Carbon Nanotube Enhanced Flexible Dry Electrodes for Wearable Brain Computer Interface Systems

Abstract

Noninvasive brain computer interfaces are limited by high electrode skin impedance, motion artifacts and poor mechanical conformity. These lower the quality of electroencephalography signals and decoding performance. This research explores carbon nanotube enhanced flexible dry electrodes to improve neural signal acquisition through increased electrical conductivity, enlarged effective surface area and improved interface stability. Carbon nanotubes are integrated into a polymer based flexible substrate using solution processed composite deposition methods like spray coating or drop casting which form a conformal electrode skin interface without the need for conductive gels. An equivalent circuit model of the electrode skin interface is developed to characterize resistive and capacitive components and compare impedance across electroencephalography relevant frequency bands. Frequency dependent impedance spectra are evaluated to find the reductions in contact resistance and improvements in interfacial capacitance associated with carbon nanotube integration. Signal acquisition is simulated under varying noise and motion conditions using synthetic and benchmark electroencephalography datasets with controlled artifact injection to represent movement induced disturbances and baseline environmental noise. The resulting data are preprocessed through filtering and normalization to be analyzed using standard feature extraction techniques which include band power analysis, spectral density estimation and time frequency decomposition. Extracted features are evaluated using machine learning methods: linear discriminant analysis and support vector machine classifiers. These help to investigate changes in signal to noise ratio, class separability and decoding accuracy. Results indicate that carbon nanotube based electrodes can reduce impedance, improve signal stability and enhance downstream decoding performance. These show that the advanced interface materials are critical for wearable closed loop brain computer interface systems.

Overview

This work looks at the sensing side of wearable BCIs: how flexible dry electrodes enhanced with carbon nanotubes could improve EEG acquisition through lower impedance, greater interface stability, and better signal-to-noise performance.

Electrode model

The work is computational and modeling based. An equivalent circuit represents the resistive and capacitive behavior of the electrode–skin interface across EEG-relevant frequencies. The analysis examines how carbon-nanotube integration could affect contact resistance, interfacial capacitance, conformity, and stability without implying that physical electrodes were fabricated.

Signal evaluation

Synthetic and benchmark EEG data are evaluated under controlled noise and motion artifacts. Filtering, normalization, band-power analysis, spectral-density estimation, and time-frequency decomposition feed classification with linear discriminant analysis and support vector machines to examine signal-to-noise ratio, separability, and decoding behavior.