Exploring Motor Learning in Acute Stroke Through Robotics

Over 3 consecutive days, the subjects will be evaluated and will train on the rehabilitation robot REAplan® (http://www.axinesis.com/). They will practice 2 serious games on the robot.

In order to differentiate the effect of motor control recovery from that of MskL, the acute stroke patients will be randomised 4/1 to "MskL" (n=120) ,to "motor control recovery" (n=30) or to "conventional" group (n=15) with a minimisation software. The experimental design will be similar except that the "motor control recovery" group will practice the serious game EASY instead of CIRCUIT (see below) and the "conventional" group will practice conventional therapy instead of CIRCUIT. The performances on both EASY \& CIRCUIT will be compared between groups (subjects in both groups will perform the EASY \& CIRCUIT tasks).The total time of rehabilitation will be the same.

The motor skill learning setup (CIRCUIT + EASY) that we developed and successfully used in healthy individuals and stroke patients has already been implemented in the REAplan environment and will be used as innovative serious games based on a speed/accuracy trade-off (SAT), allowing a detailed analysis of motor skill learning components (speed, accuracy, SAT, movement smoothness, dynamics...). For the serious game CIRCUIT, who based on motor skill learning, the subjects will have to practice a complex circuit and move as a cursor quickly and accurately as possible by controlling the handle of the robot with their affected hand/arm. For the other task EASY (a brick busters serious game), the aim will be to go back and forth between walls presented in different locations. The CONVENTIONAL therapy will consist in classical exercices focused in the upper limb administered by occupational therapist.

The subacute stroke phase is a unique opportunity to investigate the role of brain structures in motor learning/control. Compared to chronic impairments (> 6 months post-stroke), the subacute phase provides a window into how a lesion perturbs sensorimotor functions prior to reorganisation driven by plasticity and neurorehabilitation. To clarify the role of different brain structures in MskL, Voxel-based Lesion Symptom Mapping (VLSM) based on high-resolution brain magnetic resonance imaging (MRI) scans, will be used to analyse the relationship between tissue damage and MskL scores on a voxel-by-voxel basis.

In addition, several "classical" clinical scales and tests will be used to evaluate overall motor-sensory-cognitive functions. The subjects will practice serious games on the robot REAplan (R), requiring movements with the affected arm (unimanual tasks).

And, we used a Dextrain® tool which allows quantification of key components of manual dexterity : forces, selectivity (independance of fingers movement) and coactivation of the fingers.Moreover, we added a Transcranial magnetic stimulation (TMS), as a tool for predicting recovery of motor function after stroke.

In addition to the (sub)acute stroke patients, 4 others groups will be recruited for this study : a group of acute stroke patients who will receive "conventional rehabilitation" and be shortly tested on the robot (N=15), a group of chronic stroke patients (stroke > 6 months) who will not be hospitalized and will not undergo MRI (N=30), a group of healthy individuals who will not undergo MRI (N=50) and a group of patients with a transient global amnesia which is a sudden, temporary episode (<24hours) of memory loss (N=15).

Subjects in these 4 groups will be randomized 1/1 in the two arms ("MSkL" versus "motor control recovery" arms), except the patients with a TGA who will be included in the "MSkL" arm only.