A ResAE-MLP deep learning model for understanding and analyzing ADHD based on functional neuroimaging

The discovery and identification of structural or functional changes in brain regions play a crucial role in the clinical diagnosis and treatment of ADHD patients. To fully explore the functional changes in brain regions of ADHD patients, a deep learning model based on fMRI data is introduced. Through the preprocessing and calculation of functional data, the changes in brain region functions can be confirmed and verified. The model introduces an autoencoder to compress, calculate and select features of the preprocessed data, and designs a data-driven deep learning model that is convenient for clinical application. The classifier of the model can not only correctly classify ADHD patients, but also accurately identify the significantly changed regions in the brain of ADHD patients. Meanwhile, the model shows the change relationships of all brain regions and the visualization results of significant brain regions. The research results indicate that deep learning can obtain the changes in brain functions from imaging data, and better explain the clinical structural changes corresponding to brain functions from a data perspective, which demonstrates that the application of deep learning methods for the calculation and clinical application of brain function data is feasible.