Researchers aimed to diagnose schizophrenia using brain imaging data and found that schizophrenia patients exhibit significant abnormalities in brain structure and function compared to healthy controls (HC). They applied the node2vec algorithm for graph embedding to transform the brain network data into low-dimensional dense vectors, preserving the non-Euclidean spatial characteristics of the data. They also employed a transformer model to extract features from the preprocessed data and identify schizophrenia. The study demonstrated that using node2vec, in combination with the transformer model and GridMask, effectively addresses the challenge of learning brain network features with deep learning models, offering the potential for high-precision computer-aided diagnosis of schizophrenia based on brain-imaging data.
Reference: Gan A, Gong A, Ding P, Yuan X, et al. Computer-aided diagnosis of schizophrenia based on node2vec and Transformer. J Neurosci Methods. 2023 Apr 1;389:109824. doi: 10.1016/j.jneumeth.2023.109824. Epub 2023 Feb 22. PMID: 36822277.
Link: https://schizophrenia.pocn.com/diagnosis/node2vec-and-transformer-model-brain-imaging-data/
