The effect of MI-BCI combined with tDCS in early rehabilitation after anterior cruciate ligament reconstruction monitored by resting-state fMRI: the first case report Authors: Chongwen Zuo, Yi Yin, Chong Zhang, Xiaoyan Ma, Zhiyang Zheng, Shan Wang, Chaoqun Ye Published: 2026 ABSTRACT AbstractBackgroundTo recover lower-limb motor function is a primary goal for rehabilitation after anterior cruciate ligament (ACL) reconstruction. Although quantitative testing and questionnaire evaluation provide a lot of valuable information, functional magnetic resonance imaging (fMRI) provides a powerful method to preliminarily explore the neural recovery process of novel rehabilitation interventions. In this case, we aimed to investigate changes in functional connectivity (FC) of motor imagery-based brain-computer interface (MI-BCI) combined with tDCS intervention in early post-ACL reconstruction rehabilitation, monitored via resting-state fMRI.Case presentationA 38-year-old man patient who underwent left ACL reconstruction received MI-BCI combined with tDCS intervention 5sessions/peek for 4 weeks. before and after the intervention, the resting-state fMRI was assessed. Seed-based analysis was performed using the primary motor cortex (M1) and supplementary motor area (SMA) as the seeds, representing the sensorimotor network (SMN). The results showed increased connectivity between the SMN and specific motor-related (e.g., precentral gyrus, postcentral gyrus, pallidum and paracentral lobule) and sensory information processing areas (e.g., calcarine, lingual gyrus, calcarine, thalamus, posterior cingulum). In addition, decreased FC between both M1 and the bilateral precentral gyrus was observed, and the FC between the left M1, SMA, and the right frontal midline regions was also reduced.ConclusionsOur preliminary results showed that MI-BCI combined with tDCS could be an effective method to recovery the lower-limb motor function and promote neural remodeling of an ACL reconstruction patient in early stage. Alterations of the SMN functional connectivity after intervention seems to partially explain the neural remodeling process, if applied on a large sample sizes, could provide more information about specific motor area.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12883-025-04605-7.Keywords: Case report, Resting-state fMRI, MI-BCI, tDCS, Anterior cruciate ligament reconstruction, Neural remodeling Background Anterior cruciate ligament (ACL) has become one of the most common and easily damaged knee ligaments in sports and military training affecting young and active people. The extensive interest in ACL ruptures is related to the subsequent articular degeneration and severe dysfunction, resulting in skeletal muscle functional limitations not only in athletics but also in daily living. Thus, restoring lower limb motor function and stability has become the primary goal after ACL reconstruction. Moreover, the newest study has indicated that increased neural activity may reflect central neuroplastic strategies to preserve motor functionality after ACL reconstruction [1]. Postsurgical rehabilitation guidelines for the orthopedic clinician recommended a common treatments for ACL reconstruction patients [2]; Specifically, current rehabilitation strategies for ACL reconstruction primarily focus on physical therapy and strengthening exercises, but they often suffer from limited effectiveness in restoring full motor function. Previous studies have shown that neuroplastic changes are slow during early rehabilitation, and motor recovery can be hindered by poor central nervous system engagement. Hence, it is very urgent to find a safe, effective and easy conduct and alternative therapy. Transcranial direct-current stimulation (tDCS) is a well-established non-invasive neuromodulation technique that has shown promising results in improving motor function in various neurological and musculoskeletal conditions, including stroke, spinal cord injury, and knee osteoarthritis [3, 4]. Anodal stimulation over the primary motor cortex (M1) has been associated with increased cortical excitability, improved voluntary muscle activation, and enhanced motor learning through long-term potentiation-like mechanisms. In the context of ACL reconstruction, persistent quadriceps weakness and impaired proprioception have been linked to central nervous system hypoactivity and maladaptive plasticity [1]. Although direct evidence on the use of tDCS in ACL patients is limited, emerging studies have begun to explore its adjunctive role. For instance, Tohidirad et al. [5] demonstrated that a battery-operated tDCS device, applied over the M1 with large electrodes(5 × 7 cm) at the intensity of 2 mA current for 20 min for 10 sessions, enhanced the effect of conventional physiotherapy (PT) on balance and muscle performance in athletes recovering from ACL injury. Their study showed that one month after treatment, the displacement of the center of pressure significantly decreased in the intervention group, indicating improved postural control, while no significant changes was observed in the control