Pilot study for the Velocity Trajectory Optimization for Upper Limb Rehabilitation of ALS Patients

Research Objectives: To investigate the effect of active assisted arm movement on upper limb muscle activity across participants. Results will be used to determine if a human-in-the-loop optimization can be used to develop personalized exoskeletons for rehabilitation of ALS Patients

Design: A robot end-effector (KinArm) provided active assistance while participants moved between two targets, simulating a forward reaching task along fore-aft direction. Movement was repeated 10 times for 7 different velocities in randomized order (preventing ordering effect). EMG Delsys Trigno sensors placed on seven muscles: Upper Trapezius, Biceps Brachii, Triceps Brachii, Brachioradialis, Anterior Deltoid, Posterior Deltoid, Pectoralis Major.

Setting: Institutional practice laboratory

Participants: 3 Healthy males (Age: 20, 22, 27 yrs, Height: 1.84, 1.72, 1.63 m, Weight: 65, 68, 70 kg)

Interventions: Not applicable.

Main Outcome Measures: Pearson correlation coefficient between different velocities and EMG activity calculated for each muscle and each subject.

Results: Raw EMG readings processed with Neurokit2 library in Python3. Fourth order high-pass Butterworth filter of 100 Hz applied, data normalized to zero, full-wave rectified and RMS envelope taken. Angular velocity data obtained from KinArm encoders was processed to obtain velocity along the fore-aft direction. Linear correlation coefficient between linear velocity and processed EMG data obtained for each muscle obtaining different results for each subject.

Subject 1: the mean EMG activity and mean hand velocity showed a significant (p < 0.05) correlation for upper trapezius (r=0.7969), posterior deltoid(r=0.9525), triceps brachii(0.8825) and biceps (r=0.4033).
Subject 2: the correlation was significant (p < 0.05) for Posterior Deltoid (r=0.6341), Triceps Brachii (r=0.8735), Biceps Brachii (r=0.8350) and Brachioradialis (r=0.8909)
Subject 3: no correlation was found (p>0.581, r < 0.26)

Conclusions: Data analysis showed different correlation across subjects, suggesting personalized assistance using HIL is required to account for the individual differences in the personal conditions of ALS patients.

Author(s) Disclosures: Authors declare no conflict