Efficacy and safety of high-definition transcranial direct current stimulation combined with digital rehabilitation on upper limb function in stroke patients: study protocol for a randomized, double-blind, sham-controlled confirmatory trial

Abstract

Background

Stroke remains a leading cause of long-term disability worldwide, with a significant proportion of survivors suffering from persistent upper limb impairments despite conventional rehabilitation. High-definition transcranial direct current stimulation (HD-tDCS) has emerged as a promising neuromodulation technique capable of precise cortical targeting, potentially enhancing motor recovery when combined with structured rehabilitation. This study aims to confirm the efficacy and safety of HD-tDCS combined with digital upper limb rehabilitation in improving upper limb motor function among stroke survivors.

Methods

This multicenter, randomized, parallel-group, double-blind, sham-controlled superiority trial will enroll 64 participants with chronic unilateral stroke who will be randomized to receive either active HD-tDCS or sham stimulation, both combined with standardized digital upper limb rehabilitation. The intervention will consist of 10 sessions over 4 weeks (40 min per session). Each session will include 10 min of HD-tDCS at rest targeting the ipsilesional primary motor cortex and anterior intraparietal sulcus, followed by 20 min of task-based digital rehabilitation with concurrent HD-tDCS, and a final 10 min of digital rehabilitation alone. The primary outcome will be the postintervention change in the Fugl–Meyer Assessment of Upper Extremity score from baseline. Secondary outcomes will include grip and pinch strength, 9-Hole Pegboard Test, Box and Block Test, finger-tapping performance, Short Form–36, and cortical activation measured via integrated functional near-infrared spectroscopy (fNIRS). fNIRS will help characterize task-related cortical hemodynamic responses in the stimulated area, which are associated with motor recovery.

Discussion

This trial provides promising exploratory findings and is designed to rigorously assess both behavioral and hemodynamic brain responses. By integrating real-time brain monitoring, this study seeks to contribute to existing evidence on the mechanisms underlying HD-tDCS-enhanced neuroplasticity and inform the development of future precision neurorehabilitation strategies.

Trial registration

This trial has been registered at the Clinical Research Information Service (Registration number: KCT0010556). Registration date: May 28, 2025.

Administrative information

Title {1}	Efficacy and safety of high-definition transcranial direct current stimulation combined with digital rehabilitation on upper limb function in stroke patients: study protocol for a randomized, double-blind, sham-controlled confirmatory trial
Trial registration {2a and 2b}	This trial has been registered at the Clinical Research Information Service (Registration number: KCT0010556)
Protocol version {3}	Protocol version 3.4, April 30, 2025
Funding {4}	This work was supported by the Korea Medical Device Development Fund grant funded by the Korean government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, and the Ministry of Food and Drug Safety) (KMDF‑RS‑2022–00140478)
Author details {5a}	Jinuk Kim1†, Su-Hyun Lee1†, Eunmi Kim1, Seung-Rok Kang2, Yun-Ju Jo2, Ji-Eon Yun2, Seung-Hyeun Lee3, Myoung-Hwan Ko4*, and Yun-Hee Kim1,5* 1Department of Physical and Rehabilitation Medicine, Sungkyunkwan University School of Medicine, Suwon, Republic of Korea 2Biomedical Research Institue, Jeonbuk National University Hospital, Jeonju, Korea 3Global Health Technology Research Center, College of Health Science, Korea University, Seoul, Republic of Korea 4Department of Physical Medicine and Rehabilitation, Jeonbuk National University Medical School, Jeonju, Republic of Korea 5Myongji Choonhey Rehabilitation Hospital, Seoul, Republic of Korea †These authors contributed equally to this work as co-first authors *These authors contributed equally to this work as co-corresponding authors
Name and contact information for the trial sponsor {5b}	Cybermedic Inc 1 F, Bldg. 1, Free Trade Zone Standard Factory, 191–17 Yeongdeung-dong, Iksan-si, Jeollabuk-do, Korea Contact name: jeonghc@cybermedic.co.kr
Role of sponsor {5c}	The sponsor funded the study and engaged a contract research organization to support trial monitoring and regulatory compliance. The sponsor played no role in data, interpretation of the findings, or decision to publish

Introduction

Background and rationale {6a}

Stroke remains a leading cause of long-term disability, particularly among aging populations [1]. Despite the decline in mortality rates with improved acute stroke care, persistent motor impairments continue to drive long-term disability and limit functional recovery. Over 40% of stroke survivors continue to experience upper limb disability even 3 years after onset [2], which highlights the clinical need for effective interventions that facilitate recovery and minimize long-term disability. Although conventional rehabilitation remains the standard of care, novel neurorehabilitation strategies that actively promote neuroplasticity have been receiving increasing interest [3].

Recently, neuromodulation techniques using electrical or magnetic stimulation have emerged as novel strategies for regulating neuroplasticity [4]. Transcranial direct current stimulation (tDCS), one of the most widely studied noninvasive techniques [5], has been found to modulate neural excitability by either enhancing or suppressing cortical activity depending on the polarity of stimulation. In fact, accumulating evidence suggests that anodal tDCS applied over the primary motor cortex (M1) increases cortical excitability in the targeted region [6]. Moreover, one study showed that combining such stimulation with conventional motor rehabilitation promotes greater improvements in motor function than would rehabilitation alone [7]. However, conventional tDCS typically uses a single pair of large electrodes and is unable to precisely target specific cortical regions. Unintended inhibition from cathodal stimulation of nontarget areas may also reduce its overall effect.

