Visual evoked potentials estimation by adaptive noise cancellation with neural-network-based fuzzy inference system

Area of Science:

• Neuroscience
• Signal Processing
• Artificial Intelligence

Background:

• Visual Evoked Potentials (VEPs) are crucial for assessing visual pathway function but are often obscured by background electroencephalogram (EEG) noise.
• Accurate VEP signal extraction is challenging due to the low signal-to-noise ratio.

Purpose of the Study:

• To develop and validate a novel adaptive noise cancellation method for VEP signal estimation.
• To demonstrate the efficacy of a neural-network-based fuzzy inference system (NNFIS) for real-time VEP tracking.

Main Methods:

• An adaptive noise cancellation technique employing a neural-network-based fuzzy inference system (NNFIS) was designed to model VEP signals.
• The NNFIS, based on the Takagi-Sugeno fuzzy model, offers linear-in-parameter advantages for function mapping and dynamic tracking.
• The method was evaluated using simulated data and real-time processing of 150 trials.

Main Results:

• Simulated data demonstrated the proposed NNFIS method's appropriateness for VEP estimation.
• The NNFIS successfully modeled and extracted VEP signals in real-time.
• Processing 150 trials confirmed the superior performance of the proposed NNFIS-based approach.

Conclusions:

• The NNFIS-based adaptive noise cancellation is an effective and reference-free method for VEP signal estimation.
• This technique offers a robust solution for analyzing VEPs in noisy EEG recordings.
• The NNFIS's ability to track dynamic VEP behavior in real-time holds significant potential for clinical and research applications.