The brain is an intricate network of neurons. Recently the researchers have found that a deep learning approach can be applied in the treatment of other neurological conditions as well.
FREMONT, CA: According to a study, deep learning, a kind of AI, can accelerate magnetic resonance imaging’s (MRI) power in order to predict attention deficit hyperactivity disorder (ADHD). The approach is also beneficial for other neurological conditions.
Advancements in functional MRI have been very helpful in the mapping of connections within and between brain networks. The detailed brain map is also referred to as the connectome.
Progressively, the connectome is considered as the key to understanding brain disorders such as ADHD.
According to researchers, brain MRI has a potential role to play in treatment as the root cause of ADHD is some kind of breakdown or disruption in the connectome. The connectome is built from the spatial areas across the MR image called parcellations. These brain parcellations are defined on the basis of anatomical criteria, functional criteria, or both. The study of the brain can be done on different scales based on different brains.
Earlier, the studies emphasized on presumed single-scale approach in which the connectome is constructed on the basis of only one parcellation. In the latest study, researchers from the Cincinnati Children’s Hospital Medical Center and the University Of Cincinnati College Of Medicine did comprehensive research. A multi-scale technique was developed which utilized various connectome maps based on various parcellations.
The researchers took data from the NeuroBureau ADHD-200 dataset for building the deep learning model. The model made use of the multi-scale brain connectome data from the 973 participants of the project with relevant personal characteristics like gender and IQ.
Through the multi-scale approach, the researchers noticed a considerable improvement in ADHD detection performance.
By enhancing diagnostic accuracy, deep-learning-aided MRI-based treatment can be crucial in applying early interventions for ADHD patients.
In addition, this model can be useful for other neurological deficiencies as well. The scientists are already utilizing it for predicting cognitive deficiency in preterm infants. They scan them as soon as possible after their birth for predicting neurodevelopmental outcomes at the age of two years.
The researchers are anticipating seeing the deep learning model enhance in the future as it is being exposed to larger neuroimaging datasets. They are also hoping to better comprehend the precise breakdowns and disruptions in the connectome detected by the model that are related to ADHD.