Investigating Genetic Correlates of ADHD Using a Pseudo-Twin Deep Learning Architecture: The ADHD-Twin Network

Attention-deficit/hyperactivity disorder (ADHD) is a common and highly heritable neurodevelopmental disorder that is associated with a wide range of neuropsychiatric disorders. However, the lack of a comprehensive exploration of the specific genetic mechanisms of ADHD and its causal associations with these disorders, or whether these associations are simply due to confounding biases, remains a major area of unexplored territory. This paper describes a network structure called ADHD Twin, which uses the Pesudo Twin network to explore the genetic factors of attention deficit hyperactivity disorder (ADHD) through genome-wide association study (GWAS) analysis. The model is used to analyse computational analogies of genetic similarities and differences to more efficiently analyse genetic influences on ADHD.The ADHD Twin Network works well with correlations between gene-wide data (e.g. hyperactivity, inattention and impulsivity) and ADHD. The network works well to reduce the influence of non-correlating factors on the results and helps to analyse and guide the most important information in the genes associated with ADHD. To prove the validity of the model, we compared the ADHD model with a linear regression model. The results of the comparative analysis show that our ADHD results are superior to the traditional regression model, and the experimental results indicate that the model as a whole can better analyse the genetic correlation of ADHD.