Synergistic effects of aerobic exercise and transcranial direct current stimulation on executive function and biomarkers in healthy young adults

,


Introduction
In recent years, the increasing prevalence of neurodegenerative diseases and age-related executive function decline has posed significant challenges to public health and the quality of life for individuals worldwide (Livingston et al., 2017). Executive function is one of the most advanced cognitive activities that can enable the individual to complete a series of top-down mental processes occurring in maintaining concentration, resisting temptation, contending with the challenge and looking before you leap and is the activity that the individual intentionally monitors thoughts and actions, which exert a crucial effect on everyday life (Diamond, 2013). The above abilities primarily include three subcomponents: working memory, the capacity to update information held in mind; inhibition control, the ability to override entrenched habits or impulses; and cognitive flexibility, the capability to switch between tasks (Devine et al., 2019). The regions of the brain implicated in executive function encompass the frontal lobostriate circuit and the cerebellum. Specifically, the frontal lobostriate circuit includes the dorsolateral prefrontal cortex (DLPFC), the orbital frontal cortex (OFC), the anterior cingulate cortex (ACC), and the basal ganglia (Su et al., 2007). In recent years, research has focused on enhancing executive function by promoting brain plasticity and strengthening functional connections between these regions. In primates, both the DLPFC and OFC, key components of the prefrontal cortex (PFC), are implicated in this process. The DLPFC, in particular, plays a significant role in monitoring thinking and behavior. The OFC is essential in decision-making, particularly in evaluating rewards and punishments. By modulating the function and connectivity of these regions, we have the potential to improve executive functions.
Oxidative stress, defined as an imbalance between the production of reactive oxygen species (ROS) and the body's antioxidant defense system, is thought to play a key role in the pathogenesis of various neurodegenerative diseases, including impairment of executive function (Cobley et al., 2018). Emerging evidence indicates that oxidative stress is closely related to cognitive function regulation, particularly executive function (Hoyos et al., 2022;Foster et al., 2017;Valls-Pedret et al., 2015). Research indicates that ROS can affect neuronal function and synaptic plasticity through various mechanisms, including lipid peroxidation, protein oxidation, DNA damage, mitochondrial dysfunction, calcium dysregulation, inflammation, and impaired neurotransmission Knaus, 2021;Wang et al., 2020;García et al., 2020;Benusa et al., 2017;Santini et al., 2015). Together, these pathways contribute to the executive function deficits and age-related executive function decline seen in neurodegenerative diseases. Despite these findings, the exact nature of the relationship between oxidative stress and executive function remains controversial, necessitating further research. While excessive ROS production has been linked to neurodegeneration and executive function impairment, moderate ROS levels are essential for optimal neuronal growth and synaptic plasticity. In a balanced state, ROS can stimulate the release of neurotrophic factors such as nerve growth factor (NGF), which plays an important role in the growth and differentiation of neuron (Corpas et al., 2017;Massaad and Klann, 2011;Knapp and Klann, 2002).
Among these pathways, the various factors and pathways involved in ROS-induced neurotoxicity, such as brain-derived neurotrophic factor (BDNF), malondialdehyde (MDA), superoxide dismutase (SOD), glutathione peroxidase 4 (GPX4), glutamate, and serum iron ions, play critical roles (Uttara et al., 2009;Shih et al., 2007;Lu, 2003). BDNF is a key neurotrophic factor that promotes neuronal growth, differentiation and survival. Studies have found that oxidative stress can decrease BDNF expression, which in turn impacts neuronal survival and function (Usmani et al., 2023;Horvath et al., 2021;Lin et al., 2018). Therefore, maintaining BDNF levels is critical for maintaining executive function (Usmani et al., 2023). Malondialdehyde (MDA) is a lipid peroxidation product that can be used as a biomarker of oxidative stress. Oxidative stress leads to peroxidation of cell membrane phospholipids, leading to malondialdehyde production (Thangwong et al., 2022). High concentrations of MDA can cause cellular damage and neuronal apoptosis, thus potentially leading to impaired executive function (Yang et al., 2022). SOD and GPX4 are two important antioxidant enzymes that eliminate excess ROS and protect cells from oxidative damage. Oxidative stress can cause a decrease in the activity of SOD and GPX4, reducing the cell's ability to eliminate ROS and exacerbating neurotoxicity (Cobley et al., 2018;Ansari and Scheff, 2010). Glutamate is an excitatory neurotransmitter, and excess glutamate can cause excitotoxicity and cell death. Oxidative stress can increase the release of glutamate, further aggravating neuronal damage (Sabogal-Guáqueta et al., 2019;Zhao et al., 2011). The role of serum iron in neurodegenerative diseases has received much attention. Iron ions participate in the Fenton reaction, catalyzing the generation of highly active hydroxyl radicals and thereby increasing oxidative stress (Thomas et al., 2009). Imbalances in iron ion homeostasis can lead to neuronal damage and cognitive decline (Mueller et al., 2012). While many studies have focused on the individual roles of these factors in executive function and neurodegenerative diseases, few have addressed their intricate interplay and the potential synergistic effects among them (Bostanciklioglu, 2019). Furthermore, the precise mechanisms by which iron ions contribute to neuronal damage and executive function decline are still not fully understood (Ward et al., 2014), and the interplay between glutamate neurotransmission and oxidative stress in the context of executive function has not been thoroughly investigated (Lewerenz and Maher, 2015). Gaining a deeper understanding of these complex interactions, as well as the interplay between oxidative stress and neurotransmission, is critical not only for elucidating the underlying mechanisms of executive dysfunction but also for informing the development of more effective therapeutic strategies.
In recent years, researchers have proposed that multi-modal rehabilitation intervention approaches (i.e., more than one intervention technique) may induce synergistic or additive effects to achieve faster, better, more comprehensive and lasting rehabilitation effects (Menon and D'Esposito, 2022;Ward et al., 2017). such as combining brain stimulation with physical exercise such as aerobic exercise (AE) (Steinberg et al., 2018;Hendrikse et al., 2017;Moreau et al., 2015), specifically within the cognitive domain (Steinberg et al., 2018). At the same time, these multi-modal rehabilitation techniques can also improve the oxidative stress of the brain, improve neurotransmitter metabolism, and thus improve executive function.
