A Survey: Strategies for detection of Autism Syndrome Disorder

The Autism Spectrum Disorder (ASD) is known to be characterized by decreased social interactions to stimuli, difficulties in communication, repetitive behaviours and physical movements, unusual or severely limited interests. Since it affects different individuals differently and to different extents, it is very challenging to diagnose and differentiate from other neurodevelopmental disorders. The earlier it is diagnosed, the better it is for the individual, as an intervention protocol can be charted at the earliest. Research shows that early intervention can be very beneficial in the long term as the affected individuals’ symptoms can be reduced to a great extent and in some cases they may not show up on the spectrum at all, but the lack of awareness and objective diagnostic tools makes it difficult to diagnose at an early age. The objective nature of the automated approaches to the screening of Autism Spectrum Disorder would be considered as a viable option for a second opinion and by an extension also an efficient screening process after which the patients would be referred to a therapist. Tests like eye tracking, detection of physical attributes via wearable and static IoT devices, classification on MRI scans and voice prosody detection techniques are easily reproducible and more time efficient than a one on one interview session with a therapist. In this paper we explore the various innovations made in automating the screening process. This paper lists the demerits of the subjective existing diagnosis systems and also surveys and complies the various eye gazing, voice prosody, wearable IoT modules, their contributions and their impacts in making the diagnosis process of Autism Spectral Disorder more objective and efficient.