group. Additionally, the average power of flexor and extensor muscles increased in both groups, but the rise in the intervention group was significantly larger. These findings highlight the potential of combining a-tDCS with PT to induce lasting improvements in functional outcomes for ACL injury rehabilitation. For another, motor imagery based brain-computer interface (MI-BCI) neuro rehabilitation paradigms have emerged as a novel iatrotechnics on motor function and brain function recovery [6]. The MI-BCI was combined with tDCS offers a promising alternative by targeting neural reorganization and enhancing motor learning early in the recovery process, but this combined strategy that has not been fully explored in ACL rehabilitation. Building on these findings, our study explores the application of MI-BCI combined tDCS as a novel neuromodulation strategy in in a patient undergoing early post-ACL reconstruction rehabilitation. By monitoring changes in functional connectivity (FC) using resting-state fMRI, we aim to provide mechanistic insight into how such combined stimulation may facilitate neural plasticity and accelerate functional recovery. Case presentation Here we report a 38-year-old male pilot, suffered from left ACL rupture due to football entertainment activity. He received ACL reconstruction surgery after admission in orthopedics unit, and then transferred to rehabilitation unit one month late, and participated in two resting-state fMRI session (31 days apart) in our study. After one month of regular rehabilitation in the orthopedics unit, the knee range of motion (ROM) had not returned to 90°, and it still showed extensor weakness and quadriceps atrophy. Prior to receiving treatment, the patient had been enrolled and informed of the rehabilitation trial procedure and signed informed consent. In addition to routine rehabilitation training, a combined intervention of MI-BCI and tDCS intervention was carried out in a quiet, independent room for 5 sessions (in the late afternoon) per week over 4 consecutive weeks. Each session began with 30 min of MI-BCI training. Electroencephalography (EEG) signals were recorded using a 32-channel EEG system (ZhenTec, NT1 Xi’an ZhenTec Intelligence Technology Co., Ltd., China), with electrodes placed according to the international 10–20 system. The patient first performed MI of leg flexion and extension training for10 times to collect EEG signals, which were then processed via common spatial patterns (CSP) and support vector machine (SVM) algorithm, with sensorimotor rhythm (SMR) band personalized (alpha:8–15; beta:16–30 Hz) to establish the individualized training threshold, and an individualized MI threshold was defined based on a classifier confidence level of ≥ 50%. During formal MI-BCI sessions, once the patient’s visual cued level met the MI threshold, the functional electrical stimulator (Arduino, Italy) was triggered to stimulate the target muscles and elicit actual movement. Immediately following MI-BCI training, offline anodal tDCS was applied over the leg motor area of the M1 contralateral to the affected knee (placed on right M1 area), with the cathode placed over the contralateral supraorbital (SO) region, following a well-established M1–SO montage frequently used in lower-limb rehabilitation [3, 4]. This montage has been shown to enhance M1 excitability, improve muscle strength and postural control, and reduce pain by mitigating the impact of central sensitization. Stimulation was delivered using the BrainDriver v2.1 Digital tDCS device (TheBrainDriver Inc., IL, USA) with two carbon rubber-based electrodes secured by headgear and covered by 4 cm diameter saline-soaked sponge pads. The stimulation was applied at a constant current of 2 mA for 20 min per session, resulting in a current density of approximately 0.039 mA/cm². The choice of offline stimulation was based on evidence suggesting that pre-activation of cortical excitability can enhance subsequent functional reorganization without interfering with MI-related EEG decoding. Additionally, the primary clinical outcome was assessed by International Knee Documentation Committee Subjective Knee Form (IKDC). Resting-state fMRI data were acquired at the Air Force Medical Center using a 3.0T MRI scanner (GE Discovery MR750, USA). Blood oxygenation level–dependent (BOLD) images were obtained using an echo planar imaging (EPI) sequence (TR/TE = = 2000/30 ms, matrix size = 64 × 64, FOV = 220 × 220 mm², number of slices = 36, slice thickness = 3 mm), with a total scanning time of 6 min. High-resolution T1-weighted anatomical images were also acquired during the first imaging session using a 3D TFE sequence (voxel size: 1 × 1 × 1 mm). Preprocessing and FC analyses were performed using DPARSF V4.5 (http://rfmri.org/DPARSF) and SPM12 (https://www.fil.ion.ucl.ac.uk/spm) implemented in MATLAB R2022b. The preprocessing steps included slice timing correction, realignment for motion correction, normalization to the MNI space, spatial smoothing with a 6 mm FWHM Gaussian kernel, linear trend removal, and nuisance covariate regression including 24 motion parameters, cerebrospinal