NEDA-Net: A neighborhood-enhanced deformable attention network for fMRI brain disease classification

Key Points

• Classification accuracy of the neighborhood-enhanced deformable attention network surpasses traditional methods, with an emphasis on fMRI data.
• The algorithm achieved a significant increase in classification accuracy, outperforming baseline models by 15% in distinguishing brain disorders.
• Assessment conducted on various fMRI datasets, utilizing advanced neural network techniques for improved performance.
• Findings highlight the potential for enhanced classification methodologies in brain disease detection, yet validation through broader datasets is necessary.