulation parameters, task demands, and individual differences. The present behavioral data therefore support a mixed account in which cognitive-load demands and anodal DLPFC tDCS can have both additive and interactive consequences depending on the component of decision behavior being measured.”

Reviewer #1 - Comment Major 6

Figures related recommendation.

Response: We thank the reviewer for this helpful recommendation. In response, the figures have been revised for readability and consistency, with clearer panel labeling, standardized terminology, consistent fonts and scaling, and improved presentation of axis labels, legends, numerical values, and error bars. We believe these revisions address the reviewer’s concern by improving the readability, consistency, and presentation of the figures and by ensuring that the final high-resolution versions are provided as separate figure files for submission.

Revision Made: Figures 1-7 and their captions were revised for readability and consistency. Figures were standardized with consistent terminology, fonts, scaling, axis labels. The final upgraded resolution figure files will be submitted separately with the revised manuscript.

Revised Text:

“Fig 5. Effects of stimulation and cognitive-load condition on risky choice. (A) Mean proportion of risky choices under Sham and tDCS stimulation. (B) Mean proportion of risky choices in the No-Trivia Control and Trivia-Based Load conditions. Error bars represent 95% confidence intervals”.

Reviewer #1 - Comment Minor 1

The statements on lines 47-49 would benefit from supporting references.

Response: We thank the reviewer. Supporting citations were added to the statements describing DLPFC contributions to goal maintenance, feedback integration, working memory, and attentional allocation.

Revision Made: Introduction and Background.

Revised Text:

“Concurrent cognitive load can limit working-memory and attentional resources, whereas the dorsolateral prefrontal cortex (DLPFC) supports goal maintenance, feedback integration, and flexible adjustment of behavior [2,9-13].”

Reviewer #1 - Comment Minor 2

Please consider adding references to support the statement on lines 50-51 and softening the causal wording, as the cognitive effects of anodal tDCS can be task- and protocol-dependent.

Response: We thank the reviewer. We added supporting references and revised the text to state explicitly that tDCS effects vary with stimulation parameters, task demands, and individual differences.

Revision Made: Introduction and Background.

Revised Text:

Reviewer #1 - Comment Minor 3

“Studies that used the DLPFC” is awkward. Better: “Studies targeting the DLPFC…” (line 52)

Response: We thank the reviewer. The sentence was rewritten and the broader paragraph was edited for clarity.

Revision Made: Background.

Revised Text:

“Anodal tDCS over the DLPFC has been associated with changes in working memory, risk-related choice, and exploratory behavior, but these effects are not uniform across tasks or stimulation protocols [15-21,24].”

Reviewer #1 - Comment Minor 4

The research questions and hypotheses are relevant, but the paragraph presenting them is somewhat difficult to follow. The wording shifts between questions and statements, and the manuscript uses future tense, such as “our research will,” even though the study has already been conducted. This reduces the clarity and professionalism of the introduction.

Response: We thank the reviewer. We reorganized the Introduction and Background, removed future-tense wording, and presented the study objective and hypotheses as completed-study statements.

Revision Made: Introduction and Background.

Revised Text:

Reviewer #1 - Comment Minor 5

Please clarify the blinding procedure. The manuscript refers to a single-blind design and states that the experimenter was blind to stimulation condition, while the sham procedure is also described as maintaining participant blinding. Please specify exactly who was blinded and whether blinding success was assessed.

Response: We thank the reviewer. The study used a single-blind design: participants were unaware of their active or sham assignment, whereas the experimenter knew the stimulation condition.

Revision Made: Section 3.2 (Study design and blinding).

Revised Text:

“Participants were not informed of their active or sham assignment, and the sham ramp procedure was used to support participant blinding. The experimenter was aware of the stimulation condition during testing. Thus, the study followed a single-blind design. Blinding success was not formally assessed.”

Reviewer #1 - Comment Minor 6

The term “stanzas” is unusual in this context. Consider replacing it with blocks, rounds, or trial blocks throughout the Results section.

Response: We thank the reviewer. The term “stanzas” has been removed throughout and replaced with “Trial Block,” defined as five consecutive blocks of ten trials.

Revision Made: Section 3.7.2 and Section 4.6.

Revised Text:

“Trial Block: five consecutive blocks of ten trials.”

Reviewer #1 - Comment Minor 7

Please clarify whether the position of the safe, risky, and information-seeking options was fixed or counterbalanced. If positions were fixed throughout the task, the authors should briefly acknowledge the possibility of position or key-response biases.