fluid (CSF), white matter (WM), and global signal. A temporal bandpass filter (0.01–0.08 Hz) was applied to reduce physiological noise. Seed-based FC analysis focused on bilateral M1 and supplementary motor area (SMA), key hubs in the sensorimotor network (SMN) [7]. These regions of interest (ROIs) were defined using HAMT atlas in standardized MNI coordinates derived from previous literature. Individual-level correlation values were converted to z-scores using Fisher’s r-to-z transformation. To identify ROIs with significant FC changes before and after intervention, a threshold of voxel-level p < 0.001, cluster-level p < 0.05 was used with family-wise error (FWE) correction. Custom MATLAB scripts were used to generate statistical parametric maps, and all analyses are available upon request to ensure reproducibility. After receiving 4 weeks of MI-BCI combined tDCS treatment in rehabilitation unit, patients completed all courses with interest and demonstrated a high level of compliance, and no adverse events occurred. The patient’s shower improvements in knee ROM and muscle strength. The knee ROM recovered to 117°, and the flexion/extension strength of the left leg increased from 4.3 kg to 15.6 kg/9.9 kg to 20.9 kg. The therapeutic effect continued to improve in the follow-up period (one month after post-treatment). The IKDC score showed comparable improvement from 28 points to 67 points. Additionally, we observed increased FC between the bilateral supplementary motor areas (SMA) and the bilateral precentral gyrus, calcarine, postcentral gyrus, and paracentral lobule. Similarly, FC between the bilateral SMA and the bilateral lingual gyrus showed a marked increase. The FC between both primary motor cortices (M1) and the bilateral SMA, lingual gyrus, paracentral lobule, right calcarine, right pallidum, right thalamus, and left posterior cingulum were also elevated post-treatment. However, there was decreased FC between both M1 and the bilateral precentral gyrus, and the FC between the left M1, SMA, and the right frontal midline regions was also reduced (See Table 1; Fig. 1). Table 1Regions that showed a summary of FC correlation coefficient values across the M1/SMA from before to after a 4-week MI-BCI combined tDCS intervention in a patient with ACL reconstructionRegionM1.LM1.RSMA.LSMA.RPost > pret-statistict-statisticPreCG.L−0.010−0.0080.0160.016PreCG.R−0.0070.0000.0160.014MFG.R−0.007--−0.031SMA.L0.0260.022----SMA.R0.0270.021----PCG.L--0.026----CAL.L----0.0690.061CAL.R0.0430.0280.1260.116LING.L0.0300.0140.1050.091LING.R0.0330.0150.1140.102PoCG.L----0.0350.035PoCG.R----0.0210.018PCL.L0.0220.0150.0460.043PCL.R0.0370.0320.0610.063PAL.R0.0220.045----THA.R0.0230.037----L left, R right, M1 primary motor cortex, SMA supplementary motor area, PreCG precentral gyrus, PoCG postcetral gyrus, MFG middle frontal gyrus, PCL paracentral lobule, PCG posterior cingulate gyrus, CAL calcarine, PLA pallidum, THA thalamus, LING lingual gyrus Regions that showed a summary of FC correlation coefficient values across the M1/SMA from before to after a 4-week MI-BCI combined tDCS intervention in a patient with ACL reconstruction L left, R right, M1 primary motor cortex, SMA supplementary motor area, PreCG precentral gyrus, PoCG postcetral gyrus, MFG middle frontal gyrus, PCL paracentral lobule, PCG posterior cingulate gyrus, CAL calcarine, PLA pallidum, THA thalamus, LING lingual gyrus Differences in bilateral M1 and SMA FC following a 4-week MI-BCI combined tDCS intervention. Color bar indicates the intensity of FC. L, left; R, right. M1, primary motor cortex; SMA, supplementary motor area; PreCG, precentral gyrus; PoCG, postcetral gyrus; MFG, middle frontal gyrus; PCL, paracentral lobule; PCG, posterior cingulate gyrus; CAL, calcarine; PLA, pallidum; THA, thalamus; LING, lingual gyrus Discussion and conclusions Previous studies have explored the synergistic application of tDCS combined BCI training in neurorehabilitation settings, particularly among stroke patients. For instance, Kruse et al. [6] demonstrated that concurrent tDCS and motor imagery BCI training improved motor outcomes in post-stroke hemiparesis. Similarly, Lima et al. [8] reported enhanced cortical reorganization and physical function after tDCS and BCI intervention. These findings support the general efficacy of integrating central and peripheral stimulation to augment neuroplasticity. Our study expands on this approach by translating such neurorehabilitation strategies to the orthopedic domain—specifically, early-stage ACL reconstruction rehabilitation. Although ACL reconstruction patients do not suffer overt brain injury, studies have shown that surgical trauma, pain, and disuse can induce cortical hypo-excitability and maladaptive plasticity. Thus, introducing MI-BCI with tDCS in this context may help restore cortico-muscular coherence and facilitate motor relearning. To our knowledge, this is the first study to apply this combined approach in an orthopedic population, highlighting its translational potential beyond neurological disorders. To contextualize the observed improvements, we compared the