Interpretable attention-deficit hyperactivity disorder diagnosis from functional magnetic resonance imaging via hybrid convolutional neural network–transformer learning and multi-level graph-based comprehensibility

Attention-Deficit Hyperactivity Disorder (ADHD) is a disorder about the brain development with various changes in the brain structure, involving constant diagnosis using predictable neuroimaging analysis. Current deep learning methods appear to show promising performance in computerized ADHD classification using magnetic resonance imaging (MRI); however, their applicability is restricted due to ambiguity. The proposed study is an interpretable hybrid deep learning model for ADHD recognition using fMRI.

It uses a three-dimensional Convolutional Neural Network (CNN) to map brain structures and a transformer encoder to analyse transformer modules. At the voxel level, 3D Grad-CAM + + limits the characteristic feature support for ADHD classification. At the interpretative level, cross-layer attention aggregation creates reliable attention patterns during transformer layers, streamlining the identification of related brain structures.

A brain connectivity graph is created by extracting region-specific features using a neuroanatomical atlas, and this graph is then inspected using a Graph Neural Network (GNN). GNNExplainer revealed complex brain subnetworks and interregional interactions related to ADHD. Experiments conducted on the ADHD-200 dataset demonstrated that the proposed hybrid framework exceeded the CNN-only and Transformer-only baselines in terms of accuracy, sensitivity, specificity, and AUC.

These results show that the proposed approach has better diagnostic efficacy while providing substantial benefits, progressing the feasibility of deep-learning-based ADHD diagnosis from MRI. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Lakshmi Priya The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it.

The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.