Stimulus Optimization Using the Artificial Bee Colony Algorithm for Visual Neuroprostheses

Area of Science:

• Biomedical Engineering
• Computational Neuroscience
• Artificial Intelligence

Background:

• Visual neuroprosthetics aim to restore vision by electrically stimulating the retina.
• Current stimulation patterns often lack the precision of natural neural responses.
• Optimizing stimulation is key to enhancing prosthetic performance and user experience.

Purpose of the Study:

• To develop and evaluate an Artificial Bee Colony (ABC) algorithm for optimizing electrical stimulation patterns.
• To generate physiologically realistic neural responses in retinal ganglion cells (RGCs).
• To advance the development of intelligent stimulation strategies for visual neuroprosthetics.

Main Methods:

• An in silico approach using the ABC algorithm to optimize stimulus waveforms.
• Utilizing a computational model of ON and OFF RGCs.
• Matching simulated responses to recorded neural data from the rat dorsolateral geniculate nucleus (dLGN).

Main Results:

• The ABC algorithm achieved >=80% cross-correlation with target neural responses within 2000 iterations.
• Identified optimized waveforms eliciting distinct responses in neighboring neurons.
• Demonstrated superior accuracy compared to alternative optimization algorithms under computational constraints.

Conclusions:

• The ABC-based approach effectively generates natural-like neural activation patterns.
• This method supports real-time and adaptive stimulation control for visual neuroprosthetics.
• Introduces a biologically inspired framework for intelligent stimulation strategies.