Project Title: Proposing Health Indicators for Modelling and Prediction of Selected Biological Aspects in Human Communities
Funding Agency: IOE FRP Delhi University, Delhi Amount: 5L Year: 2024-25
Sanction Letter No and Date: IoE/2024-25/12/FRP dated 30.08.2024
Objectives of the Proposal
- Develop AI-driven frameworks to analyze healthcare utilization trends and predict physiological disorders using time-series and spectral analyses.
- Enhance emotion and disorder recognition by applying graph-based models on culturally specific fMRI and EEG datasets.
- Optimize predictive models for diabetes risk assessment using advanced feature selection techniques such as Cheetah, Polar Bear, and Royal Animal Optimization, supporting scalable, community-level health interventions.
Innovations / Outcomes
A key innovation is the development of the first culturally specific fMRI dataset (DOI: 10.18112/openneuro.ds005700.v1.0.2) for emotion recognition in the Indian population. DenseGCN and GCNN frameworks achieved high accuracies in autism and emotion detection using EEG and fMRI data, revealing brain-region-specific dynamics. The introduction of advanced optimization methods significantly improved diabetes diagnosis performance. The integration of Explainable AI techniques enhanced model transparency. These innovations collectively offer scalable, culturally adaptive AI healthcare tools with strong clinical and public health applications.
Publication:
- Abgeena A, Garg S, Goyal N, Raj JP. NeuroEmo: A neuroimaging-based fMRI dataset to extract temporal affective brain dynamics for Indian movie video clips stimuli using dynamic functional connectivity approach with graph convolution neural network (DFC-GCNN). accepted to publish in Computers in Biology and Medicine, IF 7.0 Elsevier E-ISSN 0010-4825.
- Pawan Kumar Patidar, Savita Shiwani, Shruti Garg, “Exploring the Potential of Royal Animal Optimization Algorithms for Diabetes prediction in Indian population data”, Journal of Information Systems Engineering and Management, Vol. 10 No. 12 Feb 2025: pp. 1-15 [Scopus] E-ISSN:2468-4376 https://doi.org/10.52783/jisem.v10i12s.1951
- Neha Prerna Tigga, Shruti Garg, Nishant Goyal, Justin Raj, Basudeb Raj, “Brain-region specific autism prediction from electroencephalogram signals using graph convolution neural network” Technology and Health Care, 01 Jan 2025 Vol. 33(1), pp. 77 – 101. IOS press IF 1.6 [SCIE, Q4] [Web of Science] E-ISSN 1878-7401, DOI: 10.3233/THC-240550.