Wang, David: Assessing the Reliability of Morphometricity as an Anatomical Index

Title: Assessing the Reliability of Morphometricity as an Anatomical Index

Name: David Wang

Major: Computer Science

School affiliation: School of Arts and Sciences

Programs: Aresty Summer Science Program

Other contributors: Yihong Zhao

Abstract: Morphometricity is a measurement that determines how much variation of observable traits can be explained by brain morphology. In other words, it is an approach that examines associations between observable traits and structural measures of the brain such as thickness, area, and volume. Currently, there is a method that calculates this measurement created in Matlab. However, the method can be improved to increase the accuracy in the results. This project used simulation tests to discover the flaws. In order to do these simulations, the Matlab code was converted into R code. Uniform randomly generated numbers were used to calculate variance statistics which were then used to compare with the results from the method. These simulation tests demonstrated that the current method had widely varying effects on the results based on the algorithm it used, which reduced the overall accuracy. Thus, the current method of calculating morphometricity can be improved upon to provide researchers with a more accurate result. Having an accurate morphometricity value will enable researchers to identify developmental risk of alcohol and substance use at brain and behavioral levels. This will also be helpful for other researchers who are studying brain morphology and observable traits as it can reveal association that might not be detectable through traditional statistical techniques.