Date of Graduation

5-2021

Document Type

Thesis

Degree Name

Bachelor of Science

Degree Level

Undergraduate

Department

Computer Science and Computer Engineering

Advisor/Mentor

Zhan, Justin

Committee Member/Reader

Gauch, John

Committee Member/Second Reader

Panda, Brajendra

Committee Member/Third Reader

.

Abstract

Analyzing the correlation between brain volumetric/morphometry features and cognition/behavior in children is important in the field of pediatrics as identifying such relationships can help identify children who may be at risk for illnesses. Understanding these relationships can not only help identify children who may be at risk of illnesses, but it can also help evaluate strategies that promote brain development in children. Currently, one way to do this is to use traditional statistical methods such as a correlation analysis, but such an approach does not make it easy to generalize and predict how brain volumetric/morphometry will impact cognition/behavior. One of the cognition behaviors that can be predicted is the IQ score. In the age of artificial intelligence and machine learning, it has become fundamental to be able to exploit techniques such as deep learning to automate and improve tasks. One of the types of data that is used to make such assessments is mean diffusivity data (MD). In this paper, I propose using a machine learning approach and use the MD data to predict the IQ score of healthy 8-year old children. These predictions will provide insight into how well MD represents the IQ score of healthy 8-year old children and they will allow experts better understand how a child’s neuropsychological score is affected by volumetric/morphometry data. In this paper I examine five different neural network models for predicting the IQ score of healthy 8-year old children. Each model is different in either the architecture and how the data is processed before it is fed to into the neural network. After analyzing these models, I found that the best performing neural network for the data set we are working with consists of using the Principal Component Analysis (PCA) for feature reduction and standardization of the data. On average, this model’s IQ score predictions deviate from the true IQ score by 8.09%. Given the small data set and the high dimensionality of the data, it is concluded that this model’s IQ score predictions are reasonable and well modeled IQ scores for healthy 8-year old children.

Keywords

deep learning; children brain imaging; analysis; mean diffusivity

Download

Share

COinS