Chronic ankle instability (CAI), characterized by recurrent sprains, is the most frequent long-term consequence following an initial ankle sprain (Delahunt et al., 2010), leading to the discontinuation of exercise in 20% to 40% of affected individuals (Ekstrand and Tropp, 1990). Individuals with CAI often demonstrate impaired postural control, whether in static or dynamic postural control (Docherty et al., 2006; Olmsted et al., 2002; Ross and Guskiewicz, 2004; Vikram et al., 2012; Wikstrom et al., 2007). These deficits may predispose individuals to repetitive trauma and exacerbation of ankle instability, thereby escalating the susceptibility to joint degeneration and arthritis (Hershkovich et al., 2015; Wikstrom and Brown, 2014). Therefore, it is imperative to optimize the rehabilitation plan to minimize postural deficits among CAI populations to the greatest extent possible.

Transcranial direct current stimulation (tDCS) is a non-invasive neurophysiological technique that employs weak currents (1-2 mA) to modulate the activity of brain neurons and alter cortical excitability (Auvichayapat and Auvichayapat, 2011). As an adjunctive therapy for movement disorders, it has been widely used in the treatment of various clinical neurological conditions (Lefaucheur et al., 2017). However, in the field of musculoskeletal injury rehabilitation, research on tDCS remains in the exploratory stage. Bruce et al. (Bruce et al., 2020) and Ma et al. (Ma et al., 2020) respectively showed that tDCS combined with eccentric training or short-foot exercises could significantly improve dynamic postural stability in patients with CAI, thereby highlighting its potential therapeutic benefits. Despite these findings, consensus regarding the optimal stimulation parameters for enhancing treatment outcomes has yet to be reached, thus leaving the protocol for achieving maximal therapeutic benefits undefined. Balance training (BT), performed in a weight-bearing position, has been extensively utilized in rehabilitation due to its proven efficacy in reestablishing neuromuscular control, thereby addressing poor postural control and repairing ankle joint injuries. BT has been a crucial element of the clinical rehabilitation protocol for CAI (Taube et al., 2008). Patients with CAI who undergo BT exhibit positive neural adaptations, which are associated with enhanced balance performance (Chung et al., 2023). However, the effectiveness of combining BT with tDCS has yet to be established in CAI individuals.

The primary purpose of this prospective, randomized, sham-controlled study is to examine the effects of a 4-week tDCS intervention, BT, and a combined tDCS and BT program on static and dynamic postural control in athletes with CAI. The secondary purpose is to explore whether the combined program yields superior effects compared to single-type interventions. We hypothesized that: (1) tDCS, BT, and the combination of both interventions would be effective in improving both static and dynamic postural control in CAI athletes; (2) the combined intervention would yield superior efficacy in postural control compared to tDCS or BT alone.

This randomized controlled trial received approval from the Sports Science Experiment Ethics Committee of Beijing Sport University (2022058H), registered in Chinese Clinical Trial Registry, and conducted in compliance with the principles outlined in the Declaration of Helsinki. Participants were randomly assigned to four groups (G1: tDCS group, G2: s-tDCS group, G3: tDCS + BT group, G4: s-tDCS + BT group) in a 1:1:1:1 ratio using a random number generator in Microsoft Excel. The randomization process was performed by SW, an independent study coordinator who was not involved in participant screening. The sham stimulation protocol can effectively facilitate single blinding, and blinding was maintained for the assessors and statisticians involved in the study.

Earlier research that combined tDCS with exercise and examined similar outcome measures, such as static and dynamic balance, reported effect sizes (η²_p = 0.096-0.154), which correspond to f values ranging from 0.33 to 0.43 (Bruce et al., 2020; McKeon et al., 2008; Xiao et al., 2022). Thus, to determine the required sample size, we utilized G*Power software (version 3.1) with the following parameters: a power of 0.90, a significance level of 0.05, and an effect size of 0.40. The software suggested that each group should have at least 7 participants. To account for an attrition rate of up to 25% and to avoid underestimating the sample size, a total of 40 participants were recruited (See Table 1).

