BALTIMORE, Md. (Diya TV) — Johns Hopkins University biomedical engineering professor Sridevi Sarma has received a VAIBHAV Fellowship from India’s Department of Science and Technology to lead a three-year research collaboration focused on developing artificial intelligence tools for earlier autism screening.
The government selected Sarma’s proposal as one of 21 projects from 227 submissions to the 2025 VAIBHAV Fellowship call. Officials placed the project in the health, medical sciences and biomedical devices category.
Sarma will work with professors George Priya Doss C and Monica Subashini M at Vellore Institute of Technology in Vellore, Tamil Nadu. Doss, a professor in VIT’s School of Bio Sciences and Technology, will serve as the Indian principal investigator. Subashini, a professor whose research includes AI in health care, will serve as a co-principal investigator.
The project, titled “Cross-Modal Intelligence for Autism Screening: EEG, AI, and Genomic Integration,” will combine several types of biological and behavioral information. The researchers plan to study electrical brain activity recorded through electroencephalography, facial images and genomic data to identify patterns that could support autism screening.
An electroencephalogram, or EEG, records the brain’s electrical activity through electrodes placed on the scalp. Sarma’s research has previously used EEG data and computational modeling to develop tools for neurological conditions, including epilepsy.
The new collaboration will examine whether machine-learning systems can find consistent relationships across EEG recordings, facial information and genetic profiles. The researchers aim to develop measurable indicators that could help identify children who should receive a more detailed developmental evaluation.
Such indicators would serve as screening tools rather than independent diagnoses. Health care providers currently diagnose autism by assessing a person’s behavior and development. The Centers for Disease Control and Prevention says no blood test or other single medical test can diagnose autism.
The CDC recommends autism-specific screening during routine pediatric visits at 18 and 24 months. A child who produces a positive screening result must still receive a comprehensive evaluation from qualified professionals.
The research team intends to explore whether a multimodal system could make screening more accessible in communities with limited access to developmental specialists. Combining several data sources may also help researchers determine whether patterns found through one method remain consistent when compared with other biological or behavioral measurements.
The project remains in the research stage. The institutions have not released information about the number or ages of participants, the sources of the training data or the standards the researchers will use to test the system’s performance. Any tool developed through the project would require clinical validation before health care providers could use it in routine screening.
Sarma serves as vice dean for graduate education at the Johns Hopkins Whiting School of Engineering and associate director of the university’s Institute for Computational Medicine. She also founded and directs the Neuromedical Control Systems Lab, which develops computational methods intended for clinical use.
Her work applies dynamical systems, control theory and data analysis to disorders of the nervous system. Her previous projects include computational tools that analyze EEG recordings to support the diagnosis and treatment of epilepsy.
Sarma also serves on the faculty steering committee of the Gupta-Klinsky India Institute at Johns Hopkins, which supports research and education partnerships between Johns Hopkins and institutions in India.
The Indian government established the VAIBHAV program to connect scientists from the Indian diaspora with universities and publicly funded research institutions in India. Fellows may spend up to two months each year at their partner institution during a project lasting as long as three years.
The fellowship will support the Johns Hopkins-VIT collaboration for three years. Neither institution has announced when participant recruitment or data collection will begin.
