Using Machine Learning to Explore Genotype Effects on Cortical Thickness of Human Brain

Abstract

The human brain is one of the most complex and unstudied parts of our body. One way to explore the cerebral cortex is to receive magnetic resonance imaging output and to calculate different measurements like cortical thickness, cortical volume, white surface total area, etc. A researcher might compare the obtained values across the defined population or through historical changes of one particular subject. Since many factors might have an impact on the brain (genetic factors, inheritance, environmental impact, lifestyle, nutrition, education) there exist limitations in the analysis. In this thesis, we aim to examine several chosen genotypes and cortical thickness in many regions of interest across the brain to understand the hidden relationship between them and possible use in early diagnostics. Since neurodegenerative diseases are not easy to diagnose in time, the preventive analysis should be introduced. For example, some genes markers (E4 allele of the APOE gene) are already known to be associated with higher chances of getting Alzheimer’s disease and people in high-risk group care more about regular health check-ups. Using machine learning techniques to examine genotype effects on cortical thickness brought some meaningful outcomes for further discussion.