Multimodal Brain Growth Patterns: Insights from Canonical Correlation Analysis and Deep Canonical Correlation Analysis with Auto-Encoder

Today’s advancements in neuroimaging have been pivotal in enhancing our understanding 1 of brain development and function using various MRI techniques. This study utilizes images from 2 T1-weighted imaging and diffusion-weighted imaging to identify gray matter and white matter 3 coherent growth patterns within 2 years from 9-10-year-old participants of the Adolescent Brain 4 Cognitive Development (ABCD) Study. The motivation behind this investigation lies in the need 5 to comprehend the intricate processes of brain development during adolescence, a critical period 6 characterized by significant cognitive maturation and behavioral change. While traditional methods 7 like Canonical Correlation Analysis (CCA) capture linear interactions of brain regions, deep Canonical 8 Correlation Analysis with autoencoder (DCCAE) extracts nonlinearly brain patterns. The study 9 involves a comparative analysis of changes in gray and white matter over two years, exploring their 10 interrelation based on correlation scores and extracting significant features using both CCA and 11 DCCAE methodologies, and finding the association between the extracted features with cognition 12 and Child Behavior Checklist. The results showed that both CCA and DCCAE components identified 13 similar brain regions associated with cognition and behavior, indicating that brain growth patterns 14 over these two-year period are linear. The variance explained by CCA and DCCAE components for 15 cognition and behavior suggests that brain growth patterns better account for cognitive maturation 16 compared to behavioral changes. This research advances our understanding of neuroimaging analysis 17 and provides valuable insights into the nuanced dynamics of brain development during adolescence. 18 Our GitHub implementation is publicly available on https://github.com/rsapkota1/Multimodal- 19 feature-extraction-using-CCA-and-DCCAE.