Magnetic Resonance Image from Children’s Brain by Evaluating IQ Estimator Using Kernel Support Vector Regression

Abstract

Human brain explains the function of the various parts of our body. The Brain has been categorized into three parts Cerebrum, cerebellum and brain stem. In this, cerebrum plays a vital role in the brain which controls reading, thinking, learning, speech, emotions, vision, hearing and\n other senses. The Prefrontal cortex (PFC) is the cerebral cortex which covers the front part of the frontal lobe and the most strongly recommended in human qualities like consciousness, general intelligence, and personality. To find out one’s intelligence, most of the IQ score is commonly\n obtained from many diverse tests with disadvantages of no single large dataset containing psychology test scores for the IQ estimation is unavailable. In this paper deals about estimating the IQ from young children Brain MRI images with Kernel Support Vector Regression (SVR) is designed. The\n experimentation is carried out to verify the performance of different feature reduction by applying four extracted features Grey Matter Volume, White Matter Volume, Gyri and Sulci Surface Area with the target only as Wechsler Abbreviated Scale Intelligence (WASI) IQ Score has been collected\n from ABIDE II MRI images of 232 children between 6 and 15 years of age. Kernel SVR is used to estimate IQ scores between the concrete and valued IQs on using reduced features obtained from PCA, Kernel PCA, t-SNE and Kernel t-SNE with root mean square error of 1.0021, 0.459, 0.976 and 1.517\n and acquired Pearson’s correlation coefficients of 0.938, 0.9864, 0.941 and 0.843 respectively.