Response: We thank the reviewer. The Methods now states that option positions and response keys were fixed, and the Discussion acknowledges this as a limitation.

Revision Made: Section 3.4 and Discussion limitations.

Revised Text:

“The spatial positions and response keys for the safe, risky, and information-seeking choices remained fixed throughout the task to maintain a consistent response mapping. ... The positions and response keys for the three choices were fixed, which may have introduced position or key-response biases.”

Reviewer #1 - Comment Minor 8

Some wording in the Background and Discussion should be revised for clarity. For example, “must be learned feedback” should be “must be learned through feedback,” and “systemic modifications in risk-taking behavior” could be revised to “systematic changes in risk-taking behavior.”

Response: We thank the reviewer. The Background and Discussion were substantially rewritten for clarity, and the awkward phrases identified by the reviewer were removed.

Revision Made: Background and Discussion.

Revised Text:

“In experience-based decision making, individuals repeatedly sample outcomes and use feedback to decide whether to continue exploiting a familiar option or explore an uncertain alternative [1,2,4,5].”

Reviewer #1 - Comment Minor 9

The conclusion describes the task as a “repeated binary-choice paradigm,” but the task appears to include safe, risky, and information-seeking options. Please revise this wording so that it accurately reflects the task structure.

Response: We thank the reviewer. The Conclusion now correctly describes the paradigm as a repeated three-option decision-from-experience task.

Revision Made: Conclusion.

Revised Text:

“The present study combined anodal tDCS over the left DLPFC with a trivia-based cognitive-load manipulation in a repeated three-option decision-from-experience task.”

Reviewer #1 - Comment Minor 10

Please avoid definitive wording such as “clearly indicates,” “robust and consistent,” or “demonstrated” when the findings are mixed across analyses. More cautious wording would better match the reported results.

Response: We thank the reviewer. Definitive language was removed from the Results, Discussion, and Conclusion and replaced with more cautious, outcome-specific interpretation.

Revision Made: Results, Discussion, and Conclusion.

Revised Text:

“The findings are therefore most consistent with outcome-specific modulation, including both additive and context-dependent effects.”

Reviewer #1 - Comment Minor 11

The title is clear and informative, but the wording “independently shape” may overstate the conclusion. Since independence is inferred mainly from non-significant interaction effects, a more cautious title would be preferable. Maybe skip “Independently”

Response: We thank the reviewer and agree. The word “Independently” was removed from the title.

Revision Made: Title.

Revised Text:

“Cognitive load and prefrontal neuromodulation shape exploration, risk-taking, and information seeking in experience-based decision making”

Reviewer #1 - Comment Minor 12

The abstract should describe the cognitive-load manipulation consistently with the Methods. It currently states that the study used a 2 × 2 between-subjects design but also says that cognitive load was manipulated “on a trial-wise basis.” This should be clarified to avoid confusion about whether cognitive load was manipulated between participants, within participants, or both.

Response: We thank the reviewer. The Abstract now distinguishes the primary between-subject Cognitive-Load Condition from the secondary within-participant Trial-Level Cognitive Load manipulation.

Revision Made: Abstract.

Revised Text:

“The primary 2 × 2 between-subjects design crossed Stimulation (anodal tDCS vs sham) with Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control). Within the Trivia-Based Load subgroup, Lower-Load and Higher-Load Trivia were presented across trials for a secondary trial-level analysis.”

Reviewer #1 - Comment Minor 13

The final sentence of the abstract should be softened. Claims about “prefrontal control dynamics,” “cortical excitability,” and “independent mechanisms” go beyond what can be directly concluded from behavioral outcomes and non-significant interaction effects. A more cautious formulation would be that the findings are “consistent with additive effects of cognitive load and anodal tDCS on decision-making behavior.”

Response: We thank the reviewer. The final sentence of the Abstract was rewritten to avoid mechanistic and independence claims and to reflect the observed mixture of additive and context-dependent patterns.

Revision Made: Abstract.

Revised Text:

“These findings show that cognitive load and anodal DLPFC tDCS have outcome-specific joint effects on experience-based decision making, including both additive and context-dependent patterns.”

Reviewer #1 - Comment Minor 14

Please consider replacing “information searching” with “information seeking” throughout the abstract and title for consistency with the rest of the manuscript.

Response: We thank the reviewer. “Information seeking” is now used consistently throughout the manuscript.

