AI can help predict aggressive behavior in people with autism

Credit: Unsplash+

A groundbreaking study, published online in JAMA Network Open, reveals that machine learning can effectively predict imminent aggressive behaviors in inpatient youths with autism.

This study, led by Tales Imbiriba, Ph.D., from Northeastern University in Boston, marks a significant step forward in understanding and managing challenging behaviors associated with autism.

The research team conducted a noninterventional prognostic study from March 2019 to March 2020, involving 70 psychiatric inpatients with confirmed autism diagnoses.

These participants, from four primary care psychiatric inpatient hospitals, exhibited self-injurious behavior, emotion dysregulation, or aggression towards others. Notably, 32 of these individuals were minimally verbal, and 30 had an intellectual disability.

Participants were equipped with a commercially available biosensor that recorded peripheral physiological signals. The study extracted and analyzed time-series features from the biosensor data to identify patterns preceding aggressive incidents.

During the study period, researchers conducted 429 observational coding sessions, amounting to 497 hours.

In these sessions, they documented 6,665 aggressive behaviors, categorized as self-injury (59.8%), emotion dysregulation (31.0%), and aggression towards others (9.3%).

The study’s most significant finding was the effectiveness of logistic regression as a classifier in predicting aggressive behavior.

It showed remarkable accuracy, with a mean area under the receiver operating characteristic curve of 0.80, particularly in predicting aggressive behavior three minutes before its onset.

The authors suggest that these findings could pave the way for developing mobile health systems that provide just-in-time adaptive interventions.

Such technology would offer new possibilities for preemptive intervention, focusing on reducing the unpredictability of aggressive behavior in autistic youths.

By enhancing the predictability and management of these behaviors, the research has the potential to significantly improve the quality of life for inpatient youths with autism, allowing them to more fully engage in their homes, schools, and communities.

This study represents a promising advancement in the field of autism research and intervention strategies.

If you care about autism, please read studies about a new cause of autism, and cats may help decrease anxiety for kids with autism.

For more information about health, please see recent studies about vitamin D that may hold the clue to more autism, and results showing strange eating habits may signal autism.

The research findings can be found in JAMA Network Open.

Copyright © 2023 Knowridge Science Report. All rights reserved.