All participants were recruited from Beijing Sport University and required to reach at least the national second-level athletes standard in China, and actively practicing and/or matching at least 3 times per week for 2 hours. Only patients with unilateral ankle sprains were included in the study. More inclusion criteria aligned with the guidelines proposed by the International Ankle Consortium (Gribble et al., 2014). (1) a history of at least one significant ankle sprain, that led to interruption of physical activity for more than one day within the last year; (2) having experienced at least 2 episodes of the ankle “giving way” and/or recurrent in the past 6 months before enrollment; (3) the initial sprain and the most recent injury should have occurred at least 12 months and 3 months before study enrollment, respectively; (4) score ≤ 24 on the Cumberland Ankle Outcome Score (CAIT).Exclusion criteria included any history of prior surgeries or fractures in the lower extremity (Gribble et al., 2014), acute injury to the musculoskeletal structures of lower extremity joints within the previous 3 months (Gribble et al., 2014), chronic musculoskeletal conditions known to induce postural control deficits (Uzlaşır et al., 2021), currently undergoing rehabilitation treatment, having previously received transcranial electrical stimulation therapy, or being inappropriate for electrical stimulation (Rossi et al., 2009).

Balance Error Scoring System (BESS)

The BESS was originally developed for the assessment of postural stability in people with concussion (Wilkins et al., 2004). Currently, its application has gone beyond the original test purpose and scope. It has been widely utilized to evaluate postural stability in diverse populations and has demonstrated good reliability and validity in static balance measurement (Bell et al., 2011). The BESS assessment protocol involves participants standing barefoot on two types of surfaces (firm and foam) while adopting three distinct stances (double-limb, single-limb, and tandem) (Docherty et al., 2006). The foam stance was conducted using a medium-density Airex balance pad measuring 50.8 × 41.7 × 6.4 cm. The test was only administered on the injured side of patients with CAI, and the order of conditions was determined through a random allocation process. Participants were asked to place their hands on the iliac crests, close their eyes, keep their heads upright, and stay as still as possible for a duration of 20 seconds. Each trial is scored by counting errors and at most one error is recorded every two seconds. The maximum error score for each condition is 10, and the total number of errors for the entire test is 60. If they cannot maintain a stable posture for at least 5 seconds with the designated 20-second time frame, it will be scored 10 (Shamshiri et al., 2020). Errors involve taking hands away from the iliac crests, opening their eyes, losing balance or falling during a step, moving the hip into abduction or flexion greater than 30 degrees, raising the forefoot or heel off the test surface, or remaining outside the proper testing posture for over 5 seconds (Riemann et al., 1999). Participants will be given one opportunity for practice, followed by a single formal trial for each of six conditions. After each participant's trial is completed, the pad will be rotated 90 degrees to prevent unevenness caused by repeated pressure. The test is administered by a panel of three raters, and the median value of the three ratings is considered as the ultimate score for this particular condition.

The YBT is a widely recognized method for assessing dynamic postural stability, demonstrating high interrater reliability (ICC 0.81-1.00) and intrarater reliability (ICC 0.85-0.91) (Plisky et al., 2021). To conduct the assessment, the YBT utilizes a testing kit (FMS, Chatham, VA, USA). A verbal and visual demonstration of the testing procedure was given to each subject by an examiner (SW). Participants practiced once and then completed 3 successful test trials for 3 directions while barefoot. They were required to stand with test foot on the starting point and utilize contralateral limb to push the indicator block as far as possible in each direction (Wilson et al., 2018). They should touch the block lightly to avoid providing extra support for maintaining equilibrium and back to the starting point while keeping both hands on their hips (DeJong et al., 2020). The test was only administered on the injured side and the testing order was applied in the following sequence: anterior (ANT), posteromedial (PM), and posterolateral (PL). Participants were deemed to have failed if they lost balance, lifted their stance foot, shifted weight onto the reach foot, were unable to execute a controlled return, failed to touch the measuring tape, or moved hands off hips (Ness et al., 2015). Leg length was measured in centimeters on the injured limb from the anterior superior iliac spine to the ipsilateral medial malleolus. The mean of the normalized reach distances in each direction and a composite (COMP) reach distance were calculated (Anguish and Sandrey, 2018; Hale et al., 2007).

Participants in G1 and G2 were required to rest for 20 minutes following electrical stimulation to avoid any potential influence from physical activity. To prevent the loosening of electrode pads caused by BT or impedance changes due to sweating, which could compromise the efficacy of tDCS, participants in G3 and G4 performed 20 minutes of BT after receiving electrical stimulation. Regardless of group allocation, all participants underwent 3 sessions per week, totaling 12 supervised intervention sessions. Intervention was performed at the Sport Medical Rehabilitation Center of Beijing Sport University.