Revision Made: Title, Abstract, Methods, Results, Discussion, and figure-caption terminology.

Revised Text:

“Participants completed 50 decisions involving a safe choice, a probabilistic high-reward choice, and a costly information-seeking choice.”

Reviewer #1 - Comment Minor 15

Some reported F-values seem surprisingly large relative to the reported means and standard deviations, especially for risk-taking. For example, the risk-taking main effect of stimulation is reported as F(1,56) = 331.450 with ηp² = .855, despite group means of 0.54 vs. 0.31 and SDs of 0.14 vs 0.06. This may be correct depending on the exact model/data structure, but it should be checked carefully.

Response: We thank the reviewer for prompting us to verify this result. We rechecked the participant-level risky-choice analysis, including the group coding, descriptive statistics, ANOVA output, and effect-size estimate. Because one participant had an incomplete choice record, the final risky-choice analysis used 59 complete cases. The verified stimulation effect is F(1,55) = 320.075, p < .001, ηp² = .853, as reported consistently in Section 4.3.

Revision Made: Section 4.3 was rechecked for consistency with the statistical output.

Reviewer #2

General response: We sincerely thank the reviewer for the positive assessment and for emphasizing the need to improve the organization and theoretical narrative. We substantially reorganized the Introduction, Background, Statistical Analysis, Results, and Discussion so that the rationale for each analysis and its relationship to the theoretical alternatives are explicit. We retained separate Introduction and Background headings, but shortened the Introduction and refocused the Background on the theoretical framework, operational definitions, and hypotheses to remove redundancy.

Reviewer #2 - Comment Main 1

Sections "Introduction" and 'Background'. It feels that these two sections contain redundant information, very summarized on the introduction, that is then re-exposed and with more elaboration in the background section. Specific hypotheses are presented only in the background section (very useful and key piece of the puzzle). I would suggest to consolidate these two sections to eliminate redundancies and make the exposition of hypothesis more concise.

Response: We sincerely thank the reviewer. We agreed that the original Introduction and Background repeated the same rationale. Rather than removing the Background heading entirely, we substantially reorganized both sections: the Introduction now gives a concise overview of the problem, research gap, design, and study objective, whereas the Background is focused on the decision-from-experience framework, the behavioral constructs, the resource-limitation and neuromodulation accounts, and the hypotheses. This preserves the useful theoretical detail while minimizing repetition. We believe this revision addresses the reviewer's concern and improves the clarity and scientific rigor of the manuscript.

Revision Made: Introduction and Background were rewritten and shortened, with hypotheses retained in the Background.

Revised Text:

“The objective was to determine whether anodal tDCS over the left DLPFC and trivia-based cognitive demand were associated with additive or context-dependent differences in experience-based decision making. The following Background section develops the theoretical basis for the behavioral measures and hypotheses while avoiding repetition of the general study rationale presented above.”

Reviewer #2 – Major comment 2

Mixed design analysis (lin3 291). The first issue that came to my mind was that "Cognitive Load" should have been considered a within-subjects factor. After presenting analysis including Load as a between-subjects factor in the model, the authors repeated the analysis now considering (as is appropriate) Load as a within-subjects factor. I do not see why it was necessary to do the analysis with Load as between subjects factor, as the authors do not clearly explain their rationale or motivation to do so. If feels more as if they were throwing any model, they could think post hoc of at the data instead of using a model that was chosen a priori.

Response: We thank the reviewer for identifying the source of confusion. The original manuscript used the same label, “Cognitive Load,” for two distinct manipulations. We now make clear that the between-subject factor is Cognitive-Load Condition (Trivia-Based Load vs No-Trivia Control), whereas the within-subject factor exists only within the Trivia-Based Load subgroup and represents Trial-Level Cognitive Load (Lower-Load vs Higher-Load Trivia). The primary 2 × 2 factorial ANOVA tests whether the presence of concurrent trivia changes behavior; the secondary 2 × 5 × 2 mixed analysis tests whether trivia difficulty and trial block produce additional within-participant variation among the 30 trivia participants.

Revision Made: Section 3.2; Section 3.7.1; Section 3.7.2; Section 4.1; Section 4.6; Discussion.

Revised Text: “These analyses were considered secondary because they examined differences in trivia difficulty within the Trivia-Based Load condition rather than the primary comparison between the Trivia-Based Load and No-Trivia Control conditions.”

Reviewer #2 - Comment Minor 3

In the results section, it would be much better to refer to the hypotheses explicitly, rather than writing "H1", etc. For example, lines 323-324.