To address the limitations of conventional tDCS, a more advanced neuromodulation technique called high-definition tDCS (HD-tDCS) has been developed [8]. By utilizing multiple small electrodes, HD-tDCS enables focal and selective stimulation of target brain regions while minimizing unintended effects on adjacent areas. This improved targeting is particularly relevant given the importance of precisely modulating motor-related cortical areas in stroke rehabilitation [9]. Although the M1 has traditionally been a key target in motor recovery, growing evidence has identified the anterior intraparietal sulcus (aIPS) as an additional region associated with motor coordination and hand function [10]. Indeed, an exploratory clinical trial applying HD-tDCS over the M1 and aIPS combined with motor training confirmed its safety and revealed positive effects on upper limb and hand function in stroke survivors [11].

Recent advances in neuroimaging, particularly in functional near-infrared spectroscopy (fNIRS), have enabled real-time monitoring of cortical activation during rehabilitation [12]. Previous exploratory studies have used fNIRS to assess hemodynamic responses induced by HD-tDCS and motor training in stroke patients [11]. These studies revealed stimulation-related activation changes over the M1 and aIPS, supporting the role of targeted neuromodulation in motor recovery. Although longitudinal data suggest potential neural adaptation, prior analyses have focused primarily on resting-state signals. To address this limitation, the present study seeks to examine task-related hemodynamic responses during motor execution following HD-tDCS, providing a more functionally relevant marker of neuroplasticity.

Objectives {7}

This confirmatory trial aims to determine the efficacy and safety of a novel multimodal intervention that combines HD-tDCS with digital upper limb rehabilitation among stroke patients. Building on previous exploratory findings, this study will retain the same stimulation parameters (target sites, intensity, and duration) while expanding the sample size to ensure sufficient statistical power. The digital rehabilitation platform enables structured, adaptive, and highly reproducible motor training that engages both proximal and distal upper limb functions. To further optimize individual responses, we will be incorporating an integrated HD-tDCS–fNIRS system to monitor task-related hemodynamic changes during each session. This approach allows for real-time assessment of cortical engagement and provides a basis for future personalization of neuromodulatory interventions.

Trial design {8}

This multicenter, randomized, parallel-group, double-blind, superiority trial has been designed to confirm the safety and efficacy of tDCS in improving upper extremity motor function among stroke patients.

Methods: participants, interventions and outcomes

Study setting {9}

The trial will be conducted at two locations within the Republic of Korea, namely Jeonbuk National University Hospital (CUH IRB 2024–03–012) and Myongji Choonhey Rehabilitation Hospital (MJCHIRB-2024–01). The study will be performed in accordance with the principles of Good Clinical Practice and the Declaration of Helsinki.

Eligibility criteria {10}

The inclusion criteria are as follows:

• Age 19–80 years

• Unilateral stroke confirmed through computed tomography or magnetic resonance imaging ≥ 3 months prior to study inclusion

• Fugl–Meyer Assessment of Upper Extremity (FMA-UE) score of 25–58

The exclusion criteria are as follows:

• Presence of pre-existing neurological disorders other than stroke

• Presence of significant psychiatric disorders, such as schizophrenia or bipolar disorder

• Korean version of the Mini-Mental State Examination (K-MMSE) score < 18

• History of botulinum toxin injection or nerve block in the upper limb 1 month prior to providing consent

• History of surgical treatment involving the peripheral nerves, muscles, or tendons in the upper limb

• Participation in another clinical trial 1 month prior to providing consent

• Participants with contraindications for tDCS, such as implanted electronic medical devices (e.g., pacemakers or cardiopulmonary assist devices) or intracranial metallic implants

• Open wounds or lesions at the electrode attachment sites

• Uncontrolled epileptic seizures

• Pregnancy or lactation

Who will take informed consent? {26a}

Informed consent from the prospective clinical trial participants or their authorized representatives will be obtained by a physician serving as an investigator. To ensure that the potential participants develop a comprehensive understanding of the study prior to enrollment, investigators are responsible for clearly explaining the contents of the finalized Institutional Review Board (IRB)–approved informed consent form. This process includes providing an explanation of the study’s purpose, its potential benefits and risks, and establishing accessible channels for participants to contact both the investigator and the IRB for any inquiries that may arise during participation. Following consent acquisition, the investigator will promptly provide participants with a copy of the signed consent form.

Additional consent provisions for collection and use of participant data and biological specimens {26b}

Not applicable. No biological specimens will be collected.

Interventions

Explanation for the choice of comparators {6b}

A sham HD-tDCS condition, which will use the same electrode montage and device interface as the active stimulation group, will be the comparator for this study. This design was selected to isolate the specific neuromodulatory effects of HD-tDCS by controlling placebo responses and participant expectations. Sham stimulation consists of a brief ramp-up and ramp-down period to mimic the sensory experience of real stimulation without delivering active current, thereby maintaining double-blind conditions for both participants and assessors.