Transcranial direct current stimulation (tDCS) is a noninvasive and safe method for modulating neuronal activity, leading to changes in neural plasticity. tDCS applied to the PFC can modulate cortical excitability in the stimulated area, while strengthening the connections between neuronal synapses and nerve fiber bundles in the brain area, thus allowing patients to obtain some beneficial results in executive dysfunction, attention, and so on. In addition, many studies have shown that tDCS also has long-term enhancement or long-term inhibition of the after-effect, which is related to the change of synaptic plasticity. Studies have shown tDCS targeting the dorsolateral prefrontal cortex in healthy individuals has been shown to enhance cognitive functions, including attention, visual memory, logical reasoning, and response times (Hoy et al., 2013). Applying anodal tDCS to the L-DLPFC for periods of 10-30 min (Hoy et al., 2013;Teo et al., 2011;Zaehle et al., 2011;Ohn et al., 2008;Fregni et al., 2005;Nosek and Banaji, 2001) has been found to enhance working memory in subjects. Ohn et al (Ohn et al., 2008). found in the study that during the stimulation process and 30 min after the stimulation, the accuracy score of the verbal 3-back task was significantly improved. Furthermore, in a study by Fregni et al. (2005), positive effects on working memory performance were observed after only 10 min of tDCS stimulation. In clinical trials of patients with Parkinson's disease, a stimulation course has similar positive effects (Fregni et al., 2010;Boggio et al., 2006). It is worth noting that in the study of Boggio et al (Boggio et al., 2006)., 2 mA stimulation produced better effects than 1 mA stimulation.The longer lasting after-effects of tDCS trigger neuroplasticity processes, as indicated by the associated regulation of several neuroplasticity markers by tDCS, such as glutamate (Nitsche et al., 2010a), gamma-aminobutyric acid (GABA) (Nitsche et al., 2004), dopamine (Nitsche et al., 2010b;Monte-Silva et al., 2009), serotonin (Paulus, 2009), acetylcholine (Kuo et al., 2007), and BDNF (Fritsch et al., 2010a). In addition, the reduction of GABA concentration induced by anodic tDCS and the concurrent reduction of GABA-gated intracortical inhibition (ICI) led to the promotion of glutamate-driven neuroplasticity (Stagg et al., 2010;A S et al., 2014). tDCS can play a neuroprotective role in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced Parkinson's disease mouse models by reducing markers of oxidative stress (Lu et al., 2015). Guo et al. (2020) found that tDCS can reduce oxidative stress by inhibiting MDA and ROS levels and by increasing SOD and GSH levels. Additionally, tDCS was observed to mitigate the expressions of IL-1β, IL-6, and TNF-α induced by hypoperfusion. This led to a decrease in hippocampal inflammation, thereby ameliorating cognitive dysfunction in rats with vascular dementia.
The effects of aerobic exercise (AE) on brain activity have been shown to be comparable to those of tDCS. Numerous studies have demonstrated that long-term moderate AE can improve central nervous system function and enhance brain plasticity. AE has been shown to notably improve executive function, with a particularly pronounced effect on inhibitory control compared to other sub-functions (Drollette et al., 2014;Katja et al., 2014;Chen et al., 2014). Research using functional near-infrared spectroscopy (fNIRS) and functional magnetic resonance imaging (fMRI) has revealed that. only a single moderate-intensity AE for 20-30 min can improve the neural activation level of L-DLPFC and prefrontal cortex (FPA), resulting in changes in the functional plasticity of brain executive control network and frontal parietal network related regions, and enhance the ability to utilize brain neural resources in the process of processing executive function tasks. AE can also improve the endogenous antioxidant defense system, stimulate the expression of superoxide dismutase, glutathione peroxidase, and glutathione reductase, while reducing the concentration of several inflammatory markers (Koloverou et al., 2018a). A single session of exercise can induce. the synthesis and release of a variety of neurochemicals, such as norepinephrine, epinephrine, dopamine, BDNF, etc. Mrakic-Sposta et al (Mrakic-Sposta et al., 2018a). found that AE can reduce the level of oxidative stress (ROS production, oxidative damage of lipids and DNA, etc.) resulting in improved the cognitive and behavioral ability of elderly patients. Long-term AE intervention has a positive effect on the proliferation and survival of cells in the hippocampal dentate gyrus and improves learning and memory. AE can improve the cognitive decline caused by the weakening of long-term potentiation (LTP) phenomenon, and can also regulate the decrease in the number of synapses and the fluidity of synaptosomes membrane, thereby affecting brain structure to a certain extent and slowing or even reversing brain structure atrophy (Cooper et al., 2018). Other studies suggest an inverted U-shaped relationship between physical activity levels and cognitive performance (Netto et al., 2018), and the activation level of subjects is highest at 66% max heart rate (Li et al., 2020) or 50% VO 2 max intensity (Zhao et al., 2020), and there is no significant neural activation at too large or too small load intensity. It can be seen that regular moderate intensity AE can regulate the stability of the neuronal environment, enhance neuroplasticity, and improve cognitive function.
The combined effects of tDCS and AE at neurophysiological, neurochemical, and behavioral levels -which include cognitive enhancement, changes in neural activity, and shifts in catecholamine levels -when applied in direct combination, have striking similarities and provide synergistic pathways that may improve brain function and neuroplasticity in both health and disease. While AE may cause a larger range of changes throughout the brain, acting as a brain primer to stimulate specific brain states, tDCS may be used to more intensively and specifically regulate brain activity and behavior. At the same time, both of them can improve cognition and executive function through oxidative stress effect, which is highly consistent with each other. The combined effects of aerobic exercise and tDCS on cognitive function and their potential synergistic effects have not been thoroughly investigated. Some studies have suggested that performing AE directly before an anodal tDCS session on the PFC could be beneficial. In these sessions, executive functions are trained or tested, potentially complementing the induction of neuroplasticity based on BDNF expression.The increase in BDNF might create an optimal environment for longer lasting tDCSinduced LTP initiation. As BDNF factors are necessary for successful and efficient LTP induction , these pathways (AE-initiated BDNF and tDCS-initiated LTP) might converge into optimal learning processes and synaptic plasticity, but the detailed mechanisms are still not clear. Accordingly, the purpose of this study is to compare the effects of aerobic exercise, tDCS, and a combined aerobic exercise and tDCS intervention on the executive function (including working memory, inhibition control and cognitive flexibility) of healthy young adults and to investigate changes in serum BDNF, MDA, SOD, GPX4, iron ions, and glutamate levels following these interventions.
This study aims to address previous research limitations by investigating the synergistic effects of aerobic exercise combined with tDCS on executive function subcomponents and associated biochemical markers, providing a more comprehensive understanding. By elucidating the intricate interplay between BDNF, SOD, MDA, GPX4, glutamate, serum iron, and oxidative stress in the context of executive function, we hope to gain insights into the mechanisms underlying executive dysfunction and inform the development of more effective therapeutic strategies to address neurodegenerative diseases and age-related executive function decline. To test these hypotheses, we will conduct a randomized controlled trial involving healthy young adults who will be assigned to one of three intervention groups: AE, tDCS, or a combined AE and tDCS intervention. Participants' executive function will be assessed before and after the interventions using a battery of psychology and behavior tests, and blood samples will be collected to measure the levels of neurotrophic support (BDNF), oxidative stress (MDA, SOD), neurotransmitter regulation (glutamate) and iron ion. The results of this study will help us understand the mechanisms underlying the executive function benefits of tDCS combined with AE, and develop more effective strategies to improve executive function and prevent neurodegenerative diseases.