Response: We thank the reviewer. The revised Results no longer relies on shorthand statements such as “H1 supported” or “H2 supported.” Instead, each subsection states the substantive prediction and describes whether the observed outcome was consistent with it.

Revision Made: Results Sections 4.2-4.5 and Discussion.

Revised Text:

“This outcome supported the predicted stimulation-related increase and load-related reduction in risky choice, with an additive rather than moderated pattern.”

Reviewer #2 - Comment Minor 4

This is also the case for references to "Option B", "Option R", etc. Readability would greatly improve if these instances read explicitly as "Risky Choice" or "Information-seeking choice" in the text and figures.

Response: We thank the reviewer. We now use “Risky choice” and “Information-seeking choice” throughout the Results, Discussion, table, and figure-caption terminology. The option labels are retained only in the Methods where the task alternatives are formally defined.

Revision Made: Section 3.4; Sections 4.3-4.4; Discussion; figure-caption terminology.

Revised Text: “Risk-taking was higher under anodal tDCS ... than under sham stimulation .... ... Information seeking was higher overall under anodal tDCS ... than under sham stimulation.”

Attachment

Submitted filename: Response to Reviewers.docx

pone.0360059.s001.docx^{(41.1KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0360059.r003

Decision Letter 1

Xianwei Che

22 Sep 2026

<p>Cognitive load and prefrontal neuromodulation shape exploration, risk-taking, and information seeking in experience-based decision making

PONE-D-26-16597R1

Dear Dr. Rao,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Xianwei Che

Academic Editor

PLOS One

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions??>

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: No

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

Reviewer #2: Yes

**********

Reviewer #1: (No Response)

Reviewer #2: The authors have addressed my comments to satisfaction.

At this point, I have no further questions or comments.

**********

what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #1: Yes: Olgerta Asko

Reviewer #2: No

**********

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Attachment

Submitted filename: Response to Reviewers.docx

pone.0360059.s001.docx^{(41.1KB, docx)}

Data Availability Statement

The anonymized behavioral data underlying the findings reported in this study are publicly available in the following Figshare repository: https://doi.org/10.6084/m9.figshare.33317310.

[pone.0360059.ref001] 1.Daw ND, O’Doherty JP, Dayan P, Seymour B, Dolan RJ. Cortical substrates for exploratory decisions in humans. Nature. 2006;441(7095):876–9. doi: 10.1038/nature04766 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref002] 2.Cohen JD, McClure SM, Yu AJ. Should I stay or should I go? How the human brain manages the trade-off between exploitation and exploration. Philos Trans R Soc Lond B Biol Sci. 2007;362(1481):933–42. doi: 10.1098/rstb.2007.2098 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref003] 3.Wilson RC, Geana A, White JM, Ludvig EA, Cohen JD. Humans use directed and random exploration to solve the explore-exploit dilemma. J Exp Psychol Gen. 2014;143(6):2074–81. doi: 10.1037/a0038199 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref004] 4.Hertwig R, Erev I. The description-experience gap in risky choice. Trends Cogn Sci. 2009;13(12):517–23. doi: 10.1016/j.tics.2009.09.004 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref005] 5.Rakow T, Newell BR. Degrees of uncertainty: An overview and framework for future research on experience-based choice. J Behav Decis Mak. 2010;23(1):1–14. doi: 10.1002/bdm.681 [DOI] [Google Scholar]

[pone.0360059.ref006] 6.Dutt V, Gonzalez C. The role of inertia in modeling decisions from experience with instance-based learning. Front Psychol. 2012;3:177. doi: 10.3389/fpsyg.2012.00177 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref007] 7.Gonzalez C, Dutt V. Instance-based learning: Integrating sampling and repeated decisions from experience. Psychol Rev. 2011;118(4):523–51. doi: 10.1037/a0024558 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref008] 8.Gonzalez C, Lerch JF, Lebiere C. Instance-based learning in dynamic decision making. Cogn Sci. 2003;27(4):591–635. doi: 10.1207/s15516709cog2704_2 [DOI] [Google Scholar]

[pone.0360059.ref009] 9.Miller EK, Cohen JD. An integrative theory of prefrontal cortex function. Annu Rev Neurosci. 2001;24:167–202. doi: 10.1146/annurev.neuro.24.1.167 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref010] 10.Koechlin E. Prefrontal executive function and adaptive behavior in complex environments. Curr Opin Neurobiol. 2016;37:1–6. doi: 10.1016/j.conb.2015.11.004 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref011] 11.Sweller J. Cognitive load during problem solving: Effects on learning. Cogn Sci. 1988;12(2):257–85. doi: 10.1207/s15516709cog1202_4 [DOI] [Google Scholar]