tDCS was administered using a constant-current device (StarStim 32, Neuroelectrics, Spain) via two saline-soaked circular sponge surface electrodes (25 cm²) placed on the scalp. The tDCS procedure was conducted with the participant seated, as this position was conducive to maintaining stable electrode contact and ensuring the effectiveness of stimulation. The anode electrode was positioned at the lower extremity motor cortex (Cz), while the cathode electrode was placed over the right orbital region (Fp2) (Needle et al., 2017). During the anodal tDCS procedure, the current intensity was gradually increased to 2 mA during the first 30 seconds, maintained for 19 minutes, and then gradually decreased to 0 mA over the last 30 seconds (Ma et al., 2020). During the sham tDCS procedure, the current was gradually increased to 2 mA over a 30-second period and then immediately turned off. The stimulation impedance maintained below 10 kΩ. Any discomfort should be promptly reported to the staff, who would determine whether to continue based on the actual circumstances and the participant's consent. Prior to the formal tDCS session, participants were informed about the potential adverse reactions and acceptable levels of this intervention.

BT refers to the progressive hop-to-stabilization balance (PHSB) training program proposed by McKeon (McKeon et al., 2008). This program included five practice contents, and each practice contained seven difficulty levels. Each training session lasted for 20 minutes. Before progressing to the next level, participants were required to demonstrate error-free completion of the task.

Data analysis was conducted using SPSS version 24.0 (IBM Corp, Armonk, NY, USA). The Shapiro-Wilk (S-W) test was utilized to assess the normality of the variables, while the Levene test was applied to evaluate the homogeneity of variances. As the data met these assumptions, parametric tests were applied for further analyses. To assess the differences between groups in terms of demographics (age, height, weight, and CAIT score), a one-way ANOVA was employed. Separate 4 (groups) × 2 (times) repeated-measures ANOVA were conducted to evaluate changes in dependent variables attributable to different interventions. The independent variables were group (G1 vs. G2 vs. G3 vs. G4) and time (Pre vs. Post). Post-hoc comparisons were conducted using the Bonferroni correction method. Mean and SDs were calculated for all continuous variables. Effect sizes were quantified using partial eta squared (η²_p), where values less than 0.01, between 0.01 and 0.06, and greater than or equal to 0.14 represent small, medium, and large effect sizes, respectively (Fritz et al., 2012). The level of statistical significance was established at p < 0.05.

No side effects or adverse events were reported. Study procedures and the number of subjects at each stage were presented in Figure 1. No significant differences were found in demographic characteristics, baseline BESS scores, and YBT scores between the groups (p > 0.05) (See Table 1).

Since the double-limb stance resulted in 0 points, this study focused on single-limb and tandem stances on firm and foam surfaces, as well as the total BESS score. We observed significant group × time interactions for single-leg stance on a firm surface (Sfi) (F = 3.958, p = 0.017, η²_p = 0.217), single-leg stance on a foam surface (Sfo) (F = 4.632, p = 0.008, η²_p = 0.303), tandem stance on a foam surface (Tfo) (F = 9.189, p < 0.001, η²_p = 0.463) and total BESS (F = 12.521, p < 0.001, η²_p = 0.540) scores, but not for tandem stance on a firm surface (Tfi) (p > 0.05). Compared to baseline measures, G1, G3, and G4 scored fewer errors on posttest measures for Sfi (G1: F = 12.100, p = 0.001, η²_p = 0.274; G3: F = 16.900, p < 0.001, η²_p = 0.346; G4: F = 10.000, p = 0.003, η²_p = 0.238), Sfo (G1: F = 21.993, p < 0.001, η²_p = 0.407; G3: F = 49.485, p < 0.001, η²_p = 0.607; G4: F = 29.086, p < 0.001, η²_p = 0.476), and the total BESS (G1: F = 24.071, p < 0.001, η²_p = 0.429; G3: F = 74.187, p < 0.001, η²_p = 0.699; G4: F = 44.232, p < 0.001, η²_p = 0.580). G3 and G4 also showed a significant decrease in Tfi (G3: F = 11.757, p = 0.002, η²_p = 0.269; G4: F = 4.452, p = 0.043, η²_p = 0.122) and Tfo (G3: F = 29.594, p < 0.001, η²_p = 0.480; G4: F = 27.077, p < 0.001, η²_p = 0.458) scores, while this change was not observed in G1 (p > 0.05). Moreover, significant differences were observed among groups on posttest measures of Sfi, Sfo, Tfi, and total BESS scores (p < 0.05). However, post-hoc comparisons indicated that no significant differences existed between G1, G3, and G4 (p > 0.05), with the inter-group differences primarily attributed to G2. (See Table 2).

