MoS2 photonic synaptic device development for non-invasivedetection of Parkinson’s disease

Abstract The brain-inspired computation using synaptic devices has opened a new avenue for neuromorphic intelligent systems. This paper presents the development of a MoS2 based photonic synaptic device using wafer-scalable microfabrication for large scale production. The synaptic device was tested using a 450 nm wavelength having a power density of 17 mW/cm2. The test results show that developed devices have enhanced characteristics such as paired pulse facilitation (PPF) index of 307% and long-term retention of 1000 sec. The optically stimulated synaptic current was ~2.5 times higher than the electrically stimulated synaptic current. The potentiation and depression data of the MoS2 device act as the synaptic interface in the artificial intelligence (AI) model. The detection of Parkinson’s disease (PD) using voice spectrograms where MoS2 behaves as photonic synapses in two dimensional (2D) convolutional neural networks (CNN) algorithm provided a promising accuracy of 96%. This neuromorphic computing platform for the detection of PD is a next-generation revolution for efficient and smart healthcare solutions.