Deep learning prediction of attention-deficit hyperactivity disorder in African Americans by copy number variation

Machine learning in attention-deficit/hyperactivity disorder: new approaches toward understanding the neural mechanisms

Attention-deficit/hyperactivity disorder (ADHD) is a highly prevalent and heterogeneous neurodevelopmental disorder in children and has a high chance of persisting in adulthood. The development of individualized, efficient, and reliable treatment strategies is limited by the lack of understanding of the underlying neural mechanisms. Diverging and inconsistent findings from existing studies suggest that ADHD may be simultaneously associated with multivariate factors across cognitive, genetic, and biological domains. Machine learning algorithms are more capable of detecting complex interactions between multiple variables than conventional statistical methods. Here we present a narrative review of the existing machine learning studies that have contributed to understanding mechanisms underlying ADHD with a focus on behavioral and neurocognitive problems, neurobiological measures including genetic data, structural magnetic resonance imaging (MRI), task-based and resting-state functional MRI (fMRI), electroencephalogram, and functional near-infrared spectroscopy, and prevention and treatment strategies. Implications of machine learning models in ADHD research are discussed. Although increasing evidence suggests that machine learning has potential in studying ADHD, extra precautions are still required when designing machine learning strategies considering the limitations of interpretability and generalization.

Attention-deficit/hyperactive disorder updates

Background

Attention-deficit/hyperactive disorder (ADHD) is a neurodevelopmental disorder that commonly occurs in children with a prevalence ranging from 3.4 to 7.2%. It profoundly affects academic achievement, well-being, and social interactions. As a result, this disorder is of high cost to both individuals and society. Despite the availability of knowledge regarding the mechanisms of ADHD, the pathogenesis is not clear, hence, the existence of many challenges especially in making correct early diagnosis and provision of accurate management.

Objectives

We aimed to review the pathogenic pathways of ADHD in children. The major focus was to provide an update on the reported etiologies in humans, animal models, modulators, therapies, mechanisms, epigenetic changes, and the interaction between genetic and environmental factors.

Methods

References for this review were identified through a systematic search in PubMed by using special keywords for all years until January 2022.

Results

Several genes have been reported to associate with ADHD: DRD1, DRD2, DRD4, DAT1, TPH2, HTR1A, HTR1B, SLC6A4, HTR2A, DBH, NET1, ADRA2A, ADRA2C, CHRNA4, CHRNA7, GAD1, GRM1, GRM5, GRM7, GRM8, TARBP1, ADGRL3, FGF1, MAOA, BDNF, SNAP25, STX1A, ATXN7, and SORCS2. Some of these genes have evidence both from human beings and animal models, while others have evidence in either humans or animal models only. Notably, most of these animal models are knockout and do not generate the genetic alteration of the patients. Besides, some of the gene polymorphisms reported differ according to the ethnic groups. The majority of the available animal models are related to the dopaminergic pathway. Epigenetic changes including SUMOylation, methylation, and acetylation have been reported in genes related to the dopaminergic pathway.

Conclusion

The dopaminergic pathway remains to be crucial in the pathogenesis of ADHD. It can be affected by environmental factors and other pathways. Nevertheless, it is still unclear how environmental factors relate to all neurotransmitter pathways; thus, more studies are needed. Although several genes have been related to ADHD, there are few animal model studies on the majority of the genes, and they do not generate the genetic alteration of the patients. More animal models and epigenetic studies are required.

Loss-of-Function Models of the Metabotropic Glutamate Receptor Genes Grm8a and Grm8b Display Distinct Behavioral Phenotypes in Zebrafish Larvae (Danio rerio)

Members of the family of metabotropic glutamate receptors are involved in the pathomechanism of several disorders of the nervous system. Besides the well-investigated function of dysfunctional glutamate receptor signaling in neurodegenerative diseases, neurodevelopmental disorders (NDD), like autism spectrum disorders (ASD) and attention-deficit and hyperactivity disorder (ADHD) might also be partly caused by disturbed glutamate signaling during development. However, the underlying mechanism of the type III metabotropic glutamate receptor 8 (mGluR8 or GRM8) involvement in neurodevelopment and disease mechanism is largely unknown. Here we show that the expression pattern of the two orthologs of human GRM8, grm8a and grm8b, have evolved partially distinct expression patterns in the brain of zebrafish (Danio rerio), especially at adult stages, suggesting sub-functionalization of these two genes during evolution. Using double in situ hybridization staining in the developing brain we demonstrate that grm8a is expressed in a subset of gad1a-positive cells, pointing towards glutamatergic modulation of GABAergic signaling. Building on this result we generated loss-of-function models of both genes using CRISPR/Cas9. Both mutant lines are viable and display no obvious gross morphological phenotypes making them suitable for further analysis. Initial behavioral characterization revealed distinct phenotypes in larvae. Whereas grm8a mutant animals display reduced swimming velocity, grm8b mutant animals show increased thigmotaxis behavior, suggesting an anxiety-like phenotype. We anticipate that our two novel metabotropic glutamate receptor 8 zebrafish models may contribute to a deeper understanding of its function in normal development and its role in the pathomechanism of disorders of the central nervous system.

Application of deep learning algorithm on whole genome sequencing data uncovers structural variants associated with multiple mental disorders in African American patients

Mental disorders present a global health concern, while the diagnosis of mental disorders can be challenging. The diagnosis is even harder for patients who have more than one type of mental disorder, especially for young toddlers who are not able to complete questionnaires or standardized rating scales for diagnosis. In the past decade, multiple genomic association signals have been reported for mental disorders, some of which present attractive drug targets. Concurrently, machine learning algorithms, especially deep learning algorithms, have been successful in the diagnosis and/or labeling of complex diseases, such as attention deficit hyperactivity disorder (ADHD) or cancer. In this study, we focused on eight common mental disorders, including ADHD, depression, anxiety, autism, intellectual disabilities, speech/language disorder, delays in developments, and oppositional defiant disorder in the ethnic minority of African Americans. Blood-derived whole genome sequencing data from 4179 individuals were generated, including 1384 patients with the diagnosis of at least one mental disorder. The burden of genomic variants in coding/non-coding regions was applied as feature vectors in the deep learning algorithm. Our model showed ~65% accuracy in differentiating patients from controls. Ability to label patients with multiple disorders was similarly successful, with a hamming loss score less than 0.3, while exact diagnostic matches are around 10%. Genes in genomic regions with the highest weights showed enrichment of biological pathways involved in immune responses, antigen/nucleic acid binding, chemokine signaling pathway, and G-protein receptor activities. A noticeable fact is that variants in non-coding regions (e.g., ncRNA, intronic, and intergenic) performed equally well as variants in coding regions; however, unlike coding region variants, variants in non-coding regions do not express genomic hotspots whereas they carry much more narrow standard deviations, indicating they probably serve as alternative markers.