Brain region localization magnetic resonance imaging data processing method based on neural stimulation target optimization

Abstract

This invention relates to a brain region localization magnetic resonance imaging (fMRI) data processing method based on neural stimulation target optimization. It acquires functional magnetic resonance imaging (fMRI) data from subjects, extracts signal response sequences of each brain region before and after neural stimulation through time-domain analysis, and identifies brain regions with characteristic response time windows based on differences in response delay. Based on the neural feedback relationship between the cortex and deep structures, it detects delayed oscillation features in cortical signals to identify potential regulatory pathways in the thalamus and other deep brain regions, thereby locating indirect stimulation targets that cannot be directly imaged. It performs decoupling analysis of cerebral blood flow signals and neural activation signals on candidate brain regions; when blood flow fluctuations are detected in a region but synchronous neural activation is lacking, it is identified as a spurious activation area and excluded. This invention calculates the time-delay correlation between the stimulation sequence and brain region signals using an improved cross-correlation integral function, and combines a nonlinear attenuation term to suppress high-frequency noise interference, making the estimation of brain region response delay more stable and repeatable.

Images