This comparator has been scientifically proven to be viable based on prior literature demonstrating the feasibility of sham-controlled neuromodulation trials in stroke populations. This approach allows for a rigorous assessment of the efficacy and safety of HD-tDCS combined with digital upper limb rehabilitation. Moreover, ethical considerations support the use of a sham control, given that all participants will be receiving the same structured rehabilitation protocol, ensuring that no group is deprived of standard care or active training opportunities.

Intervention description {11a}

Participants in both groups will undergo 10 intervention sessions over a 4-week period, with 3 sessions scheduled per week. Each session will last approximately 40 min and will be conducted in a clinical setting under the supervision of trained therapists. The intervention will comprise a structured combination of HD-tDCS and digital upper limb rehabilitation delivered in sequential phases (Fig. 1).

Fig. 1.

Study design and intervention session timeline. This figure illustrates the overall trial flow and detailed structure of each intervention session. The upper panel shows the sequence of study phases, including screening, randomization, pre-intervention evaluation (T0), intervention (10 sessions over 4 weeks), and postintervention evaluation (T1). The intervention will consist of either real or sham HD-tDCS combined with digital upper limb rehabilitation and concurrent fNIRS monitoring. The lower panel depicts the timeline of a single intervention session. Each session will begin with pre-intervention fNIRS monitoring (5 min), followed by 10 min of resting-state HD-tDCS and 20 min of concurrent HD-tDCS with digital rehabilitation. The session will conclude with 10 min of digital rehabilitation without stimulation and postintervention fNIRS monitoring (10 min). UL Rehab, upper limb rehabilitation; fNIRS, functional near-infrared spectroscopy; HD-tDCS, high-definition transcranial direct current stimulation

Each session will begin with 10 min of HD-tDCS applied while the participant is at rest. After this initial period, the next 20 min will involve concurrent HD-tDCS and task-based digital upper limb rehabilitation. This concurrent phase is designed to maximize stimulation-induced neuroplasticity by coupling cortical modulation with active motor execution. After 30 min of HD-tDCS, participants will complete an additional 10 min of digital rehabilitation without stimulation. The detailed protocols for HD-tDCS and digital rehabilitation are described below.

HD-tDCS

The HD-tDCS montage to be applied in this study will be identical to that used in our previous exploratory trial involving stroke patients [11]. The aforementioned trial revealed that the same dual-site configuration targeting the ipsilesional M1 and aIPS was feasible, safe, and preliminarily efficacious in enhancing upper limb motor function and cortical activation. Building upon those findings, the current confirmatory trial will employ an identical stimulation protocol to validate its clinical effectiveness in a larger, multicenter setting (Fig. 2).

Fig. 2.

HD-tDCS and fNIRS montage and digital rehabilitation setup. A HD-tDCS and fNIRS setup showing electrode and optode placement over the scalp. Red circles indicate anodes (C1 and CP1, 1 mA each), whereas blue circles indicate cathodes (Cz − 0.95 mA, CPz − 0.40 mA, FC3 − 0.25 mA, P3 − 0.35 mA, and CP5 − 0.05 mA) for patients with left hemisphere lesions. The montage is mirrored for right hemisphere lesions. The fNIRS montage consists of 24 sources and 24 detectors forming 80 measurement channels (green lines). B Representative digital rehabilitation tasks: (1) touchscreen-based tasks (whack-a-mole and turning off the lights), (2) handle rotation tasks (rotating a faucet, tightening a screw, starting a car, and turning on an oven), (3) air bulb squeeze tasks (bursting balloons and watering the garden), and (4) finger-tapping task (playing the piano). fNIRS, functional near-infrared spectroscopy; HD-tDCS, high-definition transcranial direct current stimulation

HD-tDCS will be administered using a multi-channel stimulator (NT Brain 48CH, Cybermedic Co., Ltd., Republic of Korea) and a seven-electrode configuration using ring-shaped electrodes (2-cm diameter, 2.55-cm² surface area). For patients with left hemisphere lesions, anodal electrodes delivering a current of 1.0 mA each will be placed at C1 and CP1 to target the hand area of the ipsilesional M1 and aIPS based on the international 10–20 electroencephalography system. Surrounding cathodal electrodes will be positioned at Cz (− 0.95 mA), FC3 (− 0.25 mA), CPz (− 0.40 mA), CP5 (− 0.05 mA), and P3 (− 0.35 mA) for even distribution of the return current. For patients with right hemisphere lesions, the montage will be mirrored symmetrically. This setup yields a current density of approximately 0.392 mA/cm² at each anodal site, consistent with safety thresholds reported in prior literature [13]. Each stimulation session includes a 30-s ramp-up at the beginning and a 30-s ramp-down at the end to minimize discomfort and avoid abrupt sensory changes.

In the sham condition, the same electrode montage and device interface will be used. Accordingly, the sham stimulation will consist of a 30-s ramp-up to 1 mA, followed by no current delivery for 29 min, and ending with a 30-s ramp-down. This protocol effectively mimics the sensation of real stimulation without providing any neuromodulatory input, thereby maintaining double-blind conditions and placebo control [14].

Digital upper limb rehabilitation

Digital rehabilitation will be performed using MotoRehab® (Cybermedic Co., Ltd., Republic of Korea), a multimodal, sensor-integrated system designed to simulate daily upper limb functions through interactive task-based training. This system incorporates diverse physical tools, including a touchscreen, rotary handles, air bulbs, and a finger-tapping module. Each tool will be equipped with sensors to detect motion, pressure, grip force, or fine motor control in real time.

Participants will complete tasks from four distinct categories, encompassing a total of nine functional task scenarios. Each category will target different aspects of upper limb motor control and will be linked to specific physical interfaces. The order of categories will be fixed across sessions to ensure consistency, whereas individual tasks within each category will be alternated to maintain participant engagement and reduce habituation effects.

1. Touchscreen-based task.

   These tasks will involve a flat touch-sensitive screen and require visuospatial control, selective attention, and rapid reaching. The following two tasks will be alternated:

   • Whack-a-mole, where participants quickly touch randomly appearing targets.
   • Turning off the lights, where participants touch illuminated circles.

   Task difficulty for both will be adjusted by increasing target appearance speed.

2. Handle rotation tasks

   These tasks will utilize interchangeable rotary handles designed for two grip types, such as cylindrical and pinch grips, to emphasize forearm supination/pronation and grip strength. Participants will alternate between grip types on each trial: odd-numbered interventions (1st, 3rd, 5th, etc.) will use a cylindrical grip, whereas even-numbered interventions (2nd, 4th, 6th, etc.) will use a pinch grip. The following tasks will be alternated within each grip type:

   • Rotating a faucet, where participants rotate a cylindrical handle to simulate turning a faucet on.
   • Tightening a screw, where participants rotate a cylindrical handle to simulate tightening a screw.
   • Starting a car, where participants rotate a handle using a pinch grip to simulate turning a car key.
   • Turning on an oven, where participants rotate a handle using a pinch grip to simulate turning an oven dial.

   Task difficulty will be increased with rotational resistance and the required range of movement.

3. Air bulb squeeze tasks

   These tasks will use a rubber air bulb equipped with pressure and duration sensors to train grasp–release motion and force modulation. The following two tasks will be alternated:

   • Bursting balloons, where participants squeeze the air bulb to simulate popping a balloon.
   • Watering the garden, where participants apply sustained pressure on the air bulb to simulate watering plants.

   Task difficulty will be adjusted by increasing the required grip force.

4. Finger-tapping task

   This task will use a finger keyboard module that will remain fixed across sessions.

   • Playing the piano, where participants tap keys to promote isolated finger control and fine motor timing.

   Task difficulty will be adjusted by increasing the speed of required tapping.

All tasks will incorporate a real-time scoring system and an automated difficulty adjustment algorithm, with each task structured across 20 progressive levels. Participants will begin at Level 1, with task difficulty being dynamically adjusted based on performance: scores below 50 trigger a level decrease, those between 50 and 79 prompt maintenance of the current level, and those 80 or higher trigger a level increase.

Each session will be supervised by trained therapists to ensure safe and accurate execution, reinforce appropriate motor strategies, and support participants’ motivation.

Intervention workflow

Each intervention session will follow a fixed sequence designed to synchronize brain stimulation, motor training, and fNIRS recording. Prior to the 40-min intervention, baseline fNIRS measurements, consisting of a 3-min resting-state recording (eyes closed) followed by two blocks of a finger-tapping task using the virtual piano interface (30-s task, 30-s rest, fixed difficulty), will be conducted.

Subsequently, 10 min of HD-tDCS will be delivered at rest, followed by a 20-min concurrent HD-tDCS and digital upper limb rehabilitation phase. During this period, participants will be asked to perform four task categories (touchscreen, rotary handles, air bulb, finger tapping), each for approximately 5 min. Each category will follow a block design with 50 s of task execution and 10 s of rest. After each block, task performance will be scored and difficulty dynamically adjusted according to the predefined algorithm.

Finally, the session will conclude with a 10–min poststimulation fNIRS recording during which participants will repeat all four task categories without stimulation. During this phase, task difficulty will be kept constant to allow consistent neural comparison across participants and sessions.

Criteria for discontinuing or modifying allocated interventions {11b}

The criteria for discontinuing the intervention are as follows:

• Voluntary discontinuation by the participant: Participants retain the right to withdraw from the trial at any time without the need for an explanation and without influencing the provision of future treatment.

• Discontinuation due to significant adverse events or as requested by the participant or legal representative due to adverse events.

• Significant protocol violation or deviation from inclusion/exclusion criteria during the clinical trial.

• Unsuitability to continue with the clinical trial based on the investigator’s judgment.

• Study adherence < 80%

  $$\text{Adherence}\left(\text{\%}\right)=\left(\frac{\text{Actual number of tDCS sessions performed}}{\text{Total required tDCS sessions}}\right)\times 100$$

Strategies to improve adherence to interventions {11c}

The stimulation protocols for all interventions have already been established as effective for patients with stroke [11]. Participants will be informed that regardless of their assigned study group, several demonstrated benefits of the digital rehabilitation program can be expected [15]. Furthermore, given that all interventions will occur within the hospital, clinical research coordinators will maintain periodic contact through phone calls or text messages to remind participants of their upcoming sessions.

Relevant concomitant care permitted or prohibited {11d}

Any changes in medication or treatment that may significantly affect the efficacy assessment during the study will be prohibited, as determined by the principal investigator or designated study personnel.

Provisions for posttrial care {30}

In accordance with the study’s compensation policy, participants of this trial will be covered by clinical trial insurance maintained by the investigator. This insurance provides coverage for any physical harm caused by the medical device utilized in the investigation, as well as for unexpected adverse events related to the study procedures. However, compensation will not be provided in situations where the device does not produce valid or beneficial outcomes, or in cases where harm results from the participant’s own negligence, among other specified exclusions.

Outcomes {12}

Pre-intervention evaluations will be conducted within 3 days prior to the first intervention session, whereas postintervention evaluations will be conducted 3 days following the final session. The primary outcome will be the postintervention difference in the FMA-UE. Secondary outcome measures will include changes in the 9-Hole Pegboard Test (9HPT), grip and tip pinch strength, the Box and Block Test (BBT), the Short Form-36 (SF-36) health survey, and reaction time and accuracy of the sequential finger-tapping test (SFTT). Additionally, the cortical hemodynamic response will be assessed by analyzing time-series changes in regional cerebral oxygenation measured via fNIRS integrated into the device used for the investigation.

The FMA is a standardized and widely used clinical scale for evaluating motor recovery after stroke, particularly in terms of motor function, balance, sensation, and joint motion [16]. Meanwhile, the FMA-UE subscale assesses voluntary movement, reflex activity, and coordination of the affected upper limb, including the shoulder, elbow, forearm, wrist, and hand. Scores can range from 0 to 66, with higher scores indicating better motor function. This subscale has been widely used in clinical trials and rehabilitation studies to monitor recovery and treatment effects on upper limb function [17]. The 9HPT is a quantitative assessment commonly used in neurological and rehabilitation research to evaluate fine motor coordination and dexterity. During the test, individuals are instructed to place and remove nine pegs into nine holes on a pegboard as quickly as possible [18]. In the current study, completion time on the affected side will be recorded for analysis. Grip and tip pinch strength will be assessed to evaluate hand strength and fine motor control. Grip strength will be measured using the Jamar® Hydraulic Hand Dynamometer (J. A. Preston Corporation, Clifton, NJ), with participants seated upright, feet flat on the floor, and the arms positioned at 90° of elbow flexion with the forearm in a neutral position. Pinch strength will be assessed using a pinch gauge (B&L Engineering, Tustin, CA), including tip, lateral, and three-point pinches to capture different thumb-finger coordination patterns. These measures serve as quantitative indicators of hand function changes induced by the intervention [19]. The BBT is a standardized, performance-based test that evaluates unilateral gross manual dexterity. Participants are instructed to move as many blocks as possible from one compartment to another within 60 s using one hand [20]. The SF-36 is a self-report survey designed to assess health-related quality of life across eight domains, including physical functioning, bodily pain, general health, and mental health. It generates physical and mental component summary scores, which offer a comprehensive view of perceived health status [21]. To evaluate intervention-related changes in individual finger motor function, reaction time and accuracy of SFTT will be assessed using a custom-developed device. Participants can respond to visual cues by tapping the designated fingers on the device.

Cortical hemodynamic response via fNIRS measurement and preprocessing

Cortical hemodynamic response will be evaluated using fNIRS integrated into the investigational HD–tDCS device. This particular system comprises 48 probes that enable both tDCS stimulation and fNIRS measurement. The fNIRS component measures regional cerebral oxygenation by detecting the absorption of near-infrared light transmitted through cortical tissue, consequently enabling the quantification of hemoglobin oxygenation levels across the entire 48-channel array. Analyses will focus on predefined regions of interest (ROIs), specifically the ipsilesional M1 and aIPS while also exploring broader cortical patterns.

fNIRS data were preprocessed following a standardized pipeline. Raw signals will undergo motion artifact detection and correction through spline interpolation and wavelet filtering. A bandpass filter of 0.01–0.2 Hz will be applied to remove high-frequency noise and physiological artifacts, such as cardiac and respiratory oscillations. Short-channel regression will be used to minimize superficial tissue contamination, considering that each channel includes a corresponding short-separation detector. Finally, hemoglobin concentration changes will be computed from the preprocessed light intensity data using the modified Beer–Lambert law. Baseline correction will rely on pre-task rest periods, and task-related signals will be extracted by block averaging of task epochs relative to baseline. For analysis, mean oxyhemoglobin levels will be calculated for each task period, and inter-channel correlations will be computed to derive functional connectivity measures.

Participant timeline {13}

Table 1 will present a comprehensive flowchart of the study process, including the allocation phase. After screening and enrollment, participants will be allocated to either an experimental or control group, followed by a pre-intervention evaluation. The first intervention will take place within 3 days of the pre-intervention evaluation. The intervention will be administered once daily, three times per week, for a total of 10 sessions within 4 weeks. A post-intervention evaluation will be conducted within 3 days after the final intervention session. The total study duration is anticipated to be approximately 4 weeks. The timelines for both the experimental and control groups will be kept identical.

Table 1.

Schedule of enrollment, interventions, and assessments

Timepoint	Study period
	Enrollment	Baseline	Intervention	Postintervention
	-V1	V0	V1–10	V11
Enrollment:
Informed consent	X
Eligibility screen	X
Allocation		X
Intervention:
Real HD-tDCS with Digital UL Rehab			X
Sham HD-tDCS with Digital UL Rehab			X
Assessment:
FMA-UE		X		X
9HPT		X		X
Grip and tip pinch strength		X		X
BBT		X		X
SF-36		X		X
SFTT		X		X
fNIRS monitoring			X
Adverse event measure		X	X	X

HD-tDCS high-definition transcranial direct current stimulation, UL Rehab upper limb rehabilitation, FMA-UE Fugl–Meyer assessment of upper extremity, 9HPT 9-hole pegboard test, BBT Box and Block Test, SF-36 Short Form-36, SFTT reaction time and accuracy of sequential finger-tapping test, fNIRS functional near-infrared spectroscopy

Sample size {14}

The primary outcome of this study will be the postintervention change in FMA-UE score from baseline. A previous randomized controlled pilot study [22] showed that the mean improvement in FMA-UE was 3.30 (SD = 3.27) and 1.20 (SD = 1.87) in the concurrent-tDCS group (n = 10) and sham-tDCS group (n = 10), respectively. Using these values, the expected mean difference (δ) between groups was set at 2.10, whereas the pooled standard deviation (σ) was calculated as 2.664. With a statistical power of 80% and a significance level (α) of 5%, the required sample size per group was estimated using the following formula:

$$N=\frac{2{\sigma }^2\times {(Z_{\frac{\alpha }{2}}+Z_{\beta })}^2}{{\delta }^2}=\frac{2\times {2.664}^2\times {(1.96+0.84)}^2}{{2.10}^2}\approx 26(\text{participants per group})$$

Considering a 20% dropout rate over the 4-week intervention period, the final target sample size was set at 32 participants per group, resulting in a total of 64 participants.

Recruitment {15}

Study participants will be recruited by posting notices on the bulletin boards of the two participating hospitals. Eligibility will not be restricted based on race or socioeconomic status. Patients who satisfy the inclusion criteria will be actively encouraged and supported to participate. Moreover, the patients will receive a clear description of the study’s purpose to promote the inclusion of a representative sample of stroke patients receiving care at each institution.

Assignment of intervention: allocation

Sequence generation {16a}

A site-stratified randomization sequence will be generated using SAS version 9.4 (SAS Institute, Cary, NC, United States). Block randomization will be applied within each site to ensure balance between the intervention and control groups.

Concealment mechanism {16b}

Allocation will be concealed using sealed opaque envelopes prepared for each participant. The envelopes will be labeled with screening identifiers and will only be unsealed to reveal group assignment after enrollment is confirmed, ensuring blinding of the treating personnel prior to allocation disclosure.

Implementation {16c}

Randomization codes will be generated centrally using SAS version 9.4. Investigators at each site will be responsible for enrolling participants and will be assigning them to either the intervention or control group by opening the sealed opaque envelope labeled with the participant’s screening identifier. An unblinded medical device administrator will be responsible for preparing the tDCS device settings according to group assignment prior to the start of each intervention session, ensuring that treating therapists and participants remain blinded throughout the trial.

Assignment of intervention: blinding

Who will be blinded {17a}

Both participants and assessors will remain blinded to the treatment allocation throughout the trial, ensuring a double-blind study. Identical tDCS devices will be used across both groups, with only the stimulation settings differing according to allocation, ensuring visual indistinguishability for blinding purposes. Assessments will be conducted in a separate space by evaluators not involved in administering the intervention, ensuring maintenance of blinding. Individual device assignment codes for each participant will be securely stored by the medical device manager and can only be unblinded with authorization from the sponsor or principal investigator.

Procedure for unblinding if needed {17b}

In the event of a serious medical emergency, unblinding will be performed only if knowledge of the stimulation protocol is deemed essential for appropriate clinical management. Participants who are unblinded due to safety concerns will discontinue participation in the study.

Data collection and management

Plans for assessment and collection of outcomes {18a}

This multicenter study will be conducted at two hospitals within the Republic of Korea. To enhance the reliability and validity of the study outcomes, investigators from both sites will be holding regular meetings to establish standardized procedures and assessment protocols. Detailed evaluation guidelines will be documented and shared across institutions to ensure consistency in data collection. Most of the assessment tools used in this study will be validated for both reliability and validity. These efforts aim to minimize inter-rater variability and maintain methodological rigor throughout the study.

Plans to promote participant retention and complete follow—up {18b}

To ensure participant compliance and retention, clinical research coordinators at each site will contact participants via telephone or text messaging to remind them of scheduled evaluations and intervention sessions.

Data management {19}

A contract research organization (CRO) will be contracted to manage key data management aspects for this study. The CRO will be responsible for developing the case report form (CRF) and implementing standardized data management procedures. All collected data will be double-entered into a centralized database, followed by consistency checks and range validations to ensure data accuracy. The data management team will code adverse events using standardized medical terminology and reconcile serious adverse events with the corresponding safety data. A predefined data validation plan will be established to guide all data handling procedures throughout the study. All data will be securely stored and managed in accordance with Good Clinical Practice guidelines.

Confidentiality {27}

In accordance with ethical regulations, participants’ personal information and study-related outcomes will be recorded on designated paper CRFs, excluding all identifiable information such as medical record numbers and participant names. To maintain confidentiality, access to these records will be strictly restricted to authorized study personnel. The identity of each participant will be protected at all stages of data handling, including analysis, presentation, and publication. All research materials, including imaging data and associated documents, will be securely stored in password-protected digital files or locked physical facilities. As mandated by the Bioethics and Safety Act of the Republic of Korea, clinical trial documents and informed consent forms are retained for a minimum of 3 years following study completion. Thereafter, any remaining records are disposed of in accordance with the Personal Information Protection Act.

Plans for collection, laboratory evaluation, and storage of biological specimens for genetic or molecular analysis in this trial/future use {33}

Not applicable. No biological specimens will be collected.

Statistical methods

Statistical methods for primary and secondary outcomes {20a}

The primary outcome will be the postintervention change in the FMA-UE score from baseline. Between-group comparisons will be conducted using analysis of covariance, with baseline FMA-UE scores being included as a covariate. Least square means, 95% confidence intervals, and p values will be reported. An adjusted mean difference exceeding 2.10 and a lower limit of the 95% confidence interval greater than 0.00 indicate a clinically superior intervention.

Secondary outcomes will include the 9-Hole Peg Test, grip and tip pinch strength, finger-tapping reaction time and accuracy, the BBT, the SF-36 Health Survey, and cortical activation. Cortical hemodynamic responses will be defined as the change in oxyhemoglobin concentration within the ipsilesional M1 and aIPS during motor task execution, relative to baseline, following HD-tDCS. These outcomes will be analyzed by comparing postintervention changes in these variables between the groups. Data that meet normality assumptions will be assessed using independent t-tests; otherwise, the Mann–Whitney U test will be applied.

Interim analyses {21b}

No interim analyses will be conducted given the relatively short study duration and predefined sample size.

Methods for additional analyses {20b}

Additional analyses will include longitudinal modeling of intervention effects across sessions to characterize the cumulative and time-dependent impact of HD-tDCS on cortical hemodynamic response. Mixed-effects models will be used to assess within-subject changes over time, accounting for repeated measures and individual variability.

Furthermore, exploratory analyses will be performed to develop a personalized intervention algorithm based on cortical activation and connectivity patterns measured via fNIRS. Specifically, unsupervised clustering and regression models will be employed to identify cortical activation and connectivity profiles associated with motor recovery, with the aim of informing the development of adaptive stimulation strategies for future individualized protocols.

Methods in analysis to handle protocol nonadherence and any statistical methods to handle missing data {20c}

Efficacy analyses will be conducted on the full analysis set (FAS), which includes all randomized participants who received at least one intervention and completed at least one postbaseline primary outcome assessment. A per-protocol set, comprising participants who adhered to the study protocol without major violations, will also be analyzed for sensitivity purposes.

Major protocol deviations include violation of the inclusion or exclusion criteria, use of concomitant interventions that may affect outcome assessments, and poor adherence to the intervention protocol (e.g., missing more than three consecutive sessions or completing less than 80% of the total number of sessions).

Missing data for the primary outcome will be handled using the last observation carried forward (LOCF) method in the FAS. Missing data for secondary outcomes will be addressed using available data without imputation. Sensitivity analyses will be performed to assess the impact of missing data on the robustness of the results.

Plans to give access to the full protocol, participant-level dataset, and statistical code {31c}

The protocol, de-identified participant-level dataset, and statistical code will be available from the corresponding author upon reasonable request and subject to ethical approval.

Oversight and monitoring

Composition of the coordinating centre and trial steering committee {5d}

The trial will be coordinated by the principal investigator and core research team with operational support provided by a CRO. The CRO will be responsible for monitoring the trial process, ensuring regulatory compliance, and supporting data integrity. No separate trial steering committee will be established. Oversight will be provided through regular study team meetings in collaboration with the CRO.

Composition of the data monitoring committee, its role and reporting structure {21a}

No independent data monitoring committee will be formed given that the intervention will be noninvasive and have minimal risk. Safety data and adverse events will be reviewed periodically by the principal investigator and research team during the study.

Adverse event reporting and harms {22}

During the clinical trial, study personnel will document details regarding adverse reactions, including symptoms, onset dates, and resolution dates, using a standardized adverse event record form. The severity of each reaction will then be rated on a scale from negligible to critical, whereas causality will be classified as obvious, probable, suspected, low, none, or indeterminable. Medical device-related interventions will be categorized into discontinuation, dose reduction, dose increase, no change, unknown, or not applicable. The outcomes of these interventions, such as symptom resolution or worsening, will also be systematically tracked.

