Method and system for positioning target in brain region

A system for positioning a target in a brain region, the system comprising:

• a processor;
• and
• a memory having programming instructions stored thereon, which, when executed by the processor, cause the processor to:
• obtain a first stroke dataset and a second stroke dataset of set N persons, wherein sample data in the first stroke dataset is N DWI images and N resting-state fMRI images of the N persons at a first time point after stroke, sample data in the second stroke dataset is N resting-state fMRI images of the N persons at a second time point after stroke, and the second time point is longer than the first time point;
• perform stroke lesion image segmentation on N DWI images in the first stroke dataset by a no-new U-Net (nnU-Net) deep learning method to obtain N lesion masks;
• register the N lesion masks obtained by the segmentation to a brain standard space by linear transformation to obtain N standard-space lesion masks;
• construct N first lesion mapping functional networks based on the N resting-state fMRI images in the first stroke dataset and the N standard-space lesion masks;
• construct an acute phase cognitive-lesion mapping functional network according to the N first lesion mapping functional networks and a preset cognitive scale, comprising:
• construct a linear model for each voxel according to the N first lesion mapping functional networks and the preset cognitive scale;
• and obtaining M linear models, wherein each first lesion mapping functional network comprises M voxels, and the preset cognitive scale comprises cognitive scores of the N persons;
• and wherein the linear model is expressed as;
• y
• =
• a
• +
• b
• *
• x
• i
• +
• c
• 1
• *
• age
• +
• c
• 2
• *
• sex
• +
• c
• 3
• *
• education
• +
• ε
• i
• ,
• =
• 1
• ,
• 2
• ,
• 3
• ,
• …
• ,
• M
• ,
• wherein y is a set of the cognitive scores of the N persons in the preset cognitive scale, xi denotes a value of an ith voxel of the N persons in a corresponding first lesion mapping functional network, age denotes an age covariant, sex denotes a sex covariant, education denotes an education level covariant, a denotes a constant, b denotes a first regression coefficient, c1 denotes a second regression coefficient, c2 denotes a third regression coefficient, ca denotes a fourth regression coefficient, and εi is a residual term;
• select voxels corresponding to linear models with credibility greater than 95%, and
• record the voxels as first cognitive impairment-related voxels;
• and
• set a value at a position of each first cognitive impairment-related voxel in a first preset standard brain network as a hypothesis testing statistic t of a linear model corresponding to the first cognitive impairment-related voxel, and set values at positions other than the positions of the first cognitive impairment-related voxels in the first preset standard brain network to 0 to obtain the acute phase cognitive-lesion mapping functional network;
• construct N second lesion mapping functional networks based on the N resting-state fMRI images in the second stroke dataset and the N standard-space lesion masks;
• construct a chronic phase cognitive-lesion mapping functional network according to the N second lesion mapping functional networks and the preset cognitive scale;
• compare the acute phase cognitive-lesion mapping functional network with the chronic phase cognitive-lesion mapping functional network to obtain a key improvement network for reflecting changes at the first time point and the second time point;
• calculate a whole-brain functional connectivity network with each voxel in a functional image to be positioned as a seed point, and perform spatial correlation calculation on the whole-brain functional connectivity network and the key improvement network to obtain a spatial correlation network, wherein the functional image to be positioned is a functional image obtained by registering an unprocessed resting-state fMRI image to the brain standard space;
• and
• determine a therapeutic target of the functional image to be positioned according to the spatial correlation network.