A BRAIN–SPINE INTERFACE TO REVERSE PARALYSIS AFTER SPINAL CORD INJURY – THINK2MOVE

The goal of this collaborative project between CEA Clinatec and the CHUV/EPFL teams is to make the use of neuroprostheses for restoring walking more accessible. We therefore built on the approved STIMO-BSI clinical trial to implement new concepts aimed at increasing the degree of voluntary control enabled by the BSI system and to explore sensorimotor neural coding during learning phases. In parallel, we developed a portable BSI system allowing participants to use it outside a clinical setting, where they were able to evaluate the opportunities and constraints of using this system in their daily lives.

To achieve these objectives, developments focused on:

- Developing a stimulation exploration procedure optimized by machine learning methods. This enables optimal selection of epidural electrical stimulation parameters (stimulation sites, intensity, duration, etc.) for each type of intended movement.

- Improving the performance of movement intention decoding through new algorithmic approaches. In addition, the ability of decoders to learn from patients’ improved control of the neuroprosthesis was studied. This included a better understanding of the mechanisms of learning and adaptation to task complexity.

- Enhancing the portability of the system through the optimization of technological components (hardware and software), in order to make the system as lightweight and user-friendly as possible, thus enabling use at home outside of a clinical environment.

The THINK2MOVE project demonstrated, in a clinical trial, the effectiveness of artificial intelligence–based algorithms to optimize spinal cord stimulation parameters. The approaches developed made it possible to obtain stimulation settings equivalent to (or even surpassing in some cases) those chosen by an expert, and in a significantly shorter time (only 1h30 in the worst case).

Decoding performance was also improved, enabling both more responsive detection when the patient attempts to initiate a movement, and more robust control (fewer unintended stimulations).

Finally, the system was optimized to be easier to use and fully portable, fitting into a backpack or being integrated onto a walker. The portable system is now used by patients at home, outside the controlled clinical environment.

These results have led in particular to a major publication in Nature (Lorach et al., Nature 2023), as well as the filing of several patents.

The results obtained within the Think2Move project are very promising for the future of brain–spine interfaces and will serve as a foundation for the continuation of clinical trials based on epidural stimulation controlled by a brain–machine interface system. These trials will aim to restore movements not only in paraplegic patients but also in tetraplegic patients, for the control of upper limbs. To date, two clinical trials have already begun, with the implantation of two paraplegic patients (Think2Go clinical trial) and two tetraplegic patients (UP2 clinical trial).

A technology transfer agreement for the developments achieved within the Think2Move project has been signed between CEA and the company ONWARD Medical. This agreement will enable the transition to a larger-scale, multi-center clinical trial and, ultimately, the availability of a neuroprosthesis that can be used by a greater number of patients.

Problem: Nearly 500.000 people sustain a spinal cord injury (SCI) every year, with dramatic human, societal and economical cost. In the most severe cases, the SCI leads to complete paralysis of both legs, which binds patients to a wheelchair for the rest of their life, with no treatment perspective.

Current solution: Recent advances in the field of neuroprosthetics opened prospects for solutions to alleviate some of the motor deficits associated with SCI, and thus increase the autonomy of affected patients. For example, we recently showed that the application of epidural electrical stimulation over the lumbosacral spinal cord restored standing, walking, cycling and swimming in patients with clinically incomplete and clinically complete SCI. This recovery required real-time control of the stimulation, since the individual dorsal roots of the spinal cord must be targeted with a spatial and temporal sequence that coincides with the intended movements.

Current barriers: The detection of the intended movements is critical to synchronize the stimulation with the intended movements. We have used motion sensors to detect intended movements from movements of body parts that have not been completely paralyzed by the SCI. This approach is sufficient to detect ongoing movements, but it fails to anticipate task transitions or volitional adjustment of movement—as necessary to support mobility in everyday life. Complex motor control cannot be achieved with external sensors only and would trigger unnatural compensatory movements.

Hypothesis and approach: We previously developed a brain-spine interface (BSI) that links motor intentions decoded from recordings of motor cortex activity to the delivery of epidural electrical stimulation targeting the dorsal roots of lumbar segments. This BSI immediately restored voluntary control of walking in a non-human primate model of leg paralysis, and augmented neuroplasticity and functional recovery when combined with gait rehabilitation. Our aim is to translate this BSI concept into a treatment to overcome paralysis and augment recovery in humans with SCI. This BSI is contingent on a system capable of decoding motor intention from brain recordings. We have developed WIMAGINE, the first fully-implantable system that enables safe long-term wireless recording of motor cortex activity with high reliability in humans. WIMAGINE allowed two individuals with complete paralysis to control computer programs and a full-body exoskeleton with their motor intent.
We have interfaced WIMAGINE to an implantable neurostimulator for real-time control of epidural electrical stimulation, establishing the first fully implantable BSI for humans. We obtained approval from the Swiss Ethics committee (CER-VD2020- 01814) to implant this fully implantable BSI in three patients with chronic paralysis. In this study, we will explore the safety preliminary efficacy of this BSI to restore voluntary movement and mobility, and to increase neurological recovery in response to rehabilitation.
In this joint proposal between the CEA/Clinatec and the CHUV/EPFL teams, we propose to build on the approved proof-of-concept STIMO-BSI trial and implement new concepts to increase the degree of voluntary control allowed by the BSI and explore the exquisite details of sensorimotor neural coding. In parallel, we will develop a user-friendly portable system that will allow the participants to operate the BSI outside hospital settings. This portable BSI will support a transition to home use where the participants will test the opportunities and constraints of integrating the BSI into their everyday lives.

Expected results and impact: This project represents the unique opportunity to gather highly synergistic teams of engineers, scientists and clinicians in France and Switzerland to combine cutting-edge neurotechnologies into an integrated system that will restore voluntary control of leg movements and mobility in patients with chronic paralysis.