Cognitive detection support method and system based on Stroop normal form and near-infrared data

The invention provides a cognitive detection support method and device based on a Stroop normal form and near-infrared data, and the method comprises the steps: taking the coupling strength of an HbO signal in fNIRS data in a brain interval on each frequency band as a brain effect connection diagram, extracting the data features of the brain effect connection diagram through a convolutional neural network, and obtaining the brain effect connection features; extracting data features of the fNIRS data through a convolutional neural network to obtain space-time spectrum features; and splicing the brain effect connection feature, the space-time spectrum feature and the data feature of the task execution result to obtain a fusion feature, and obtaining the probability that the training user belongs to AD, MCI and health by a classifier according to the fusion feature as a cognitive detection support result. According to the method, the cognitive state of the user can be comprehensively evaluated through intelligen