No abstract

No abstract

Functional magnetic resonance imaging (fMRI) has helped characterize the pathophysiology of autism spectrum disorders (ASD) and carries promise for producing objective biomarkers for ASD. Recent work has focused on deriving ASD biomarkers from resting-state functional connectivity measures. However, current efforts that have identified ASD with high accuracy were limited to homogeneous, small datasets, while classification results for heterogeneous, multi-site data have shown much lower accuracy. In this paper, we propose the use of recurrent neural networks with long short-term memory (LSTMs) for classification of individuals with ASD and typical controls directly from the resting-state fMRI time-series. We used the entire large, multi-site Autism Brain Imaging Data Exchange (ABIDE) I dataset for training and testing the LSTM models. Under a cross-validation framework, we achieved classification accuracy of 68.5%, which is 9% higher than previously reported methods that used fMRI data from the whole ABIDE cohort. Finally, we presented interpretation of the trained LSTM weights, which highlight potential functional networks and regions that are known to be implicated in ASD.

A deep learning model trained on some labeled data from a certain source domain generally performs poorly on data from different target domains due to domain shifts. Unsupervised domain adaptation methods address this problem by alleviating the domain shift between the labeled source data and the unlabeled target data. In this work, we achieve cross-modality domain adaptation, i.e. between CT and MRI images, via disentangled representations. Compared to learning a one-toone mapping as the state-of-art CycleGAN, our model recovers a manyto-many mapping between domains to capture the complex cross-domain relations. It preserves semantic feature-level information by finding a shared content space instead of a direct pixelwise style transfer. Domain adaptation is achieved in two steps. First, images from each domain are embedded into two spaces, a shared domain-invariant content space and a domain-specific style space. Next, the representation in the content space is extracted to perform a task. We validated our method on a crossmodality liver segmentation task, to train a liver segmentation model on CT images that also performs well on MRI. Our method achieved Dice Similarity Coefficient (DSC) of 0.81, outperforming a CycleGAN-based method of 0.72. Moreover, our model achieved good generalization to joint-domain learning, in which unpaired data from different modalities are jointly learned to improve the segmentation performance on each individual modality. Lastly, under a multi-modal target domain with significant diversity, our approach exhibited the potential for diverse image generation and remained effective with DSC of 0.74 on multi-phasic MRI while the CycleGAN-based method performed poorly with a DSC of only 0.52.

Finding the biomarkers associated with ASD is helpful for understanding the underlying roots of the disorder and can lead to earlier diagnosis and more targeted treatment. A promising approach to identify biomarkers is using Graph Neural Networks (GNNs), which can be used to analyze graph structured data, i.e. brain networks constructed by fMRI. One way to interpret important features is through looking at how the classification probability changes if the features are occluded or replaced. The major limitation of this approach is that replacing values may change the distribution of the data and lead to serious errors. Therefore, we develop a 2-stage pipeline to eliminate the need to replace features for reliable biomarker interpretation. Specifically, we propose an inductive GNN to embed the graphs containing different properties of task-fMRI for identifying ASD and then discover the brain regions/subgraphs used as evidence for the GNN classifier. We first show GNN can achieve high accuracy in identifying ASD. Next, we calculate the feature importance scores using GNN and compare the interpretation ability with Random Forest. Finally, we run with different atlases and parameters, proving the robustness of the proposed method. The detected biomarkers reveal their association with social behaviors and are consistent with those reported in the literature. We also show the potential of discovering new informative biomarkers. Our pipeline can be generalized to other graph feature importance interpretation problems.