Muscat – A groundbreaking study conducted at AI Research Centre of National University of Science and Technology offers promising advancements in early detection of schizophrenia using artificial intelligence and brain MRI scans.

Led by Dr K Suresh Manic, Associate Professor at the university’s College of Engineering, the research demonstrates how AI can improve diagnostic accuracy and potentially transform clinical practices for mental health disorders.

Manic’s team developed a novel framework that integrates deep learning features with handcrafted data, refined through the Mayfly Algorithm, to detect schizophrenia in MRI slices. “Our goal is to enhance the reliability of automatic schizophrenia detection addressing the critical need for early diagnosis,” Dr Manic stated.

The study found that combining deep features (DF) with handcrafted features (HF) yielded an accuracy rate exceeding 95%. By comparison, DF alone achieved over 91% accuracy, while HF alone reached over 85%. This increase in precision underscores the potential for real-world clinical applications, particularly for identifying schizophrenia in its early stages and improving treatment outcomes.

Manic highlighted the need for further validation of the framework through larger datasets and clinical trials before its integration into routine clinical practice. “This success shows how AI can bridge critical gaps in mental healthcare, but the next steps will involve rigorous testing in clinical settings,” he added.

Titled ‘Framework to Detect Schizophrenia in Brain MRI Slices with Mayfly Algorithm-Selected Deep and Handcrafted Features’, the study was published in the peer-reviewed journal Sensors, with an impact factor of 3.4 and a CiteScore of 7.3.

The research team included experts from Oman, India, Australia and Norway, reflecting a collaborative international effort.

The project was supported by Ministry of Higher Education, Research and Innovation’s Publication Support Programme, which aims to promote high-quality research in Oman.

With potential to revolutionise mental health diagnostics, the study signals a significant step forward in addressing the challenges posed by schizophrenia in clinical and research settings.