Optimal Multichannel Artifact Prediction and Removal for Neural Stimulation and Brain Machine Interfaces

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Neural implants for neurological disorders increasingly require real-time neural function assessment.
• Multi-site electrical stimulation is crucial for studying neural circuits but faces challenges with stimulus-evoked artifacts obscuring neural signals.

Purpose of the Study:

• To develop a versatile artifact removal method for multi-site electrical stimulation and recording.
• To enhance the quality of neural recordings in closed-loop systems and brain-machine interfaces.

Main Methods:

• Developed a novel method leveraging linear electrical coupling between stimulating currents and recording artifacts.
• Estimated a multi-channel linear Wiener filter to predict and subtract stimulus-evoked artifacts.
• Validated the method across various recording modalities including in vitro sciatic nerve, cochlear implants, and auditory midbrain-cortex recordings.

Main Results:

• Demonstrated significant artifact reduction, typically 25-40 dB, leading to vastly enhanced recording quality.
• Confirmed the linearity assumption underpinning the artifact removal technique.
• Showcased the method's efficiency and scalability for large-scale arrays and complex neural interfaces.

Conclusions:

• The novel artifact removal method is versatile and effective for multi-site stimulation and recording.
• This technique is ideal for advanced applications like closed-loop implants and high-resolution brain-machine interfaces.
• Improved neural recording quality facilitates better understanding and treatment of neurological disorders.