• spectrogram
• EEG
• ECoG
• SCI
• healthy
• locomotion
• EES
• motor control
• BSI
• kinematics
• EMG
• TDM
• gait events

Spinal Cord Injury (SCI) has a devastating impact on the motor control and affects many essential physiological functions such as locomotion function. A partial or total incapability of safe upright standing and walking capability causes an important loss of independency and autonomy in humans daily life. In latest years, new technologies have emerged giving hope for rehabilitation therapies that aim to alleviate paralysis using neuroprosthetic systems.

This Master thesis project is part of a research study, it aims to establish a digital bridge between the motor cortex and the dormant neurons located in the spinal cord that are deprived of brain inputs due to a spinal cord injury (SCI). This digital bridge, also called a brain-spine interface (BSI), aims at restoring voluntary mobility and promote neurological recovery in people who are chronically paralyzed. In this contest, I focus my efforts on extraction of best cortical features in healthy and SCI participants during different walking condition, in order to demonstrate all reasons because the signals’ acquisition through a minimal invasive recording device, EpiECoG electrodes could be an appropriate trade-off between invasiveness and quality of signals and the best choice to decode gait cycle events.

As a first step, we investigated cortical dynamics reconstructed from high-density EEG signals (128 electrodes) during a walking task in seven healthy volunteers. In order to track and study gait, we recorded electromyography (EMG) activity from seven leg muscles in each leg, and used a motion capture system to acquire three-dimensional kinematics during walking. We identified locomotor gait events (foot strike and foot off) by analysing EMG signals and kinematic parameters. In accordance with previous literature, we found low gamma (24-40 Hz) amplitude modulations related to the gait cycle, mainly in sensorimotor electrodes.

Then we perfume same analysis on four SCI patients during different locomotion conditions (treadmill and overground walking, with EES ON, with EES OFF, with and without intention) and, also tanks to accurate reconstruction of source localization, we get interesting results about similar cortical dynamics in healthy and SCI participants but also some different behaviour.

At the end, we analyse also one SCI subject with an epicortical implantation thank to WIMAGINE device. The results based on minimally-invasive fully implantable epidural ECoG-based system called are clearer and the muscular artifacts are removed. It let us to highlight also a conversely correlation between high and low gamma amplitudes. Furthermore, a perfect synchronization between ERS/ERD and gait cycle phase in low gamma amplitudes. These two evidences could reveal the presence of two different pathways, sensory feedback information and cortical dynamics relate to motor action itself.

So, while EEG is a non-invasive technique, it does not allow stable signals in daily life conditions due to the low signal-to-noise ratio and the presence of exogenous signals that contaminate brain signals. We investigated whether semi-invasive recordings of electrocorticographic (ECoG) signals enable more robust decoding of leg movements and can be integrated with spinal cord stimulation for daily use as a BSI in patients with severe SCI.

At the end, to bypass the spinal cord lesion, we prefer decode motor intention from ECoG recordings, and translate the decoded events into commands to trigger EES, which will activate lower limb muscles following the natural flow of an intended movement.