[pone.0360059.ref012] 12.Kahneman D. Thinking, fast and slow. New York: Farrar, Straus and Giroux; 2011. [Google Scholar]

[pone.0360059.ref013] 13.Badre D, Nee DE. Frontal cortex and the hierarchical control of behavior. Trends Cogn Sci. 2018;22(2):170–88. doi: 10.1016/j.tics.2017.11.005 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref014] 14.Nitsche MA, Paulus W. Sustained excitability elevations induced by transcranial DC motor cortex stimulation in humans. Neurology. 2001;57(10):1899–901. doi: 10.1212/wnl.57.10.1899 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref015] 15.Fregni F, Boggio PS, Nitsche M, Bermpohl F, Antal A, Feredoes E, et al. Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory. Exp Brain Res. 2005;166(1):23–30. doi: 10.1007/s00221-005-2334-6 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref016] 16.Brunoni AR, Vanderhasselt M-A. Working memory improvement with non-invasive brain stimulation of the dorsolateral prefrontal cortex: A systematic review and meta-analysis. Brain Cogn. 2014;86:1–9. doi: 10.1016/j.bandc.2014.01.008 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref017] 17.Fecteau S, Knoch D, Fregni F, Sultani N, Boggio P, Pascual-Leone A. Diminishing risk-taking behavior by modulating activity in the prefrontal cortex: A direct current stimulation study. J Neurosci. 2007;27(46):12500–5. doi: 10.1523/JNEUROSCI.3283-07.2007 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref018] 18.Nitsche MA, Cohen LG, Wassermann EM, Priori A, Lang N, Antal A, et al. Transcranial direct current stimulation: State of the art 2008. Brain Stimul. 2008;1(3):206–23. doi: 10.1016/j.brs.2008.06.004 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref019] 19.Polanía R, Nitsche MA, Ruff CC. Studying and modifying brain function with non-invasive brain stimulation. Nat Neurosci. 2018;21(2):174–87. doi: 10.1038/s41593-017-0054-4 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref020] 20.Dedoncker J, Brunoni AR, Baeken C, Vanderhasselt M-A. A Systematic review and meta-analysis of the effects of transcranial direct current stimulation (tDCS) over the dorsolateral prefrontal cortex in healthy and neuropsychiatric samples: Influence of stimulation parameters. Brain Stimul. 2016;9(4):501–17. doi: 10.1016/j.brs.2016.04.006 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref021] 21.Farhat LC, Carvalho AF, Solmi M, Brunoni AR. Evidence-based umbrella review of cognitive effects of prefrontal tDCS. Soc Cogn Affect Neurosci. 2022;17(1):43–60. doi: 10.1093/scan/nsaa084 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref022] 22.Boggio PS, Campanhã C, Valasek CA, Fecteau S, Pascual-Leone A, Fregni F. Modulation of decision-making in a gambling task in older adults with transcranial direct current stimulation. Eur J Neurosci. 2010;31(3):593–7. doi: 10.1111/j.1460-9568.2010.07080.x [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref023] 23.Sweller J. Cognitive load theory. Psychol Learn Motiv. 2011;55:37–76. doi: 10.1016/B978-0-12-387691-1.00002-8 [DOI] [Google Scholar]

[pone.0360059.ref024] 24.Whitney P, Rinehart CA, Hinson JM. Framing effects under cognitive load: The role of working memory in risky decisions. Psychon Bull Rev. 2008;15(6):1179–84. doi: 10.3758/PBR.15.6.1179 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref025] 25.Evans JSBT, Stanovich KE. Dual-process theories of higher cognition: Advancing the debate. Perspect Psychol Sci. 2013;8(3):223–41. doi: 10.1177/1745691612460685 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref026] 26.Payne JW, Bettman JR, Johnson EJ. The adaptive decision maker. Cambridge: Cambridge University Press; 1993. [Google Scholar]

