Current evidence and clinical applications of robotic gait training devices for motor function recovery post-stroke are increasingly available. Although existing research demonstrates that robotic gait training can improve patients' gait and balance, there remains a lack of in-depth investigation into its specific mechanisms of action concerning central nervous system (CNS) reorganization - notably, changes in activity within the motor cortex and associated neural networks. The intrinsic changes within the CNS have received insufficient attention, limiting a comprehensive and profound understanding of the rehabilitation outcomes. Therefore, this study aims to elucidate the potential mechanisms underlying robotic gait training-induced neuroplasticity by integrating functional near-infrared spectroscopy (fNIRS) technology with multi-dimensional lower limb motor function assessment tools (such as FAC, BBS, AMEDA, 10MWT, 6MWT, TUGT). It will systematically investigate the effects of robotic gait training on both the central nervous system and lower limb motor function in stroke patients. Furthermore, the study will compare the differences in functional recovery efficacy between robotic gait training and conventional rehabilitation therapies.

In this study, participants will be randomly allocated into two groups: the Welwalk training group and the conventional rehabilitation therapy group.

Welwalk Training Group: Each session will consist of 30 minutes of Welwalk robot-assisted training, followed by 15 minutes of gait training and 15 minutes of supplementary exercises.Control Group (Conventional Rehabilitation Therapy): Each session will consist of 45 minutes of gait training and 15 minutes of supplementary exercises.The intervention period will span 3 weeks, with sessions administered six times per week. Each session will last 1 hour.

Clinical assessments will be conducted by certified healthcare professionals at four time points: at baseline (prior to the commencement of formal training), and after the 1st week, 2nd week, and 3rd week of treatment.

Inclusion Criteria:

• Patients or family gave written informed consents to participate in this study.
• Patients with first hemiplegia caused by primary supratentorial intracerebral hemorrhage or cerebral infarction.
• Within 1 year of stroke onset
• Aged ≥ 20
• Body weight is between 40 and 80 kg
• No excessive spasticity in hip, knee, and ankle joints (Modified Ashworth Scale <3)
• sufficient cognition to follow simple instructions and to understand the content and purpose of the study (Chinese version-MOCA ≥ 20 points)
• Patients who have risks of giving-way when they walk with Ankle-Foot orthosis (AFO)

Exclusion Criteria:

• A history of myocardial infarction
• Muscular or neurological disorder including diabetic neuropathy
• Symptomatic angina or arrhythmia
• Symptomatic respiratory disorder
• Communicable infection
• Joint contracture or limb deformity that affects walking (Range of motion of hip extension < 5 degree, knee extension < -5 degree (can be flexible), ankle dorsiflexion with knee extension position < 5 degree)
• Heterotropic ossification that restrict the range of motion of joints of lower extremities
• Being vulnerable to fracture like severe osteoporosis of spine or lower extremities
• Incontinence of urine or feces that may deface the robotic knee-ankle-foot device of Welwalk
• Inadequate control of hypertension (resting systolic blood pressure ≥ 180 mmHg or diastolic blood pressure ≥ 120 mmHg)
• Inadequate control of tachycardia (heart rate at rest ≥ 120 bpm)
• Training restriction due to reduced cardiac function or respiratory dysfunction
• Visual or auditory impairment hindering training
• Pregnant patients
• Recent participation in other clinical trials
• Patient whom examination doctor judge improper as a trial subject
• Anyone not able to sustain the training protocol with Welwork or regular training

We will share the demographic information and baseline clinical data of all participants.

Starting 6 months after publication

Authorized professional researchers, including but not limited to researchers engaged in neuroscience research who have obtained data access permission from their affiliated institutions, and clinical doctors from other medical institutions that have a cooperative relationship with this study and have signed data confidentiality agreements.

They can access the detailed clinical medical histories of the participants, including past disease histories and treatment process records; neurological function assessment scale data; as well as imaging data collected during the study, such as brain magnetic resonance imaging (MRI) results. However, sensitive information related to participants' privacy, such as names, ID numbers, and contact information, will be strictly anonymized to ensure that such information cannot be obtained.

They can contact the corresponding author or the first author via email.