Study design and settings
The current study was a randomized controlled trial (RCT) (Fig. 1A). The experimental protocol was thoroughly explained to the participants on the day of the experiment. Before participating, they were asked to complete a tDCS questionnaire assessing their eligibility for the study and any potential use of psychostimulan prior to the experiment. Upon completion of the questionnaire, they were asked to sign an informed consent form. Participants were then randomly assigned to the tDCS+AE, tDCS or AE groups. Behavioral evaluation and blood sample collections were performed at two time points, i.e., one day before and one day after the experiment (Fig. 1C). All experiments were conducted either in the early morning (8-10 am) or early afternoon (1-3 pm). Following the experiment, participants were asked to complete a tDCS experience questionnaire, which aimed to report any side effects of the stimulation. These potential side effects included minor discomfort, mild headache, neck pain, tingling, itching, burning sensation, sleepiness, metallic/iron taste, fatigue, trouble concentrating, acute mood change, investigator, skin redness, and skin irritation. This study was approved by the Ethics Committee of Wuxi Mental Health Center (Ethics WXMHCIRB2021LLky145) and conducted in accordance with the Helsinki Declaration. The trial was registered in the Chinese Clinical Trial Registry (ChiCTR2200057170). All participants signed informed consent forms before being randomly assigned to their respective groups.

Participants
The sample size for this study was calculated using G*Power software based on the results of a previous study that indicated an effect size of − 0.73 (Conceição et al., 2021) . We anticipated that the target effect size had 90% power with a type I error of 5% (α = 0.05). Thus, the sample size for each group was determined to be at least 16. Taking into account an estimated dropout rate of 20%, the required sample size increased to 20 participants per group. Therefore, a total of 60 participants (20 per group) were targeted for this study. The selection of n = 60 (20 per group) also ensured an adequate size for within-group analyses. A total of 70 college students (23 male and 47 female) expressed interest in participating in the trial, and 67 of them (22 male and 45 female) were recruited through advertisements based on the following inclusion criteria: being between 18 and 25 years of age, being right-handed, having normal or corrected vision and color vision, having a body mass index (BMI) less than 25, no history of mental or neurological disease, cardiovascular disease or physical disability, and lack of regular exercise habits. To avoid the potential ceiling effect of exercise, participants should not have regular exercise habits [defined as exercising at least 3 times a week, each session lasting over 30 min at moderate exercise intensity (heart rate >110 times/min) and maintained for more than a year]. Participants completed the Physical Activity Readiness Questionnaire (PAR-Q) (Warburton et al., 2019), the Edinburgh handedness inventory (Oldfield, 1971), the international physical activity questionnaire (IPAQ) and the shortened version of Raven's Advanced Progressive Matrices test (sRAPM) (Marzecová et al., 2013). Participants were randomly assigned to the tDCS+AE group, tDCS group, or AE group, with an equal number of males and females in each group. Participants were instructed to stop exercising 24 h before the experiment and to wear comfortable clothes and shoes for all interventions. The trial should be discontinued if treatment is not performed according to the study protocol or if fatigue, headache, dizziness, and tension persist without relief. If the subject voluntarily requested withdrawal, the test was terminated.

Randomization
SPSS v.22.0 was utilized for random group assignment. Males were assigned numbers from 1 to 22 according to the order of enrollment. Initially, a starting point was established in random number generators, followed by the use of Rv. Uniform(0,1) in Compute Variable to generate a random number for each participant. Subsequently, a cutoff point was set using Equal Percentiles Based on Scanned Cases in Visual Binning, allowing males to be randomly assigned to the tDCS+AE, tDCS, AE experiment groups. These steps were repeated to ensure that females were also randomly assigned to the three experimental groups. Unfortunately, due to reasons such as having a cold or needing to return to school, 1 woman in the tDCS group, 4 women in the AE group, and 2 women in the tDCS+AE group were unable to complete the experiment. This resulted in a total of 60 participants who successfully completed the experiment. The final number of participants in each group who completed the experiment was 20 (7 male and 13 female) in the tDCS group, 20 (7 male and 13 female) in the AE group, and 20 in the tDCS+AE (8 male and 12 female) group. The specific test methods were as follows (Fig. 1A).

tDCS
Facilitatory anodal tDCS was used in the tDCS and AE + tDCS sessions. The tDCS was administered using a battery-driven stimulator (YZB/Chuan 0185-2014, Sichuan Machinery Note 20142210040, Sichuan Smart Electronics Industry Co., Ltd.) through a pair of salinesoaked sponge electrodes (7 cm × 5 cm). The stimulation targeted the L-DLPFC. The anode electrode was positioned at the F3 position according to the 10-20 international EEG system, with the cathodal electrode placed in the contralateral supraorbital area (FP2). The stimulation current intensity was 2 mA for 30 min with an accelerated ramping at the beginning of stimulation and a decelerated ramping at the end, each lasting 30 s. Sham (placebo) stimulation included ramping up and down the current for 30 s, but no current was delivered during the exercise (AE session). Participants were kept blind to the type of tDCS during the AE + tDCS and AE sessions (Fig. 1B).