There were no significant group × time interactions found for ANT, PM, PL, and COMP (p > 0.05); however, there was a significant main effect of time (ANT: F = 5.235, p = 0.029, η²_p = 0.141; PM: F = 12.889, p = 0.001, η²_p = 0.287; PL: F = 15.741, p < 0.001, η²_p = 0.330; COMP: F = 17.832, p < 0.001, η²_p = 0.358). No significant changes were found in either G1 or G2 across all directions between the pretest and posttest measures (p > 0.05). Both G3 and G4 exhibited significantly greater PM (G3: F = 10.347, p = 0.003, η²_p = 0.244; G4: F = 8.506, p = 0.006, η²_p = 0.210), PL (G3: F = 11.712, p = 0.002, η²_p = 0.268; G4: F = 9.167, p = 0.005, η²_p = 0.223), and COMP (G3: F = 13.995, p = 0.001, η²_p = 0.304; G4: F = 11.460, p = 0.002, η²_p = 0.264) reach distances in the posttest compared to the pretest. Only G3 revealed a farther ANT reach distance after 4 weeks of intervention (F = 4.692, p = 0.038, η²_p = 0.128). Also a significant main effect of group was observed (ANT: F = 3.598, p = 0.024, η²_p = 0.252; PM: F = 4.846, p = 0.007, η²_p = 0.312; PL: F = 3.281, p = 0.033, η²_p = 0.235). Consistent with the main effect finding, significant differences were observed among groups on posttest measures (p < 0.05). However, post-hoc comparisons indicated that no significant differences existed between G1, G3, and G4 in terms of ANT, PM, PL, and COMP (p > 0.05), with the inter-group differences primarily attributed to G2 (See Table 3).

After a 4-week intervention with BT, or the combination of tDCS and BT, CAI athletes exhibited significant reductions in the total BESS score and error scores for Sfi, Sfo, Tfi, and Tfo. These findings collectively suggest an enhancement in static balance ability, which aligns with prior research outcomes. Following a 4- to 6-week BT program, subjects with CAI or FAI demonstrated significant improvements in static postural control, as evaluated by the center of pressure excursion, BESS, and time-to-boundary (Bernier and Perrin, 1998; Hall et al., 2018; McKeon et al., 2008; Mettler et al., 2015; Rozzi et al., 1999). BT could diminish proprioceptive deficits associated with ligamentous injury to the ankle (Freeman et al., 1965). However, a systematic review highlights a paucity of evidence supporting the effectiveness of BT in improving instrumental postural control among patients with CAI (McKeon and Hertel, 2008). The potential reasons for this outcome include: the measurement instruments may lack the sensitivity to detect improvements from BT in patients with CAI, and static single-limb standing may not provide sufficient challenges to elicit discernible changes.

In the current study, we found that after 4 weeks of tDCS, patients with CAI demonstrated significant improvements in Sfi, Sfo, and total BESS scores. tDCS, either as a standalone intervention or in conjunction with other therapeutic approaches, has been shown to enhance balance in adults with neurological disorders (Beretta et al., 2022). CAI patients experience diminished somatosensory input, leading to a greater reliance on visual cues. During the closed-eye test, the lack of visual input exacerbates the reduced proprioceptive information from the lower extremities, resulting in higher balance error scores (Kwon, 2018). Anodal tDCS can positively affect cortical excitability and thus address the maladaptive neuroplasticity following joint instability (Jeffery et al., 2007; Nitsche and Paulus, 2001). This improvement in neural plasticity may contribute to more efficient neural processing of sensory information relevant to balance, such as proprioceptive and vestibular cues, and enhance the generation of appropriate motor responses for maintaining balance. It has been shown that tDCS can enhance the long-term potentiation effects in the motor cortex, promoting the formation of new synaptic connections and strengthening existing ones (Monte-Silva et al.,2013). Additionally, it could enhance the balance within asymmetric neural networks between brain hemispheres and modify the functional connectivity of various brain regions involved in both direct and indirect pathways of postural control (Beretta et al., 2022). These mechanisms likely contribute to the observed improvements in the static balance ability of CAI athletes in our study following a 4-week tDCS intervention.

We found that both tDCS combined with BT and BT alone significantly improved PM, PL, and COMP reach distances, compared to tDCS or sham tDCS alone, neither of which resulted in any significant changes. These improvements indicate a substantial enhancement in dynamic postural control. BT is beneficial for enhancing dynamic postural control among patients with CAI (Anguish and Sandrey, 2018; Burcal et al., 2019; Hale et al., 2007; McKeon et al., 2008). Specifically, participants in the BT group exhibited significant improvements in PM and PL reach distances of the Star Excursion Balance Test compared to their pretest measurements (Cruz-Diaz et al., 2015; McKeon et al., 2008). These enhancements might be attributed to the diminished restrictions on the sensorimotor system resulting from BT (McKeon et al., 2008). A significant improvement in COMP was observed following BT in the present study. However, the increase in ANT was observed only in the tDCS combined with BT group, with a moderate effect size (η²_p = 0.128). Given that this direction is primarily associated with reductions in dorsiflexion mobility and posterior talar glide (Vicenzino et al., 2006), and no significant differences were found among different interventions in this study, we should cautiously conclude that tDCS combined with BT may improve ANT reach distance in CAI athletes.

