Heterogeneous Structured Federated Learning with Graph Convolutional Aggregation for MRI-Based Mental Disorder Diagnosis

To relieve the growing burden of mental disorders, deep learning techniques have emerged as a promising tool to aid clinicians by detecting abnormal patterns in neuroimaging data. However, the efficacy of such models is contingent upon access to vast pools of patient data, which is impractical for individual healthcare institutions. Moreover, the privacy-preserving policy regulations governing medical images further complicate the pooling of information necessary for training robust models. Federated Learning (FL) offers a solution to this dilemma by aggregating the local model updates without compromising patient privacy. However, current studies fail to adequately account for the need to personalize models according to the diverse structures of local data. In this work, an effective heterogeneous structured FL framework using graph convolutional aggregation dubbed GAHFL is proposed to diagnose mental disorders on functional magnetic resonance imaging data. In addition, we propose to perform the global model self-evaluation to enable the training to emphasize the samples that are difficult to classify. To solve the catastrophic forgetting problem, we build a historical logit pool to awaken the global model’s recognition ability by performing a server knowledge self-distillation. Empirical evaluations demonstrate that the proposed framework achieves averaged diagnosis AUC values of 69.01% and 69.04% with different sizes of public datasets of ABIDE-I and ADHD-200 datasets, respectively. The ablation studies and robustness validation test further demonstrate the superior performance of our framework.

I. Introduction

Mental disorders, such as autism spectrum disorder (ASD) and attention deficit/hyperactivity disorder (ADHD), have become a growing global public health concern. Their high prevalence gradually poses a huge pressure on the health center services [1], [2]. In the past decades, computer-aided diagnosis (CAD) approaches are developed to address the psychiatrist shortage by automatically analyzing high-resolution medical images, e.g., functional magnetic resonance imaging (fMRI) [3], [4]. fMRI can investigate aberrant neurobiological functions in mental disorders by detecting tiny changes in blood flow [5], [6]. Recently, deep learning-based CAD approaches (DL-CAD), e.g., long short-term memory network (LSTM) [7], gated recurrent units (GRU) [8], and hopfield neural network [9] et al., achieved decent performance in mental disorder diagnosis. However, the successful training of deep learning models tends to require sufficient training samples.