This invention relates to the field of medical
signal processing technology, specifically disclosing an AI-assisted diagnostic method and
system for Parkinson's
disease based on brain networks. The method includes: acquiring continuously collected temporal signals from the brain and establishing a discrete sequence; establishing an effective window set through
signal activity screening; evaluating the temporal
connectivity of the effective window set to obtain a steady-state
data stream; analyzing the
phase synchronization index and determining significant functional
coupling to construct a topological confidence structure of the
brain network; establishing an equivalent resistance model based on the topological confidence structure, outputting global
conductivity, and tracking
potential energy drops to establish a transmission cost subgraph; and performing gridded
density analysis based on the transmission cost subgraph to obtain spatial damping distribution and evaluate the
brain network's operational state. The
system includes: a
signal acquisition module, a temporal evaluation module, a confidence analysis module, a blockage analysis module, and a state evaluation module. This invention is beneficial for assisting in the diagnosis of Parkinson's
disease and providing data support.