AI emerges as an effective tool for Autism Spectrum Disorder detection, says report

A recent piece of study published in the journal 'Scientific Reports' emphasises the potential of artificial intelligence (AI) in the diagnosis of Autism Spectrum Disorder (ASD). The complex nature of ASD creates diagnostic issues that necessitate highly trained physicians. The study looked at brain imaging data from 500 people, including 242 with ASD.

The researchers used machine learning techniques to design a new methodology using this data. The research was headed by Francisco Rodrigues, a professor at the University of Sao Paulo's Institute of Mathematics and Computer Science.
 

What is Autism Spectrum Disorder?

Autism Spectrum Disorder (ASD) is a condition that affects how people think, feel, and interact with others. It is called a ‘spectrum’ because it varies from person to person. People with ASD might find it harder to communicate, understand social cues, or express their feelings. They might have certain repetitive behaviours or interests.

Some individuals with ASD might need extra support in daily life, while others might be very independent. It's important to remember that every person with ASD is unique, and they can have different strengths and challenges.

How AI can detect ASD?

The research introduced an innovative approach by focusing on brain network organisation through cortical networks, or brain maps. These maps illuminate connections among brain regions, providing insights into brain alterations associated with ASD. The study found that ASD patients exhibited distinct alterations in cognitive, emotional, learning, and memory-related brain regions.

Neuroscientist’s approach to ASD diagnosis

Francisco Rodrigues explained that they first developed their methods by gathering functional magnetic resonance imaging (fMRI) and electroencephalogram (EEG) data. His participation in the study was funded by FAPESP, and he is a professor at the Institute of Mathematics and Computer Science (ICMC-USP) of the University of So Paulo in So Carlos, Brazil.

"We compared the maps of people with and without ASD and found that diagnosis was possible using this methodology," Rodrigues stated.

He further added, These maps were supplied to a machine learning algorithm by the researchers. With greater than 95 percent accuracy, the algorithm was able to identify which brain abnormalities were connected to ASD based on the taught instances.

Opening doors of hope for ASD patients

Though the methodology is still in its developmental stage, it holds significant promise. It could potentially aid specialists in cases of diagnostic uncertainty.

While further research is needed, the study's findings offer new avenues for understanding ASD's link to brain alterations. The interdisciplinary effort involved physicists, statisticians, physicians, and neuroscientists from various centres, underscoring the complexity of the subject and its potential impact on medical understanding and treatment.