Concatenating framework in ASD analysis towards research progress

Abstract

Autism Spectrum Disorder (ASD) is highly complicated neurodevelopment disorder whose increasing prevalence is 1 in 68 individuals (survey of Centers for Disease Control and Preventions). There are various influential’s for ASD. The root cause is not known predominantly even today. But the state of the art of autism in research is, due to autism risk genes showcasing structural & functional brain differences and behavioral features of ASD. Some of the key measuring tools which are multifaceted indicators help to diagnose autism are like: 1.Physiological Detection (emotion assessment from autistic individual), which uses 4 Physiological signals namely electrocardiogram (ECG), skin conductance (SC), respiration and skin temperature. Outcomes were addressed by rating on three scales: arousal, valance and dominance. This approach is non invasive and economical. 2. Exploring the network connectivity in brain, the magnetic resonance imaging (MRI) and functional magnetic resonance imaging (f-MRI) fetches a non invasive approach to map the ordinal patterns of interaction in brain regions to better understand the pathology. 3. Most common machine learning classifier applied to diagnose ASD is Support vector machine (SVM) algorithm. The further implication of Robust SVM (variant of the single SVM) in research progress has improved the accuracy of diagnosing ASD from control group (CG). 4. Last but not the least Deep learning models helps in building model of profound classification accuracy. Early and accurate diagnosis of ASD intensity level leading to selection of correct treatment procedures and thus helps the autistic individual to undergo worth therapies or other relevant treatments.