When reporting study results, the principal investigators (PIs) will provide a detailed description and evaluation of all symptoms observed during the trial. In the event of a serious adverse event, the PIs will notify the IRB to assess whether the study should proceed or be terminated. Critical events, such as death or life-threatening conditions, will be reported to the Ministry of Food and Drug Safety (MFDS) of the Republic of Korea within 7 days. Events involving hospitalization, prolonged hospitalization, irreversible damage, severe disability, or functional impairment will be reported within 15 days. All intervention outcomes will be recorded according to the following predefined categories: resolved, resolving, unresolved, resolved with sequelae, death, or unknown.

Frequency and plans for auditing trial conduct {23}

No pre-established schedules will be established for auditing the clinical trial. Nevertheless, audits may be carried out at any time by the MFDS of the Republic of Korea or by an internal auditing body within the institution conducting the study. These audits will be conducted independently of both the investigators and the sponsor.

Plans for communicating important protocol amendments to relevant parties (e.g., trial participants, ethical committees) {25}

In the event of significant protocol amendments, the PIs at each participating institution will engage in mutual consultations to reach a consensus before implementation. They will be responsible for subsequently reporting these amendments to the MFDS of the Republic of Korea and their respective IRBs. Furthermore, they are expected to communicate the changes to their research teams in detail and, when necessary, inform the research participants.

Dissemination plans {31a}

The trial has been registered at CRIS (KCT0010556). We will continuously update the trial status on the site throughout the study. The academic paper will be scheduled for publication within 2 years after completing data collection.

Discussion

This protocol outlines a rigorously designed, multicenter, double-blind, sham-controlled trial that aims to confirm the efficacy and safety of HD-tDCS combined with structured digital upper limb rehabilitation among stroke survivors. This trial builds on promising exploratory data showing that dual-site HD-tDCS targeting the ipsilesional M1 and aIPS, in combination with task-specific upper limb training, can be a safe and well-tolerated approach for improving upper limb function [11]. By integrating fNIRS, this study provides an opportunity to link stimulation-induced cortical activation patterns with motor recovery in real time, thereby addressing a key limitation of prior trials focusing primarily on behavioral endpoints [23, 24].

A major strength of this study is its robust methodology, which involves site-stratified randomization, standardized intervention protocols, and the use of a validated digital rehabilitation system that combines task variability and automated difficulty adjustment [15]. Integrating HD-tDCS with a 48-channel fNIRS array allows for precise monitoring of neuroplastic changes in predefined regions, such as the M1 and aIPS, while exploring broader cortical patterns. This multimodal approach may guide the future development of personalized neuromodulation strategies, an emerging priority in poststroke rehabilitation [25, 26].

Despite these strengths, some inherent limitations of the study need to be noted. The short-term follow-up limits the establishment of definitive conclusions regarding the durability of intervention effects, underscoring the need for future longitudinal studies. Furthermore, the use of a specific HD-tDCS–fNIRS system, which has yet to be widely adopted outside research settings, may limit the generalizability of the findings. Nonetheless, the trial is positioned to provide high-quality evidence addressing critical gaps in stroke neurorehabilitation research.

In summary, this confirmatory trial seeks to not only validate a promising multimodal intervention but also deepen our understanding of its neurophysiological mechanisms. The findings from this study may help pave the way for precision neurorehabilitation strategies that tailor stimulation protocols based on individual cortical response patterns.

Trial status

Protocol version: Version 3.3 (04 APR 2025).

Study start: 20 SEP 2025 (actual).

Study completion: 31 DEC 2025 (estimated).

Abbreviations

aIPS

Anterior intraparietal sulcus

CRF

Case report form

CRO

Contract research organization

FAS

Full analysis set

FMA-UE

Fugl-Meyer Assessment of Upper Extremity

fNIRS

Functional near-infrared spectroscopy

IRB

Institutional Review Board

HD-tDCS

High-definition transcranial direct current stimulation

LOCF

Last observation carried forward

M1

Primary motor cortex

MFDS

Ministry of Food and Drug Safety

PI

Principal investigators

ROI

Regions of interest

WSO

World Stroke Organization

Authors’ contributions {31b}

JK and Su-HL contributed to the development and organization of the protocol and wrote the first draft of the manuscript. EMK, SRK, YJJ, and Se-HL conceptualized and organized the protocol. HCJ, MHK, and YHK conceptualized, developed, and organized the protocol. MHK and YHK, as the corresponding authors, also conceived the study. All authors read and approved the final manuscript.

Funding {4}

This work was supported by the Korea Medical Device Development Fund grant funded by the Korean government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health and Welfare, the MFDS) (KMDF‑RS‑2022–00140478).

Data availability {29}

The datasets generated and analyzed during the current study will be available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate {24}

The study protocol was approved by the IRB of the Jeonbuk National University Hospital and Myongji Choonhey Rehabilitation Hospital (IRB No. CUHIRB 2024–03-012 and MJCHIRB–2024–01, respectively). Written informed consent to participate will be obtained from all participants.

Consent for publication {32}

Not applicable. No identifying images or other personal or clinical details of participants will be presented herein or will be presented in reports of the trial results. Participant information materials and informed consent form will be available from the corresponding authors on request.

Competing interests {28}

Ho Choon Jeong is the President of CyberMedic Co., Ltd., the sponsor of this study. The remaining authors declare that they have no competing interests.