Decoding individuated finger movements using volume-constrained neuronal ensembles in the M1 hand area

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Individuated finger and wrist movements are controlled by widely distributed neuronal populations in the primary motor cortex (M1).
• Previous brain-machine interface (BMI) studies utilized large neuronal ensembles for decoding movement, but the efficacy of spatially-constrained recordings remains less explored.

Purpose of the Study:

• To investigate the feasibility of decoding dexterous finger movements using spatially-constrained neuronal ensembles, mimicking microelectrode array recordings.
• To determine if decoding accuracy is influenced by the location or configuration of the microelectrode array within the M1 hand area.

Main Methods:

• Recorded single-unit activities from task-related neurons in two rhesus monkeys performing individuated finger and wrist movements.
• Simulated simultaneous neuronal ensembles by constraining recorded activities to microelectrode array dimensions.
• Utilized artificial neural network (ANN) based filters to decode movement trajectories from neuronal ensemble data.

Main Results:

• ANN filters achieved over 90% accuracy in decoding individuated finger movements using as few as 20 neurons.
• Decoding accuracy showed no significant differences across various recording volume locations within the M1 hand area (p < 0.01).

Conclusions:

• Dexterous control of individuated finger and wrist movements can be achieved using brain-machine interfaces with microelectrode arrays.
• The findings suggest that microelectrode arrays can be broadly placed within the M1 hand area for effective BMI implementation, offering flexibility in device design and placement.