AE
AE was conducted on a power bike (model motion cycle 600med, Germany) at a moderate intensity, personalized to each individual's maximum heart rate (HRmax). HRmax was estimated using the following formula: HRmax = 220 -Age. The AE program lasted 40 min and consisted of 3 phases: (1) a 5-minute warm-up with heart rate maintained between 55% and 60% of HRmax; (2) a 30-minute exercise at intensity between 60% and 70% of HRmax; and (3) 5 min of finishing The study timeline is also depicted (C). First, participants provided demographic information. The 2-back, 3-back, Staircase task, Stroop task, T test, and HOJ were measured one day before and again after the completion of all intervention sessions. During the intervention period, the experimental group underwent a 4-week intervention, consisting of 5 sessions per week, each lasting 30 min, spanning a total of 4 weeks. Abbreviations: HR, heart rate; RD, ramp down; RU, ramp up; tDCS, transcranial direct current stimulation. exercise, keeping the heart rate below 65% of HRmax. HR was recorded using a heart rate monitor (model DB18, Philips, Suzhou Erda Medical Equipment Co., Ltd.) throughout the exercise period (Fig. 1B).

Sample collection
One day before and after the intervention, peripheral blood samples were taken aseptically, and serum samples were obtained by centrifuging (3000 rpm, 10 min) the blood samples and stored at − 80 • C until testing.

Assessment tools
We used the following assessment tools to evaluate different aspects of the subjects' executive function abilities one day before the test and after 4 weeks of the test: (1) 2-back, 3-back, and staircase working memory tasks were applied to evaluate working memory function. (2) The Word-color Stroop task was used to evaluate inhibition function. (3) The T test (Lambourne and Tomporowski, 2010) and the hexagonal obstacle jump test (HOJ) (Kibele et al., 2014) were used to evaluate task switching ability in the exercise state. The 2-back, 3-back, Staircase, and Word-color Stroop experimental tasks were all written with E-prime 3.0 software and run on a Lenovo Thinkpad X13 Ge laptop. The T test and the HOJ were both carried out by two special personnel at the same time to ensure the accuracy of the test.

2,3-back
The 2-and 3-back tasks required subjects to respond by pressing the "F" or "J" key with both forefingers by comparing the current stimulus (number) with the second or third previous stimulus. The keystroke correspondences were balanced and random among subjects, i.e., the "F" key was the same stimulus for half of the subjects, and the "J" key was the same stimulus for the other half. Subjects were asked to complete a 2back task and a 3-back task. In the 2-back task, subjects were asked to compare the current stimulus to the second number before it, and in the 3-back task, they were asked to compare the current stimulus to the third number before it. In each trial, the digital stimulus was presented for 600 ms, followed by an empty screen for 1400 ms, and subjects were allowed to perform keystrokes between the present stimulus and the subsequent blank screen. Before the scan, all subjects were given sufficient task instructions and received corresponding task training to ensure that they understood how to complete the task. The 2-back task and the 3-back task each consisted of 100 trials, of which 50 were consistent and 50 were inconsistent (Fig. 2 A).

Spatial working memory
At the beginning of each trial, a 500 ms "+ " fixation point is presented in the center of the white screen, followed by a 2000 ms target stimulus coding interface consisting of a 6 × 6 square matrix with n random small squares filled in green, each 60 × 60 pixels in size. This is followed by a memory maintenance interface in which all the small squares are filled in white and the interface appears for 2000 ms. The initial interface of the recall interface is the same as that of the memory maintenance interface. Subjects needed to click the square position of all the green squares in the target stimulation interface with the mouse, and each square was filled with green after being clicked. After subjects click n positions, an 800 ms feedback interface will be presented. The feedback interface contains two pieces of information: 1. The number of correct squares in the current trial response m; 2. Correct square ratio (m/n * 100%) for the current test response.
This task uses Staircase to dynamically adjust the task difficulty. If the number of squares that the subject responds correctly in the current trial (m) is less than the number of squares that need to be memorized (n), the number of squares that need to be memorized in the next trial (m) will become n-1; if the subject gets 100% correct on two consecutive attempts, the number of squares needed to be memorized on the third attempt becomes n + 1. This task consists of a total of 40 tests. A practice test will be given before the formal task to enable subjects to understand the test methods. When a subject's performance at a specific task difficulty level cannot be continuously improved or continuously decreased, the task difficulty level is called a turning point. The average of turning points 2-7 was taken as the indicator of the visual working memory span of the subjects (Fig. 2B).

Word-color Stroop task
The experiment in this task stimulated Chinese characters written in different colors, including "red," "green," "yellow," and "blue." The colors used for writing also included red, green, yellow, and blue. Subjects needed to respond to the corresponding buttons for the writing color used in the currently presented Chinese characters while ignoring the semantic meaning of the presented Chinese characters. They were asked to place their left middle finger, left index finger, right index finger and right middle finger on the DFJ and K keys, each representing a color. In each trial, a fixation point for 500 ms was presented, followed by a Chinese character, which was presented until the subject responded. The block consisted of 120 trials, of which 60 were consistent and 60 were inconsistent, and the presentation of the trials was random. Subjects' accuracy rate and reaction time each time were recorded as corresponding evaluation indexes (Fig. 2 C).

T test
The T test began with the subject standing at point A. Upon hearing the start signal, the subjects quickly ran to point B, touched cone B, ran 5 yards to the left (4.6 m) to touch cone C, ran 10 yards to the right (4.6 m) to touch cone D, ran 5 yards to the left (4.6 m) to touch cone B, and ran back to point A. The stopwatch stopped. Two tests were conducted, and the best result was obtained (Fig. 2D).

Hexagonal obstacle jump task
The subject stood in the middle of a hexagon with each side length of 0.6 m and a cone placed at each corner. After hearing the start signal, the subject jumped from the center to each corner with his feet in the clockwise direction, then jumped back to the center again, repeating this for each corner three times (approximately three times), and finally jumped back to the origin. The subjects were asked to face the same direction throughout the test. Two tests were conducted, and the best score was obtained (Fig. 2E).

Determination of serum cytokines
Serum levels of BDNF and GPX4 were determined using enzymelinked immunosorbent assay (ELISA) kits provided by Lapuda Biotechnology Co., Ltd, Nanjing, China. Serum samples were diluted as recommended by the manufacturer, and the assay was performed in 96well plates. For the determination of MDA, glutathione, SOD, and iron ions, we utilized colorimetric assay kits from Jiancheng Bioengineering Institute, Nanjing, China. The assays were carried out following the manufacturer's recommended protocol, and the absorbance was measured using a microplate reader set at the appropriate wavelength. The lower detection limits (LODs) for each cytokine were BDNF, 0.16 ng/mL; GPX4, 1.25 pmol/mL; MDA, 0.50 nmol/L; glutamate, 1.50 μmol/L; SOD, 2 U/mL; and iron ions, 1.79 μmol/L. Values lower than the LOD were reported as half of the LOD.

Statistical analysis
First, we conducted separate analyses for the accuracy of the 2-back and 3-back tasks, number of spatial working memories, response time of the Stroop task, time of T-test and HOJ, and biomarker levels. A 3 (Treatments: tDCS+AE vs. tDCS vs. AE) × 2 (Times: Pre vs. Post) repeated-measures ANOVA was used with Greenhouse-Geisser epsilon corrections applied when necessary to meet the assumption of sphericity. For significant results, post hoc comparisons were conducted with familywise alpha levels set at 0.05 using Bonferroni significant difference tests. Next, we calculated the differences in cognitive task scores and biomarker levels before and after the intervention for each group. One-way ANOVA was performed to compare these differences among the three groups, followed by post hoc pairwise comparisons using Bonferroni correction for multiple comparisons if a significant difference was detected. To assess the relationship between changes in cognitive task scores and changes in biomarker levels, we conducted Pearson's correlation analysis. Statistical significance was set at an alpha level of 0.05, and all tests were two-tailed. All statistical analyses were performed using SPSS 26.0.

Basic information and adverse effects
Participants' characteristics are summarized in Table 1. One-way analysis of variance (ANOVA) indicated that there were no significant differences between tDCS+AE, tDCS and AE in terms of sex, age, height and weight. Regarding the adverse effects of tDCS and AE, most of the reported sensations came from the tDCS+AE and tDCS groups. In the tDCS group, the reported sensations included mild tingling (100% in tDCS+AE, 100% in tDCS), itching (95.00% in tDCS+AE, 95.00% in tDCS), burning sensation (60.00% in tDCS+AE, 65.00% in tDCS), sleepiness (5.00% in tDCS+AE, 0% in tDCS), skin redness (95.00% in tDCS+AE, 95.00% in tDCS), and skin irritation (85.00% in tDCS+AE, 90.00% in tDCS) ( Table 1). The Wilcoxon test also revealed no significant difference between the tDCS+AE and tDCS intervention sessions for participants' guesses about the experiment condition.

Executive function performance
According to the experimental design, we divided participants into three groups: the combined group (tDCS + AE), the tDCS group, and the AE (sham tDCS + AE) group. We conducted a series of executive function tasks before and after the intervention for each group to assess their executive functions. These tasks included the working memory tasks (2back, 3-back and spatial working memory tasks), inhibitory control task (Stroop task with congruent and incongruent conditions), and flexible conversion tasks (T-test and HOJ task). Behavioral data and serum biomarker concentrations(mean ± SE) across treatment and tasks are shown in Tables 2 and 3  and serum biomarker concentrations(mean ± SE) across treatment and tasks are shown in Tables 4 and 5, Figs. 5 and 6. Next, the results of these tasks are discussed.
Before the intervention, there was no significant difference in accuracy rates among the three groups. After the intervention, the accuracy rate in the tDCS+AE group was significantly higher than that in the tDCS group (F 2,57 =3.904, p=0.039, η 2 p=0.120), but there were no significant differences between the tDCS+AE group and the AE group (F 2,57 =3.904, p=0.086, η 2 p= 0.120) or between the tDCS group and the AE group (F 2,57 =3.904, p=1.000, η 2 p=0.120) (Fig. 3A).

3-back
Using the accuracy rate of 3-back tasks before and after intervention as the dependent variable, the results showed a significant interaction effect of Treatments ( There was a significant increase in accuracy from pre-intervention (0.64 ± 0.03) to post-intervention (0.84 ± 0.02) in the tDCS+AE group (F 1,57 =10.274, p= 0.000, η 2 p= 0.265), from pre-intervention Before the intervention, there were no significant differences in accuracy rates among the three groups. After the intervention, the accuracy rate in the tDCS+AE group was significantly higher than that in the tDCS group (F 2,57 =10.274, p = 0.000, η 2 p = 0.265) and the AE group (F 2,57 =10.274, p = 0.004, η 2 p = 0.265), while there was no significant difference between the tDCS group and the AE group (F 2,57 =10.274, p = 1.000, η 2 p = 0.265) (Fig. 3B).
Before the intervention, there was no significant difference among the three groups in the number of correct Staircase responses. After the intervention, there was no significant difference between the tDCS+AE group and the tDCS group (F 2,57 =3.304, p = 0.246, η 2 p = 0.104), but there was a significant difference between the tDCS+AE group and the AE group (F 2,57 =3.304, p = 0.046, η 2 p = 0.104); there was no significant difference between the tDCS group and the AE group (F 2,57 =3.304, p = 1.000, η 2 p = 0.104) (Fig. 3E).
There was no statistical difference in the congruent reaction time among the three groups pre-and post-intervention (Fig. 3C).
There was no statistical difference in the T-test task among the three groups pre-and post-intervention (Fig. 3F).
There was no statistical difference in the HOJ task among the three groups pre-and post-intervention (Fig. 3G).
Post-intervention, there was no statistical difference in the concentration of MDA among the three groups (Fig. 4B).

SOD
The concentration of SOD in the blood pre-and post-intervention was used as the dependent variable. The results showed a significant interaction for Treatments (tDCS+AE vs. tDCS vs. AE) × Times (Pre vs. There was no statistical difference in the SOD concentration among the three groups pre-and post-intervention (Fig. 4D).
There was no statistical difference in the GPX4 concentration among the three groups pre-and post-intervention (Fig. 4E).
There was no statistical difference in the iron ions concentration among the three groups pre-and post-intervention (Fig. 4F).

