Autism spectrum disorder (ASD) is a heterogenous neurodevelopmental disorder which is characterized by impaired communication, and limited social interactions. The shortcomings of current clinical approaches which are based exclusively on behavioral observation of symptomology, and poor understanding of the neurological mechanisms underlying ASD necessitates the identification of new biomarkers that can aid in study of brain development, and functioning, and can lead to accurate and early detection of ASD. In this paper, we developed a deep-learning model called \emph{ASD-SAENet} for classifying patients with ASD from typical control subjects using fMRI data. We designed and implemented a sparse autoencoder (SAE) which results in optimized extraction of features that can be used for classification. These features are then fed into a deep neural network (DNN) which results in superior classification of fMRI brain scans more prone to ASD. Our proposed model is trained to optimize the classifier while improving extracted features based on both reconstructed data error and the classifier error. We evaluated our proposed deep-learning model using publicly available Autism Brain Imaging Data Exchange (ABIDE) dataset collected from 17 different research centers, and include more than 1035 subjects. Our extensive experimentation demonstrate that \emph{ASD-SAENet} exhibits comparable accuracy (70.8\%), and superior specificity (79.1\% ) for the whole data set as compared to other methods. Further, our experiments demonstrate superior results as compared to other state-of-the-art methods on 12 out of the 17 imaging centers exhibiting superior generalizability across different data acquisition sites and protocols. The implemented code will be available on GitHub portal of our lab at \url{https://github.com/pcdslab/ASD-SAENet}.

中文翻译：

---

ASD-SAENet：稀疏自动编码器，用于使用fMRI数据检测自闭症谱系障碍（ASD）

自闭症谱系障碍（ASD）是一种异质性神经发育障碍，其特征是沟通障碍和社交互动受限。当前仅基于症状的行为观察的临床方法的缺点，以及对ASD潜在神经机制的了解不足，因此需要鉴定新的生物标记物，以帮助研究大脑发育和功能，并能导致准确，早期的诊断。检测ASD。在本文中，我们开发了一种名为\ emph {ASD-SAENet}的深度学习模型，用于使用fMRI数据对典型对照对象中的ASD患者进行分类。我们设计并实现了一种稀疏自动编码器（SAE），该算法可优化提取可用于分类的特征。然后将这些特征输入到深度神经网络（DNN）中，从而对fMRI脑部扫描进行更好的分类，从而更容易出现ASD。我们提出的模型经过训练可以优化分类器，同时基于重构数据误差和分类器误差改进提取的特征。我们使用从17个不同研究中心收集的公开可用的自闭症脑成像数据交换（ABIDE）数据集评估了我们提出的深度学习模型，其中包括1035多个受试者。我们广泛的实验表明\ emph {ASD-SAENet}与其他方法相比，在整个数据集上具有可比的准确性（70.8 \％）和更高的特异性（79.1 \％）。进一步，在17个成像中心中的12个中，我们的实验证明了与其他最新方法相比的出色结果，这些成像中心在不同的数据采集站点和协议上均具有出色的通用性。实施的代码将在我们实验室的GitHub门户上提供，网址为\ url {https://github.com/pcdslab/ASD-SAENet}。