A deep learning artificial intelligence model developed by researchers from Cincinnati Children's Hospital Medical Center enhances the power of MRI readings to predict attention deficit hyperactivity disorder in children, a new study suggests.
The study, published in Radiology: Artificial Intelligence, described the construction of an AI model built on nearly 1,000 participants' brain connectomes — essentially, comprehensive maps of the connections within their brains — as well as relevant data such as gender and IQ. In contrast to previous, more limited studies, the researchers' multiscale approach was significantly more successful in predicting ADHD: The model achieved nearly 80 percent accuracy.
While ADHD is diagnosed based on a variety of symptoms and behavioral exams, the study's authors said their findings show the potential of brain MRI scans in early diagnosis and interventions for children with ADHD. Additionally, they noted, the model could potentially be generalized to predict other neurological deficiencies in young children.
More articles on AI:
Physicians express doubts about Babylon Health's symptom-checking AI
Mount Sinai spinout partners with AI voice assistant Suki to tackle physician burnout
Harvard Medical School, Johns Hopkins leaders join advisory board of AI startup