Changs of executive function performance
Based on the results of the single-factor variance analysis of the differences between each task before and after intervention, the following results were observed.

Changes of serum biomarker concentrations
Based on the results of the single-factor variance analysis of the differences between each task before and after intervention, the following results were observed.

The relationships between the change in biomarker levels and executive function performance
The change of BDNF levels and the reaction time of the Stroop incongruent task showed a significant negative correlation (r = − 0.295, p = 0.022). This indicates that an increase in BDNF levels after the intervention is associated with an improvement in Stroop incongruent task performance (Fig. 7A).
The change of BDNF levels and the reaction time of the HOJ task showed a significant negative correlation (r = − 0.338, p = 0.008). This indicates that an increase in BDNF levels after the intervention is associated with an improvement in HOJ task performance (Fig. 7B).
The change of MDA levels and the reaction time of the Stroop congruent task showed a significant positive correlation (r = 0.325, p = 0.011). This suggests that a decrease in MDA levels after the intervention is associated with an improvement in Stroop congruent task performance (Fig. 7C). The change of MDA levels and the reaction time of the Stroop incongruent task also demonstrated a significant positive correlation (r = 0.408, p = 0.001). This indicates that a decrease in MDA levels after the intervention is related to an improvement in Stroop incongruent task performance (Fig. 7D).
The change of glutamate levels and 3-back task performance showed a significant negative correlation (r = − 0.307, p = 0.017). This suggests that an increase in glutamate levels after the intervention is associated with an improvement in 3-back task performance (Fig. 7E).
In addition, the change of BDNF levels and MDA levels showed a significant negative correlation (r = − 0.353, p = 0.006). This suggested that an increase in BDNF levels after the intervention is associated with a decrease in MDA levels and vice versa (Fig. 7F).