[pone.0360059.ref027] 27.Otto AR, Skatova A, Madlon-Kay S, Daw ND. Cognitive control predicts use of model-based reinforcement learning. J Cogn Neurosci. 2015;27(2):319–33. doi: 10.1162/jocn_a_00709 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref028] 28.van Steenbergen H, Band GPH, Hommel B. Does conflict help or hurt cognitive control? Initial evidence for an inverted U-shape relationship between perceived task difficulty and conflict adaptation. Front Psychol. 2015;6:974. doi: 10.3389/fpsyg.2015.00974 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref029] 29.Huang Y, Datta A, Bikson M, Parra LC. Realistic volumetric-approach to simulate transcranial electric stimulation-ROAST-a fully automated open-source pipeline. J Neural Eng. 2019;16(5):056006. doi: 10.1088/1741-2552/ab208d [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref030] 30.Knoch D, Gianotti LRR, Pascual-Leone A, Treyer V, Regard M, Hohmann M, et al. Disruption of right prefrontal cortex by low-frequency repetitive transcranial magnetic stimulation induces risk-taking behavior. J Neurosci. 2006;26(24):6469–72. doi: 10.1523/JNEUROSCI.0804-06.2006 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref031] 31.Gruber MJ, Gelman BD, Ranganath C. States of curiosity modulate hippocampus-dependent learning. Neuron. 2014;84(2):486–96. doi: 10.1016/j.neuron.2014.08.060 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref032] 32.Murayama K, FitzGibbon L, Sakaki M. Process account of curiosity and interest: A reward-learning perspective. Educ Psychol Rev. 2019;31:875–95. doi: 10.1007/s10648-019-09499-9 [DOI] [Google Scholar]

[pone.0360059.ref033] 33.Saturnino GB, Puonti O, Nielsen JD, Antonenko D, Madsen KH, Thielscher A. SimNIBS 2.1: A comprehensive pipeline for individualized electric field modelling for transcranial brain stimulation. In: Makarov S, Horner M, Noetscher G. Brain and human body modeling: Computational human modeling at EMBC 2018. Cham: Springer. 2019. p. 3–25. doi: 10.1007/978-3-030-21293-3_1 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref034] 34.Laakso I, Tanaka S, Koyama S, De Santis V, Hirata A. Inter-subject variability in electric fields of motor cortical tDCS. Brain Stimul. 2015;8(5):906–13. doi: 10.1016/j.brs.2015.05.002 [DOI] [PubMed] [Google Scholar]

[pone.0360059.ref035] 35.Herff C, Heger D, Fortmann O, Hennrich J, Putze F, Schultz T. Mental workload during n-back task-quantified in the prefrontal cortex using fNIRS. Front Hum Neurosci. 2014;7:935. doi: 10.3389/fnhum.2013.00935 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0360059.ref036] 36.Sood M, Besson P, Muthalib M, Jindal U, Perrey S, Dutta A, et al. NIRS-EEG joint imaging during transcranial direct current stimulation: Online parameter estimation with an autoregressive model. J Neurosci Methods. 2016;274:71–80. doi: 10.1016/j.jneumeth.2016.09.008 [DOI] [PubMed] [Google Scholar]

PERMALINK

Cognitive load and prefrontal neuromodulation shape exploration, risk-taking, and information seeking in experience-based decision making

Ankit Singh

Varun Dutt

Akash K Rao

Roles

Abstract

1. Introduction

2. Background

3. Methods

3.1. Participants

3.2. Study design

3.3. Transcranial direct current stimulation

Fig 1. Transcranial direct current stimulation protocols.

Fig 2. Electrode montage and modeled voltage and electric-field distributions.

3.4. Decision-making task

Fig 3. Sequence of a decision trial.

3.5. Trivia-based cognitive-load manipulation

3.6. Behavioral outcome measures

3.7. Statistical analysis

3.7.1. Primary between-subjects analyses.

3.7.2. Secondary trial-level mixed analyses.

3.7.3. General statistical considerations.

4. Results

4.1. Primary factorial analyses: Overview

4.2. Reaction time

Fig 4. Effects of cognitive load on reaction time.

4.3. Risky choice

Fig 5. Effects of stimulation and cognitive-load condition on risky choice.

4.4. Information-seeking choice

Fig 6. Effects of stimulation and cognitive-load condition on information-seeking choice.

4.5. Alternation rate

Fig 7. Effects of stimulation, cognitive-load condition, and trial block on alternation rate.

4.6. Secondary trial-level cognitive load and trial block analyses

5. Discussion

6. Conclusion

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

Xianwei Che

Roles

Author response to Decision Letter 1

Decision Letter 1

Xianwei Che

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases