The Research Question
Can we reliably record individual motor unit activity from the abductor pollicis brevis (APB) - the thumb muscle? If so, we could enable more precise prosthetic control based on neural signals.
The Challenge
Standard surface EMG decomposition algorithms struggled with the APB muscle:
-
Inherent instability of the small muscle
-
Insufficient motor units detected in 1-2 hour sessions
-
Signal quality issues from the muscle's location
The Approach
Adaptive EMG Decomposition
We started implementing an adaptive algorithm that could:
-
Track motor units as they drift over time
-
Compensate for electrode movement
-
Maintain stable recordings over longer sessions
Alternative Validation
While developing the adaptive algorithm (a time-consuming process), we conducted parallel studies using the tibialis muscle as an alternative, ensuring research progress continued.
Future Goals
The project aims to demonstrate:
-
Stable recordings from the APB muscle
-
Effectiveness of adaptive decomposition
-
Applicability to prosthetic control
Why Motor Units Matter
Motor units are the fundamental units of movement control:
-
Each motor unit = one motor neuron + the muscle fibers it controls
-
Decoding motor unit activity = understanding movement intention
-
More precise than bulk EMG signals
Connection to Prosthetics
If we can reliably decode motor unit activity from residual muscles, we could create prosthetics that respond to the user's precise intentions - not just gross muscle activity, but the actual neural commands the brain is sending.