Identifying ADHD Individuals From Resting-State Functional Connectivity Using Subspace Clustering and Binary Hypothesis Testing

Abstract

Objective: This study focused on the ADHD classification through functional connectivity (FC) analysis. Method: An ADHD classification method was proposed with subspace clustering and binary hypothesis testing, wherein partial information of test data was adopted for training. By hypothesizing the binary label (ADHD or control) for the test data, two feature sets of training FC data were generated during the feature selection procedure that employed both training and test data. Then, a multi-affinity subspace clustering approach was performed to obtain the corresponding subspace-projected feature sets. With the energy comparison of projected feature sets, we finally identified ADHD individuals for the test data. Results: Our method outperformed several state-of-the-art methods with the above 90% average identification accuracy. By the discriminative FC contribution analysis, it also proved the reliability of our method. Conclusion: Results demonstrate the remarkable classification performance of our method and reveal some useful brain circuits to identify ADHD.

Separated Channel Attention Convolutional Neural Network (SC-CNN-Attention) to Identify ADHD in Multi-Site Rs-fMRI Dataset

“…The first reason was that we directly inputted the preprocessed time-series signals into the model to learn the discriminative features. Unlike the traditional hand-crafted features [ 12 , 14 , 41 ], the resting-state BOLD signal reflected the special activated patterns of the brain [ 42 ], which carried more information. Furthermore, the hand-crafted features mainly represented the single and static brain activity measures, while the resting-state time-series signals measured dynamic brain activity [ 16 , 43 ].…”

“…Bernhardt et al [ 13 ] used an overall LEFMSF algorithm to fuse structural texture features and FC and achieved an accuracy of 67% on the ADHD-200 data with the support vector machines (SVM) classifier. Tang et al [ 14 ] applied a multi-affinity subspace clustering approach to FC feature for identifying ADHD and obtained the best performance of 96.2% for the single-site New York University Medical Center (NYU) dataset.…”

High-Accuracy Classification of Attention Deficit Hyperactivity Disorder With l 2,1-Norm Linear Discriminant Analysis and Binary Hypothesis Testing

“…Different from the training and testing scheme, the binary hypothesis approach provides an alternative way to detect ADHD individuals [18], [19]. Its basic idea is to let the FCs of test data (without seeing its label) affect the selected FCs of training data.…”

“…Subspace projection algorithms are also well developed in feature extraction. Several subspaces are effectively designed in consideration of intra-and inter-class relationship of subjects to improve the classification accuracy [18], [19]. Moreover, since FCs can describe a topographic map of brain, some special graph-based methods are performed for ADHD classification.…”

“…Though ADHD classification with machining learning and deep learning has obtained some impressive achievements, its accuracy performance still has a gap to meet the requirement of clinical diagnosis. In fact, the accuracies of most state-ofthe-art methods fluctuate between 65% and 87% [11], [18], [22], [27], [30]- [32]. The reason comes from two aspects.…”