Attention Deficit HyperActivity Disorder or ADHD in short, has been a relatively common condition amongst the youth and old people alike all across the world. This condition, as the name suggests, is responsible for stunting a person’s attention span. People who are diagnosed with this disorder generally only have the option of learning to live with it as it has no concrete cure. But that being said, trying to understand it better by extracting relevant information about the brain’s functioning could potentially put us on the path to solving it. And according to a study published in Radiology: Artificial Intelligence, this might just be possible thanks to the power of Deep Learning.
As one can imagine, it is no easy task to study the inner workings of the brain as it such an important and delicate organ. Luckily, advances in MRI technology has opened up the possibility to take a look into the complex set of networks within the human brain. By generating images through the detection of changes in blood flow, MRIs can help train brain activity and map connections within and between brain networks. This detailed depiction of the connections within the brain is referred to as the connectome.
Since the connectome gives us a maplike layout of the brain, it makes it easier to study issues that are associated with the brain. So, this makes it the perfect tool to help understand the workings of a mind that is affected by ADHD.
However, the connectome only gives researchers a starting point for analysing ADHD affected brains. It is not a complete solution on its own. ADHD is not a disease that can be definitively diagnosed based on a single test or medical image analysis. It requires long term examinations wherein the candidate is tested for symptoms and is subjected to a series of behavior based tests.
Therefore, performing studies over longer periods while examining changes that happen within the connectome could potentially give us a clearer picture of ADHD. Moreover, the plausibility of this theory has been confirmed as well as research has suggested that ADHD results from a breakdown or malfunction in the connectome.
These connectomes are basically constructed from spatial regions across the MR image known as parcellations. Criteria such as the anatomy and functions are key towards defining these spatial regions. And by studying different brain parcellations, it is possible to uncover details about the workings of the brain at different scales.
Till date, studies have only been carried out on connectomes that were modeling from a single parcellation. But a new study conducted at the University of Cincinnati College of Medicine and Cincinnati Children’s Hospital Medical Center took it a step further. The researchers took advantage of the multiple underlying parcellations and developed a multiscale approach
By utilizing this multi scale approach the researchers went on to develop a deep learning model that could better predict and detect ADHD symptoms. This model incorporated data from the NeuroBereau dataset that included brain connectome data from 973 different participants along with relevant personal identifiers such as gender and iq.
This model has been viewed as a potential breakthrough in terms of Neuroscience. This is because, apart from providing essential data that can help diagnose and treat ADHD; through more research, its applications could be further extended to help with other neurological issues as well. Even recently, this technology delved into new territories and was used to predict cognitive deficiencies in preterm infants. The spectrum of possibilities this new imaging technology could offer seem to be wide ranging. So, with more exposure to new neuroimaging datasets and other developments, this deep learning model could essentially become one of the most useful tools within the realm of neuroscience.