Cerebral cortex surface uniform segmentation method based on spectral clustering algorithm

Abstract

The invention relates to a cerebral cortex surface uniform segmentation method based on a spectral clustering algorithm. The cerebral cortex surface uniform segmentation method comprises the followingsteps: firstly, reconstructing a triangularized cerebral cortex surface for structural nuclear magnetic resonance imaging data by utilizing Freesurfer software; secondly, modeling the surface of thecortex into an undirected graph composed of vertexes and edges according to topological information of the surface of the cortex, and constructing a sparse similarity matrix between vertexes of the cortex; and finally, realizing uniform segmentation of the cerebral cortex by using a spectral clustering algorithm. The method is mainly suitable for uniform segmentation of the cerebral cortex surfacegenerated by Freesurfer reconstruction, and a cerebral cortex segmentation map is provided for structure and function analysis. The method has the advantages that different levels of understanding ofcerebral cortex can be realized by setting different clustering numbers and combining functional connection and structural connection analysis; a relatively uniform segmentation map can be generatedby utilizing a local topological relation between vertexes; and the method can be converged to a globally optimal solution in a sample space of any shape.

Description

technical field [0001] The invention belongs to the field of computational neuroscience in medical image processing, and relates to a method for uniform segmentation of cerebral cortex surface based on spectral clustering algorithm, which is mainly suitable for uniform segmentation of cerebral cortex surface reconstructed based on three-dimensional structural nuclear magnetic resonance structural imaging data . Background technique [0002] Functional separation and functional integration are two basic principles of brain functional organization. The principle of functional separation believes that different cognitive functions are realized by different heterogeneous brain regions; while the principle of functional integration believes that the brain is a whole network, and different cognitive functions are realized through long-distance connections between different brain regions. These two principles complement each other and form the neurophysical basis of higher brain f...

AI Technical Summary

Problems solved by technology

The main defect of this type of method is that it does not take into account that the brain is a multi-level organization, and there is no specific "gold standard" for setting the number of clusters, which will have a crucial impact on the clustering results

Images