Brain Functional Connectivity Classification Method for Voice Pitch Processing in the Deaf Based on One-Dimensional Convolutional Networks

This study aims to investigate the differences in brain functional connectivity between deaf children and hearing children in processing three different tones (first tone, second tone, third tone) using deep learning methods. Additionally, the study aims to identify the disparities in brain activation regions during tone processing between deaf and hearing children. Additionally, the goal is to identify the variations in activated brain regions during tone processing between deaf and typically developing children. Five deaf children and two typically developing children were selected as participants. Resting-state functional magnetic resonance imaging (fMRI) scans were conducted using a functional MRI scanner. Subsequently, data preprocessing and fMRI data analysis were performed. A one-dimensional convolutional network was employed to extract features and classify the input brain functional connectivity. Through a series of theoretical and comparative analyses, it was determined that the one-dimensional convolutional network is capable of capturing local patterns within sequential data. It can extract information related to connectivity patterns from the one-dimensional sequence of connection weights, making it more suitable for processing brain functional connectivity data in deaf children.