Machine learning reveals symptoms of aggression in childhood ADHD


Child psychiatric disorders such as oppositional defiance disorder and attention deficit / hyperactivity disorder (ADHD) can include outbursts of anger and physical aggression. A better understanding of the causes of these symptoms could help develop treatment strategies. Yale researchers have now used a machine learning-based approach to uncover brain connectivity disorders in children who exhibit aggression.

While previous research focused on specific brain regions, the new study identifies patterns of neural connections throughout the brain that are associated with aggressive behavior in children. The results published in the journal Molecular Psychiatry, build on a novel model of brain function called a “connectome” that describes this pattern of brain-wide connections.

Maladaptive aggression can harm yourself or others. This challenging behavior is one of the main reasons for referrals to child psychiatry, ”said Denis Suchodolsky, Senior Writer and Associate Professor at the Yale Child Study Center. “Connectome-based modeling offers a new representation of brain networks involved in aggressive behavior.”

For the study, which is the first of its kind, the researchers collected functional magnetic resonance imaging (fMRI) data while children performed an emotional facial awareness task where they observed faces that made calm or fearful expressions. Seeing faces that express emotions can trigger brain states relevant to emotion generation and regulation, both of which have been linked to aggressive behavior, the researchers said. The scientists then applied machine learning analysis to identify neural connections that characterized children with and without aggressive behavior.

They found that patterns in brain networks involved in social and emotional processes – such as frustration with homework or understanding why a friend is upset – predicted aggressive behavior. To confirm these results, the researchers then tested them in a separate dataset and found that the same brain networks predicted aggression. In particular, abnormal connectivity to the dorsolateral prefrontal cortex – a key region involved in the regulation of emotions and higher cognitive functions such as attention and decision-making – has been shown to be a consistent predictor of aggression when used in subsets of children with aggressive behavior and disorders such as like anxiety, ADHD and autism.

These neural connections to the dorsolateral prefrontal cortex could represent a marker of aggression that occurs in several childhood psychiatric disorders.

This study suggests that the robustness of these large-scale brain networks and their connectivity to the prefrontal cortex could represent a neural marker of aggression that can be used in clinical trials, ”said Karim Ibrahim, Associate Research Scientist at the Yale Child Study Center and first author of the article. “The human functional connectome describes the enormous networking of the brain. Understanding the connectome is at the forefront of neuroscience as it can provide us with valuable information for developing brain biomarkers for psychiatric disorders. “

Sukhodolsky added, “This connectome model of aggression could also help us develop clinical interventions that can improve coordination between these brain networks and nodes like the prefrontal cortex. Such interventions could include learning the emotion regulation skills necessary to modulate negative emotions such as frustration and anger. “


K. Ibrahim, S. Noble, G. He et al. Large-scale functional brain networks of maladaptive aggression in childhood, identified by connectoma-based predictive modeling. Mol psychiatry. Published online October 25, 2021: 1-15. do:10.1038 / s41380-021-01317-5

This article was republished from the following materials. Note: The material may have been edited for length and content. For more information, please refer to the source cited.


Leave A Reply