Toward optimal target placement for neural prosthetic devices.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Neural prosthetic systems, including motor and communication prostheses, decode neural activity to control devices.
• Current communication prostheses often use fixed, radially symmetric target geometries, irrespective of neuron tuning properties.
• Optimizing target placement is crucial for enhancing the accuracy and efficiency of neural decoding.

Purpose of the Study:

• To develop and validate an optimal target placement algorithm for neural prostheses.
• To automate the selection of target locations based on neural population characteristics.
• To improve decoding accuracy in communication prostheses and potentially motor prostheses.

Main Methods:

• An optimal target placement algorithm was developed to maximize decoding accuracy with respect to target locations.
• The algorithm was tested using simulated neural spiking data from two monkeys.
• Experimental neural data from a trained monkey was used to validate simulation findings.

Main Results:

• The optimal target placement algorithm demonstrated statistically significant improvements in decoding accuracy, up to 8% for two targets and 9% for sixteen targets in simulations.
• For four and eight targets, the algorithm's performance gains were modest as its proposed layouts closely resembled canonical layouts.
• Validation with experimental neural data confirmed simulation results, showing the algorithm's practical applicability.

Conclusions:

• The developed algorithm is the first of its kind for optimizing neural prosthesis target placement.
• This approach can generate novel target layouts that outperform traditional ones and confirm or select among existing layouts.
• The algorithm promises to enhance decoding accuracy and automate target placement for neural prosthetic systems.