This study aims to explore the possibility of using a multilayer artificial neural network for the classification between children with autism spectrum disorder (ASD) and typically developing (TD) children based on short-time spontaneous hemodynamic fluctuations. Spontaneous hemodynamic fluctuations were collected by a functional near-infrared spectroscopy setup from bilateral inferior frontal gyrus and temporal cortex in 25 children with ASD and 22 TD children. To perform feature extraction and classification, a multilayer neural network called CGRNN was used which combined a convolution neural network (CNN) and a gate recurrent unit (GRU), since CGRNN has a strong ability in finding characteristic features and acquiring intrinsic relationship in time series. For the training and predicting, short-time (7 s) time-series raw functional near-infrared spectroscopy (fNIRS) signals were used as the input of the network. To avoid the over-fitting problem and effectively extract useful differentiation features from a sample with a very limited size (e.g., 25 ASDs and 22 TDs), a sliding window approach was utilized in which the initially recorded long-time (e.g., 480 s) time-series was divided into many partially overlapped short-time (7 s) sequences. By using this combined deep-learning network, a high accurate classification between ASD and TD could be achieved even with a single optical channel, e.g., 92.2% accuracy, 85.0% sensitivity, and 99.4% specificity. This result implies that the multilayer neural network CGRNN can identify characteristic features associated with ASD even in a short-time spontaneous hemodynamic fluctuation from a single optical channel, and second, the CGRNN can provide highly accurate prediction in ASD.

Yaya Liu, Lingyu Xu*, Jun Li*, Jie Yu and Xuan Yu. Exp Neurobiol 2020;29:27-37. https://doi.org/10.5607/en.2020.29.1.27

Functional near-infrared spectroscopy (fNIRS) was used to investigate spontaneous hemodynamic activity in the temporal cortex for typically developing (TD) children and children with autism spectrum disorder (ASD). Forty-seven children participated in the experiments including twenty-five with ASD. Compared with TD children, children with ASD showed weaker bilateral resting-state functional connectivity (RSFC), but much stronger fluctuation magnitude in terms of oxy-hemoglobin (HbO₂) and deoxy-hemoglobin (Hb). Differentiating between ASD and TD based on a support vector machine (SVM) model including bilateral RSFC and the fluctuation power of HbO₂ and Hb as variables could achieve high accurate classification with sensitivity of 81.6% and specificity of 94.6%. This study demonstrates optical brain imaging has the potential for screening children with risk of ASD.

A method for optical image hiding and for optical image encryption and hiding in the Fresnel domain via completely optical means is proposed, which encodes original object image into the encrypted image and then embeds it into host image in our modified Mach-Zehnder interferometer architecture. The modified Mach-Zehnder interferometer not only provides phase shifts to record complex amplitude of final encrypted object image on CCD plane but also introduces host image into reference path of the interferometer to hide it. The final encrypted object image is registered as interference patterns, which resemble a Fresnel diffraction pattern of the host image, and thus the secure information is imperceptible to unauthorized receivers. The method can simultaneously realize image encryption and image hiding at a high speed in pure optical system. The validity of the method and its robustness against some common attacks are investigated by numerical simulations and experiments.