Detection of autism in infant via EEG using optimization algorithms

Abstract

Autism Spectrum Disorder ASD stands as a neuro psychiatric disorder distinguished by the impairments in the social reciprocity behavior and interaction or language and signs of sensory function deficits Finding scalable bio markers for its early discovery is challenging on account of the inconsistency in the disease s presentation and the requirement for easy measurements which could be applied habitually during well baby healthrelated check ups Given that the Electroencephalogram EEG recording along with its analysis is a basic device in recognizing and also diagnosing the problem in neuro physiology It uses the signals of EEG for diagnosing persons with ASD These signals encompass lots of data that mirror the behavior of the brain functions that thereby grab the indications for autism assist in early analyzing and also accelerate the treatment The chief goal of this particular thesis is to spot the ASD in children via EEG using optimization algorithms Accordingly this thesis made two significant contributions The first contribution is to develop a method of PSO ANFIS which is an integration of Adaptive Neuro Fuzzy Inference System ANFIS classifier as well as Particle Swarm Optimization PSO for categorizing the diagnosing signals of EEG Initially Savitzky Golay S G is utilized to preprocess the input EEG signal Next Variational Mode Decomposition VMD is utilized to disintegrate the signal newline