Spinal cord injury (SCI) often results in persistent motor deficits that are inadequately addressed by conventional rehabilitation. Transcutaneous spinal cord stimulation (tSCS) is a promising non-invasive neuromodulatory approach that can enhance motor activation and gait performance; however, current tSCS systems rely on static, pre-programmed stimulation parameters that do not adapt to real-time motor output or task demands. This limitation may reduce muscle selectivity and disrupt the spatiotemporal dynamics of spinal network activation required for functional movement.

This study aims to develop and evaluate an AI-powered closed-loop multielectrode tSCS system that dynamically adjusts stimulation parameters in real time based on kinematic and surface electromyography (EMG) feedback during walking in individuals with incomplete SCI. The study will compare immediate muscle recruitment and motor performance between conventional static tSCS and dynamic, targeted tSCS guided by real-time physiological signals.

The investigators hypothesize that AI-driven closed-loop tSCS will be safe and feasible, and will result in superior muscle activation patterns and improved gait performance compared with static stimulation. Findings from this study will provide foundational evidence for adaptive neuromodulation strategies and support the advancement of next-generation, data-driven spinal cord stimulation technologies for neurorehabilitation in SCI.

Spinal Cord Injury (SCI) is a debilitating condition affecting approximately 20.6 million individuals worldwide, with an annual incidence of 0.9 million. In Singapore, the prevalence of SCI increased from 5 to 6 per 100,000 in 1990 to 13 to 15 per 100,000 in 2019. SCI often results in profound and long-term neurological impairments, particularly in motor function, leading to significant limitations in activities of daily living (ADLs) and ambulation. The costs associated with SCI acute treatment, rehabilitation, and long-term care are substantial. While conventional rehabilitation strategies remain essential, their effectiveness in restoring lost motor function is limited, leaving many individuals with permanent disability.

Spinal cord stimulation (SCS) has emerged as a promising neuromodulation technique to improve neurological recovery following SCI. SCS delivers electrical impulses to activate afferent fibers, enhancing interneuronal connections, motor neuron excitability, and communication between spinal networks SCS can be broadly classified into epidural SCS (eSCS) and transcutaneous SCS (tSCS). Owning to its non-invasive nature and therapeutic potential, tSCS has gained significant attention, with numerous studies demonstrating its effectiveness in improving motor function following SCI. Most current tSCS protocols for lower limb motor control involve placing one or two active electrodes over the thoracolumbar spine (T10 - L2), with fixed stimulation sites and parameters throughout each session. This approach is grounded in the hypothesis that tSCS enhances overall excitability of the neural network by increasing sensory input. However, the lack of muscle-specific stimulation may lead to unwanted co-contraction of antagonistic muscles, hindering functional movement and reducing overall gait efficiency.

Emerging evidence indicates that spinal cord excitability responds dynamically to variations in stimulation sites and parameters. Spatially selective eSCS with real-time processing has been shown to rapidly restore voluntary motor control even in individuals with chronic, motor-complete SCI. In animal studies, integrated approaches combining epidural spinal cord stimulation with peripheral muscle stimulation designed to mimic sensory feedback and feedforward muscle contraction loops demonstrated synergistic effects, providing a framework for the development of neuromodulation systems to enhance motor recovery following SC. In human studies, another study team has reported that multielectrode tSCS with continuous stimulation to engage central pattern generator (CPG) networks in combination with spatiotemporal alternating stimulation targeting dorsal roots projecting to the leg flexor and extensor motor pools, can induce alternating locomotor activity. Remarkably, this approach enabled immediate recovery of locomotor function in individuals with severe lower limb motor deficits even in clinically complete SCI.

To date, the relative efficacy of combing continuous stimulation with spatiotemporal modulation, compared with continuous stimulation alone, has not yet been systematically evaluated in humans. This study aims to address these gaps by developing and evaluating an AI-powered closed-loop multielectrode tSCS system that integrates continuous midline stimulation with real-time, feedback-driven spatiotemporal modulation. The system leverages wearable kinematic sensors and surface electromyography (EMG) to dynamically adjust stimulation timing and intensity based on ongoing gait and muscle activation patterns. By aligning stimulation delivery with physiological motor demands, the proposed approach seeks to enhance muscle selectivity, optimize lower limb motor recruitment, and improve gait performance in individuals with incomplete SCI.

If successful, this study will provide critical evidence supporting adaptive, AI-driven neuromodulation strategies and establish a foundation for next-generation tSCS systems that more effectively engage spinal sensorimotor circuits to promote functional recovery after SCI.

Inclusion Criteria:

Healthy volunteers:

1. Adults aged 21 years or older with the mental capacity to provide informed consent.
2. No prior diagnosis of neuromuscular or neurological conditions affecting the lower limbs.
3. Able to walk independently with a normal gait pattern, as confirmed by clinical observation done by the study team.

SCI patients:

1. Age 21 -75 years;
2. Non-progressive, incomplete SCI (traumatic or non-traumatic).
3. International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) /ASIA Impairment scale (AIS) grade B, C, or D, with Lower Extremity Motor Scores (ISNCSCI-LEMS) less than or equal to 40;
4. Injury level at/above L1 (above conus medullaris) with intact segmental reflexes below level of lesion;
5. Able to provide informed consent;
6. No contraindication for tSCS, such as pace makers or other implantable electrical devices;
7. Eligible for body-weight support treadmill training;
8. Able to stand with body weight support and/or assistance.

Exclusion Criteria:

Healthy voluteers:

1. Presence of significant or unstable medical comorbidities, including uncontrolled cardiopulmonary disease or severe cognitive impairment, as determined by the study team.
2. Diagnosis of any neuromuscular or musculoskeletal disorders (e.g., congenital skeletal deformity, limb amputation, neurological disorders, myopathy).
3. Abnormal gait pattern due to underlying medical co-morbidities, as identified through physical examination and clinical assessment, and confirmed by the study team.
4. Current pregnancy.

SCI patients:

1. Significant or unstable medical co-morbidities, including uncontrolled cardiopulmonary disease, severe cognitive impairment, or severe dysautonomia, as determined by the study team;
2. Uncontrolled neuropathic or musculoskeletal pain, or contractures affecting participation in therapy;
3. Known history of peripheral nerve injury (e.g., traumatic nerve injury, entrapment neuropathy);
4. Pregnancy;
5. Active malignancy or ongoing cancer treatment;
6. Skin conditions (e.g., ulcers, infections, malignant lesions) that limit the application of tSCS electrodes;
7. Prior exposure to tSCS or eSCS interventions.