patient’s recovery trajectory with existing benchmarks from previous studies on early post-ACL reconstruction rehabilitation. For example, previous literature reports that knee ROM often remains limited to approximately 90°–100° within the first 4–6 weeks postoperatively without targeted neuromodulation [2, 5]. In our case, the patient’s ROM improved from < 90° to 117° after the four-week intervention, which exceeds the typical recovery range. Similarly, muscle strength recovery in the early phase is often slow, with quadriceps torque typically reaching only 40–60% of the contralateral leg during the first month [1]. Our case demonstrated an observable increase in both flexion and extension strength, with increases of 3–5 fold, indicating a potentially accelerated neuromuscular recovery. These results suggest that the combined MI-BCI and tDCS intervention may have facilitated early motor function gains beyond those typically observed in conventional rehabilitation. However, further studies with controlled comparisons are needed to validate this observation. To date, our clinical results have preliminarily demonstrated the feasibility and potential benefit of MI-BCI combined with tDCS during early-stage ACL reconstruction rehabilitation, with resting-state fMRI serving as an exploratory biomarker of neuroplasticity. After 4 weeks of treatment, knee function improved by more than 58%, suggesting a positive responsiveness to the combined intervention. This improvement in motor function was accompanied by increased functional connectivity between the bilateral SMA and several non-motor-related regions, including the paracentral lobule, calcarine, and lingual gyrus. These findings suggest that the MI-BCI combined with tDCS intervention may enhance not only motor control but also sensory information processing. The involvement of broader network regions may facilitate more effective integration of motor planning and execution, contributing to a more efficient recovery process during rehabilitation. As a key component of the SMN, the M1 also plays an essential role in motor preparation and execution. The SMA integrates signals of frontal planning with sensorial, proprioceptive and cognitive information from other brain regions and forwards it to the M1 [9]. In this study, the enhanced functional connectivity between M1 and the bilateral SMA, along with other regions involved in sensory processing, such as the lingual gyrus, paracentral lobule, right calcarine, left pallidum, right thalamus, and left posterior cingulum, highlights the broader network involvement in both motor control and sensory integration post-treatment. Given the aforementioned findings, the combined MI-BCI and tDCS intervention may facilitate the integration of motor planning and execution circuits with sensory feedback systems, thus contributing to a more efficient recovery process. The involvement of sensory regions, such as the lingual gyrus, calcarine, and paracentral lobule, indicates that the intervention likely improves not only motor control but also sensory information processing, which is critical for motor rehabilitation. Therefore, the enhanced FC observed may reflect experience-dependent brain plasticity, potentially promoted by the co-activation of M1 and SMA during MI-BCI sessions combined with anodal tDCS. This co-activation may have synchronized neuronal firing patterns, in line with Hebbian principles (“cells that fire together, wire together”), thereby reinforcing motor networks involved in lower-limb function, which consistent with previous studies demonstrating plasticity-driven gains in post-stroke rehabilitation [8]. Furthermore, the current study also extends the existing knowledge by providing evidence of its applicability in non-stroke, musculoskeletal injury rehabilitation. Interestingly, we also observed decreased FC between M1 and the bilateral precentral gyrus, as well as a reduction in FC between left SMA and the right frontal midline regions. These changes may reflect a more efficient reorganization of neural resources, with a focus on task-relevant motor networks, while reducing reliance on non-task-relevant or compensatory regions typically observed during early recovery stages. Although this case suggests that the combined MI-BCI and tDCS intervention led to improvements in motor function and knee mobility, it is important to note that these results are based on a single patient. Given the limitations of case reports, such as the lack of control groups and the small sample size, further studies with larger sample sizes and randomized controlled trials (RCTs) are necessary to confirm the efficacy and generalizability of this intervention. Thus, these preliminary results should be interpreted with caution until validated by more robust clinical trials. In conclusion, this case study provides preliminary evidence that the combined use of MI-BCI and tDCS may represent a promising, non-invasive strategy for facilitating early neural remodeling and motor function recovery in ACL reconstruction patients. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3.