No significant changes were observed in any direction in the tDCS group after 4 weeks of intervention. This result diverges from prior research findings, which indicated that anodal tDCS applied to the leg area of the primary motor cortex enhances dynamic performance in young adults (Kaminski et al., 2016; Hou et al., 2022). However, to date, for patients with CAI, tDCS has been combined with other forms of training, including eccentric training (Bruce et al., 2020) and short-foot exercises (Ma et al., 2020). tDCS over leg motor enhances the excitability of the motor cortex and modifies the firing strategies of motor units (Dutta et al., 2015), and enhances locomotor adaptation aftereffects (Kaski et al., 2012). This suggests that tDCS has the potential to serve as an effective adjunctive rehabilitation treatment to enhance postural control. The reason tDCS alone did not positively affect YBT results in CAI athletes in this study may be attributed to the multifaceted nature of dynamic postural stability. This function relies on the integration of multiple sensory inputs (e.g., proprioceptive, vestibular, visual) and motor outputs. In the absence of complementary effects from other training modalities, tDCS alone might have been insufficient to effectively coordinate and optimize the complex neuro-muscular processes required for maintaining dynamic balance in CAI patients. When tDCS was combined with BT, a significant increase in the reach distance could be observed, which also indirectly verified this reason.

Regarding the results of static and dynamic balance, no intergroup differences were found in this study. Specifically, tDCS alone, BT alone, and the combination of these two produced comparable effects.This finding diverges from our hypothesis and previous research Previous studies have indicated that combining tDCS with other interventions may offer greater benefits for postural control compared to using only one type of intervention. Bruce et al. revealed that a 4-week tDCS in conjunction with eccentric training improved the dynamic postural stability index by enhancing neural drive to stabilizing muscles in patients with CAI (Bruce et al., 2020). Jafarzadeh et al. additionally noted that combining 2 weeks of tDCS with physical training led to more significant enhancements in dynamic stability, as measured by the Biodex Balance System (Jafarzadeh et al., 2019). In older adults at high risk of falling, the combination of bilateral cerebellar and primary motor cortex anodal tDCS along with postural training led to significant enhancements in postural control and balance. However, neither the 2-week regimen of postural training by itself nor the standalone application of bilateral cerebellar anodal tDCS resulted in comparable levels of improvement (Yosephi et al., 2018). Additionally, high-definition tDCS combined with short-foot exercise resulted in significantly greater improvement in the performance on the YBT in CAI individuals (Ma et al., 2020). However, a combination of 4-week anodal tDCS with foot core exercise did not affect static balance in healthy young adults. Perhaps because of the stimulus target effect (Xiao et al., 2022). Although the structural and functional alterations in the motor cortex may influence balance, the motor cortex’s role in postural control is not as crucial as that of the cerebellum, the primary balance regulation center (Takakusaki, 2017). This perspective is supported by the findings of Yosephi (Yosephi et al., 2018). Additionally, a meta-analysis indicated no significant difference between tDCS alone and its combination with another intervention in treating neurological disorders (Beretta et al., 2022), which supports our findings. These discrepancies could be attributed to methodological variations in study design, leading to divergent outcomes. Another potential reason could be the high motor abilities of the subjects in this study, which may have caused a ceiling effect, thereby making further improvements in balance ability challenging.

Firstly, we did not conduct a longer follow-up after the intervention, which could provide valuable insights into the long-lasting impacts of different interventions targeting postural control for individuals with CAI. Secondly, all participants were collegiate athletes competing at the national level or higher. Despite experiencing recurrent ankle sprains, their adaptability and athletic abilities may partially compensate for the postural impairments, making it challenging to discern changes in their postural control. Finally, the paper failed to include posturography as a measurement tool, which is widely regarded as the gold standard for assessing postural control. Future studies should consider integrating posturography to achieve more precise and reliable observations of changes in postural control.

Four weeks of tDCS, BT, and the combination of these two interventions significantly improved static postural stability in athletes with CAI. However, only intervention methods incorporating BT were effective in enhancing dynamic stability. Notably, the combined program did not offer additional benefits over the individual interventions.