Discussion
In recent years, more and more research has shown that oxidative stress is closely linked to cognitive and executive-related disorders, The onset of neurodegenerative diseases related to executive function such as mild cognitive impairment (MCI), Alzheimer's disease (AD), Parkinson's disease (PD), stroke.etc are all associated with a series of oxidative stress-enhanced responses such as increased free radicals, lipid peroxidation, dysregulation of calcium homeostasis, cytochrome C release, and increased iron deposition. Excessive oxidative stress eventually leads to neuronal apoptosis/death, which is the basis of these neurodegenerative diseases related to executive function. Therefore, how to improve oxidative stress response to improve executive function is worthy of our exploration and research.
The combination of tDCS and AE proposes a promising strategy for cognitive enhancement, leveraging synergistic effects on cortical excitability and neuroplasticity. Simultaneously, both tDCS and AE could improve subjects' executive function through modulating oxidative stress response. Specifically, tDCS can increase the release of glutamate (Nitsche et al., 2010a), gamma-aminobutyric acid (GABA) (Nitsche et al., 2004), dopamine (Nitsche et al., 2010b;Monte-Silva et al., 2009), serotonin (Paulus, 2009), acetylcholine (Kuo et al., 2007), and BDNF (Fritsch et al., 2010a) in the brain by regulating neuroplasticity，In animal experiments, tDCS can also reduce oxidative stress by inhibiting the levels of MDA and ROS, increasing the levels of SOD and GSH, and inhibiting the expressions of IL-1β, IL-6 and TNF-α induced by low perfusion, thereby reducing the inflammatory response of hippocampus and improving cognitive function (Guo et al., 2020) Likewise, systemic AE in animals has also been found to increase arousal and promote the release of brain neurotransmitters such as norepinephrine, serotonin, dopamine (Meeusen and De Meirleir, 1995), brain-derived neurotrophic factor (BDNF), insulin-like growth factor 1 (IGF-1), lactate, and vascular endothelial growth factor (VEGF) (Basso and Suzuki, 2017;Mcmorris, 2016;Mcmorris and Hale, 2015). Moderate AE can also bolster the endogenous antioxidant defense system, stimulate the expression of superoxide dismutase, glutathione peroxidase and glutathione reductase, while reducing ROS production, mediating cell remodeling through lipid interaction, reducing oxidative damage, and improving cognitive function (Mrakic-Sposta et al., 2018b;Koloverou et al., 2018b). However, the joint effects of AE and tDCS on executive function, as well as their potential synergies on biomarkers such as oxidative stress and neurotransmitter have not been thoroughly investigated.This highlights the importance of further research into these potential synergistic effects, which could unveil new strategies beneficial for enhancing executive function.
In this study, we randomly divided 60 healthy adults into three groups: a group receiving combined tDCS and AE (tDCS+AE), a group receiving only tDCS, and a group undergoing only AE. Over a 20-day period, we evaluated executive functions including working memory, conflict inhibition, and flexibility, and analyzed serum biomarkers such as BDNF, MDA, SOD, GPX4, glutamate, and iron ions, both before and after the intervention. At the same time, the correlation between the changes of executive function and serum biomarkers before and after intervention was analyzed. The main findings of this study is that the combined intervention of AE and tDCS led to more significant improvements in executive function compared to either intervention alone. Furthermore, these improvements were accompanied by notable changes in serum biomarkers associated with oxidative stress and executive function regulation.
Our study extends previous findings on individual interventions by demonstrating the synergistic effect of combining tDCS with aerobic exercise. A repeated measures ANOVA (3 Treatment groups: tDCS+AE vs. tDCS vs. AE, and 2 Time points: Pre vs. Post) revealed significant enhancements in the executive function after 4 weeks of combined intervention. Specifically, the tDCS+AE group showed greater accuracy on the 2-back working memory task than the AE group, and on the 3back task than both the tDCS and AE groups. Moreover, the tDCS+AE group also showed improved reaction times in the conflict suppression task, Stroop, compared to the other two groups. While there was no statistically significant difference among the three groups in their flexibility conversion ability as measured by repeated measure ANOVA, the tDCS+AE group demonstrated a statistically significant improvement in the performance of the HOJ task compared to the AE group.
Regarding biomarkers, we performed a repeated measures ANOVA (3 Treatment groups: tDCS+AE, tDCS, AE; 2 Time points: Pre-vs. Postintervention). Our findings indicate that only the tDCS+AE group demonstrated a statistically significant reduction in glutamate concentration post-intervention compared to the tDCS group. Furthermore, we observed significant increases in the concentrations of MDA and SOD in the tDCS+AE group compared to both the tDCS and AE groups. Additionally, the BDNF concentration in the tDCS+AE group was significantly higher than in the tDCS group. Lastly, the tDCS+AE group had a significantly lower glutamate concentration than the AE group.
Our study demonstrated significant correlations between the changes in BDNF, MDA, and glutamate levels and various executive function task performances post-intervention. Notably, BDNF levels were negatively correlated with MDA levels and positively correlated with Stroop incongruent task performance. Conversely, MDA levels were positively correlated with Stroop congruent and incongruent task performances. Additionally, glutamate levels were negatively correlated with 3-back task performance. These findings suggest that the intervention may have influenced the participants' working memory function and inhibitory control function performance through changes in these biomarkers.
The improvements in executive function observed in our study correlated with changes in BDNF levels, suggesting that the combined intervention may bolster executive performance by modulating neurotrophic support. This supports the potential superiority of combined tDCS and AE interventions in optimizing executive function, underscoring the significance of targeting biomarkers such as BDNF in these strategies. Our findings align with previous research demonstrating associations between BDNF levels and cognitive performance, particularly within inhibitory control function and working memory tasks (e.g., Stroop and n-back tasks) (Hillman et al., 2014;Leckie et al., 2014;Szuhany et al., 2015). Studies by Hillman et al. (2014) and Leckie et al. (2014) demonstrated that exercise interventions improved specific executive functions, with these improvements mediated by changes in BDNF levels. Szuhany et al.'s (Szuhany et al., 2015) meta-analysis further strengthened this evidence, revealing a significant positive relationship between exercise interventions and increased BDNF levels. Similarly, tDCS interventions have been shown to improve executive function performance through the modulation of BDNF levels and the balance of neurotransmitters (Hone-Blanchet et al., 2016;Smith et al., 2013;Fritsch et al., 2010b). Hone-Blanchet et al. (2016) and Agarwal et al. (2013) provided evidence that tDCS could lead to cognitive improvements, including in working memory and attention, which are key components of executive function. Moreover, Fritsch et al. (2010b) showed that tDCS can promote BDNF-dependent synaptic plasticity in vitro, suggesting a potential mechanism through which tDCS may enhance executive function.
Our findings underscore the advantage of combined tDCS and AE interventions in reducing MDA levels and augmenting executive performance, notably in tasks involving inhibition control functions such as the Stroop congruent task. These observations align with earlier research demonstrating a correlation between MDA levels and executive function performance (Smith et al., 2013;Gomez-Cabrera et al., 2008). For instance, Smith et al (Smith et al., 2013). and Gomez-Cabrera et al (Gomez-Cabrera et al., 2008). illustrated how exercise interventions reduced oxidative stress, as signified by reduced MDA levels, thereby enhancing cognitive performance. Similarly, Peruzzotti-Jametti et al. (2013) reported that tDCS interventions led to reductions in MDA levels and enhanced cognitive function. The significant positive correlation observed between the change in MDA levels and Stroop congruent task performance in our study supports the notion that the combined interventions might enhance executive performance, particularly in inhibition control function, through the modulation of oxidative stress markers such as MDA. Moreover, our study found a significant negative correlation between the changes in BDNF levels and MDA levels, suggesting that an increase in BDNF levels after the intervention is associated with a decrease in MDA levels and vice versa. This finding highlights the potential interplay between neurotrophic support and oxidative stress in modulating executive performance and underscores the importance of targeting both BDNF and MDA to optimize intervention strategies (Knaepen et al., 2010;Radak et al., 2006). For example, Radak et al. (2006) and Knaepen et al (Knaepen et al., 2010). demonstrated that combined interventions targeting both BDNF and MDA led to greater improvements in executive performance, including inhibition control functions and working memory, than single interventions. In our study, we also explored the associations between MDA and glutamate levels and executive performance. These correlations suggest that the changes in oxidative stress and neurotransmitter levels induced by the intervention might mediate improvements in executive function. The negative correlation between BDNF and MDA levels might indicate that the intervention effectively mitigated oxidative stress, potentially contributing to a surge in BDNF levels. Increased BDNF levels have been linked to enhanced neuroplasticity and improved executive function. The positive correlation between MDA levels and Stroop task performance might indicate that reducing oxidative stress is essential for improving cognitive performance in tasks that require executive functioning. Our findings highlight the superiority of combined tDCS and AE interventions in reducing glutamate levels and improving executive performance, especially in working memory tasks like the 3-back task. Previous research has demonstrated that exercise and tDCS can independently modulate brain glutamate levels and improve cognitive performance, encompassing attention, memory, and executive functions (Lengu et al., 2021;Maddock et al., 2016). Maddock et al. (2016) demonstrated that a single session of moderate-intensity aerobic exercise could acutely modulate cortical glutamate and GABA content in healthy adults, as measured by proton magnetic resonance spectroscopy (1 H-MRS). The study found exercise led to significant decreases in glutamate and GABA levels in the visual cortex, a region implicated in visual processing and attention. These neurotransmitter changes were associated with improvements in attentional performance, suggesting exercise can influence executive performance by modulating neurotransmitter levels in the brain. Similarly, Lengu et al. (2021) applied high-definition transcranial direct current stimulation (HD-tDCS) over the left dorsolateral prefrontal cortex (DLPFC), a brain region associated with working memory and executive functions, in older adults with or without mild cognitive impairment. The study used 1 H-MRS to measure local GABA and glutamate levels and found that HD-tDCS led to significant decreases in glutamate levels in the DLPFC. The reduction in glutamate levels was associated with improvements in working memory performance, suggesting that tDCS can modulate executive performance, particularly working memory performance, by altering glutamate levels in specific brain regions.
In terms of SOD, we found significant differences between the combined group and both the tDCS group and the AE group. This finding supports the notion that the combined intervention of tDCS and exercise has a more prominent effect on SOD levels than either intervention alone. SOD is an essential antioxidant enzyme that helps protect cells from oxidative stress (Fukai and Ushio-Fukai, 2011). Improved SOD activity may contribute to a reduction in oxidative stress, which could potentially underlie the observed executive improvements in the combined intervention group. Radak et al (Radak et al., 2008). discussed how exercise can lead to both oxidative stress and the activation of cellular antioxidant signaling pathways, such as the upregulation of antioxidant enzymes such as SOD. The authors highlighted that moderate exercise increases SOD activity, which helps combat oxidative stress and promotes positive health outcomes. Similarly, Ji et al (Ji, 2006). investigated the influence of exercise on cellular antioxidant pathways, including the activation of antioxidant enzymes such as SOD. Similarly, Yoon et al (Yoon et al., 2012). showed that anodal tDCS in a rat model of cerebral ischemia resulted in increased SOD levels and promoted the ability of the brain to regulate motor functions. These studies provide evidence for the positive impact of exercise or tDCS on SOD levels.
The results of the present study show no significant differences in GPX4 levels between the combined intervention group and the tDCS group or the AE group. Similarly, no significant differences were observed in serum iron ion levels among the three intervention groups. These findings suggest that the combined intervention of AE and tDCS may not have a synergistic effect on GPX4 levels or iron ion homeostasis. Aerobic exercise has been previously reported to increase GPX4 activity, which helps protect cells from oxidative damage (Ansari and Scheff, 2010). However, our findings did not show significant differences in GPX4 levels between the groups, possibly due to variations in exercise protocol, intensity, or duration (Daussin et al., 2012). The relationship between tDCS and GPX4 levels has not been extensively studied, and the potential mechanisms underlying tDCS-induced changes in GPX4 levels remain to be elucidated (Yoon et al., 2012). Regarding iron ion levels, previous studies have shown that exercise can influence iron metabolism, which may play a role in cognitive function (Belaya et al., 2021). Nevertheless, our results did not show significant changes in iron ion levels among the groups. This could be attributed to the relatively short duration of the intervention or the possibility that the effects of exercise and tDCS on iron ion homeostasis may be more subtle than those on other biomarkers such as BDNF, MDA, and SOD (Belaya et al., 2021). Additionally, the precise mechanisms by which iron ions contribute to neuronal damage and cognitive decline are still not fully understood (Mueller et al., 2012), warranting further research.

