Real-time removal of stimulation artifacts in closed-loop deep brain stimulation

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Closed-loop deep brain stimulation (DBS) offers improved therapeutic outcomes and reduced side effects.
• Large stimulation artifacts in DBS recordings hinder the development of adaptive stimulation strategies.
• Accurate neural signal detection is crucial for effective closed-loop DBS.

Purpose of the Study:

• To develop and validate a real-time artifact removal method for closed-loop DBS.
• To address the challenge of large stimulation artifacts obscuring neural signals.
• To facilitate the advancement of closed-loop DBS systems with varied stimulation parameters.

Main Methods:

• An irregular sampling technique was proposed to detect and remove stimulation artifacts.
• Artifact peaks were identified using a threshold, and contaminated samples were interpolated.
• The method was validated using simulated signals and in vivo closed-loop DBS in Parkinsonian animal models.

Main Results:

• The irregular sampling method effectively removed stimulation artifacts from simulated signals with low relative errors (2.14%-7.97%) across various frequencies.
• Real-time artifact removal was successfully demonstrated in vivo during closed-loop DBS with continuously changing and adaptive stimulation frequencies.
• The method proved effective for stimulation frequencies ranging from 20 Hz to 180 Hz and variable low/high frequencies.

Conclusions:

• The proposed irregular sampling method offers an effective approach for real-time artifact removal in closed-loop DBS.
• This technique supports closed-loop DBS applications with low, high, and variable frequency stimulation.
• The method is a significant step towards developing more sophisticated and adaptive closed-loop DBS strategies.