Endovascular Brain-Computer Interface for Upper-Limb Paralysis

Participants with post-stroke hemiplegia who still have rehabilitation potential. After endovascular implantation of the venous BCI Type-A sensor, they receive decoder calibration during week 1, followed by daily 45-minute closed-loop functional electrical stimulation (FES) training during weeks 2 to 4, in which imagined or attempted hand movements decoded from brain signals trigger electrical stimulation of the paralyzed forearm muscles with real-time visual feedback. Stimulation intensity is adjusted dynamically, with a mandatory 2-minute rest every 10 minutes.

A self-expanding nitinol stent (diameter 6 mm) carrying pure platinum electrodes, implanted under general anesthesia via the subclavian vein into the superior sagittal sinus or a cortical vein overlying the hand area of the motor cortex, using fusion of 3-dimensional venography with MRI for targeting. The sensor is connected to a chest-wall implanted relay and telemetry unit for wireless signal transmission.

Closed-loop functional electrical stimulation of the paralyzed forearm muscles, triggered in real time by decoded movement intention. Daily 45-minute sessions during post-implantation weeks 2 to 4.

Participants with ALS or severe paraplegia with no realistic prospect of recovering voluntary movement. After endovascular implantation of the venous BCI Type-A sensor, they complete decoder calibration during week 1, followed by daily 45-minute robotic arm training during weeks 2 to 4, in which decoded movement intention is used to control a bionic robotic arm for simulated daily tasks such as eating, grasping objects, and opening a door.

A self-expanding nitinol stent (diameter 6 mm) carrying pure platinum electrodes, implanted under general anesthesia via the subclavian vein into the superior sagittal sinus or a cortical vein overlying the hand area of the motor cortex, using fusion of 3-dimensional venography with MRI for targeting. The sensor is connected to a chest-wall implanted relay and telemetry unit for wireless signal transmission.

A bionic robotic arm controlled by decoded movement intention, used to perform simulated daily tasks. Daily 45-minute training sessions during post-implantation weeks 2 to 4. Response latency, task success rate, and false trigger rate are recorded.

Number and percentage of participants with treatment-emergent adverse events, including device- or procedure-related events such as cerebral venous sinus thrombosis, cerebral hemorrhage or infarction, device migration, infection, and epileptiform discharges. Events are coded with MedDRA version 24.1.

Number and percentage of participants in whom the endovascular venous BCI Type-A sensor is successfully deployed in the target venous sinus and remains functional (wireless signal transmission) through post-implantation day 60.

Effective bandwidth of neural signals recorded by the endovascular venous BCI Type-A sensor, in hertz (Hz), assessed by power spectral analysis of recorded signals. Higher values indicate better signal fidelity.

Percentage of successfully triggered system commands in 10 randomly selected motor-imagery tasks per assessment session (protocol-defined target >=90%). Higher percentages indicate better decoding performance.

Change in activation volume of upper-limb motor areas during motor tasks on functional MRI or magnetoencephalography (activation volume in cubic centimeters; greater volume indicates greater activation; protocol target >=15% increase).

Change in Barthel Index score from baseline, a measure of activities of daily living (range 0-100; higher scores indicate greater independence).

Change in Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) scores from baseline. (Scale and summary scores range 0-100; higher scores indicate better health-related quality of life)

Relative band power of delta, theta, alpha, and beta bands; relative alpha variability; and cerebral state index derived from quantitative electroencephalography. Each index is reported separately.

Glasgow Outcome Scale score at the end of follow-up (range 1-5; higher scores indicate better neurological outcome; GOS 1-3 unfavorable, GOS 4-5 favorable).

Change in integral surface electromyography value (in microvolt-seconds) of the affected flexor carpi radialis, reflecting muscle recruitment (protocol target >=50% increase; higher values indicate greater muscle recruitment).

Change in National Institutes of Health Stroke Scale total score from baseline (range 0-42; higher scores indicate more severe neurological impairment).

Percentage of triggered commands classified as false triggers (lower percentages indicate better decoding specificity).

Electrode impedance of the endovascular sensor in ohms, measured monthly. Lower impedance indicates better electrode-tissue contact.

Percentage of recording data lost or discarded per session. Lower percentages indicate more reliable data acquisition.

Consistency of signal amplitude across electrode channels, expressed as an inter-channel correlation coefficient (range 0-1; higher values indicate greater consistency).

Change from baseline, assessed by manual muscle testing (Medical Research Council scale, 0-5; higher grades indicate greater strength). Protocol target: improvement of at least 1 grade.

Change from baseline (0-34 points as specified in the protocol; higher scores indicate better upper-limb motor function). Protocol target: increase of at least 40%.

Change from baseline under the assisted condition, in seconds (no fixed range; shorter times indicate better performance). Protocol target: reduction of at least 10%.

Percentage of key daily tasks (eating, opening a door, grasping objects) successfully completed using the robotic arm (0-100%; higher percentages indicate better task performance). Protocol target: at least 80%.

Percentage of planned training sessions completed, recorded in training logs.

Change from baseline. The HADS is a 14-item self-report questionnaire with anxiety and depression subscales, each scored 0 to 21 (total 0-42); higher scores indicate worse anxiety and depressive symptoms.

Satisfaction with the assistive technology system assessed by the QUEST 2.0, a 12-item questionnaire scored on a 5-point satisfaction scale (1 = not satisfied at all to 5 = very satisfied); higher scores indicate greater satisfaction.