Conclusions
In this study, we observed through randomized controlled trials that a combined tDCS and AE intervention over a 4-week period could enhence the executive function of healthy young people more effectively than either tDCS or AE alone. The most notable improvements were found in the accuracy of 2-back and 3-back tasks, as well as a reduction in response times for the Stroop task and the duration of HOJ. Moreover, lower concentrations of MDA and higher concentrations SOD were recorded in the tDCS+AE group compared to the tDCS and AE groups. The tDCS+AE group also exhibited a higher concentration of BDNF compared to the tDCS group, and a lower concentration of glutamate compared to the AE group. Correlation analysis revealed negative correlation between changes in BDNF concentration and changes in Stroop incongruent response time and HOJ task duration, a positive association between changes in MDA concentration and changes in Stroop congruent and incongruent response times, and a negative association between changes in glutamate concentration and changes in 3-back accuracy. Furthermore, we found a negative correlation between BDNF and MDA concentrations.
In summary, our study supports the hypothesis that a combined intervention of tDCS and AE is more effective in enhancing inhibition control function and working memory than either intervention alone. The observed improvements in executive functions were associated with changes in serum biomarkers related to neurotrophic support (BDNF), oxidative stress (MDA, SOD), and neurotransmitter regulation (glutamate). These findings suggest that the combined intervention may enhance executive performance by modulating the biomarkers, emphasizing the importance of targeting multiple physiological pathways to optimize intervention strategies. Furthermore, our study contributes to the growing body of evidence supporting the potential of noninvasive brain stimulation and exercise interventions improving executive function across various populations, including healthy adults and individuals with cognitive impairments.

Limitation and perspectives
While our findings provide valuable insights into the combined effect of tDCS and AE interventions on executive function and related biomarkers, they should be interpreted in the context of several limitations. First, our study sample was relatively small and comprised only healthy young adults, which may limit the generalizability of our findings to older individuals or those with cognitive impairments. Future research should consider a larger and more diverse sample to validate these findings. Second, the duration of our study was relatively short, which may have influenced the changes we observed in certain biomarkers, such as GPX4 and iron ion levels. Longer-term studies are needed to better understand the impact of combined tDCS and AE interventions on these biomarkers. Third, we did not control for potential confounding factors such as diet, sleep, and other lifestyle factors that could affect biomarker levels and cognitive performance. Future studies should consider these variables to more accurately isolate the effects of the intervention. Despite these limitations, our study provides preliminary evidence of the potential synergistic effects of tDCS and AE on executive function and associated biomarkers, opening up new avenues for future research in this area.
Our current study proposes the use of centrally-acting tDCS in conjunction with the peripheral regulation of AE, to provide a synergistic rehabilitation therapy for enhancing executive function that could be more effective than either therapy alone. As both tDCS and AE are simple, safe, and clinically convenient, their combined use could offer benefits beyond those of singular use, such as a shortened rehabilitation treatment time, resistance to oxidative stress, increased brain BDNF, and reduced release of inhibitory neurotransmitters. This makes it a suitable rehabilitation option for patients with neurodegenerative diseases related to executive dysfunction, such as MCI, AD, PD, and stroke. The multi-modal rehabilitation therapy of tDCS combined with AE has the potential to improve the executive function of patients, improve their daily life ability, and augment the effect of rehabilitation therapy. Given its positive effect on the recovery from diseases, its convenience, and its potential for widespread application, this combined therapy has a promising future.

Future research directions
Future research should focus on investigating the long-term effects of combined tDCS and exercise interventions on executive function and the underlying physiological mechanisms. Moreover, it would be valuable to explore the optimal parameters for tDCS and exercise interventions, such as stimulation intensity, duration, and frequency, as well as exercise intensity and modality, to maximize their cognitive benefits. Additionally, studies examining the efficacy of combined interventions in other populations, such as individuals with neurodegenerative diseases, could provide valuable insights into potential therapeutic applications.

Ethics approval and consent to participate
This study was approved by the Ethics Committee of Wuxi Mental Health Centre, with the grant number of WXMHCIRB2021LLKY145.

Consent for publication
The patient understood the report, and signed informed consents.

Funding
The work is supported by the National Key R&D Program of China (Grant No.: 2018YFC 2001600, 2018YFC 2001603)

Declaration of Competing Interest
The authors declare no conflict of interest.

Data Availability
Data will be made available on request.