1. Home
  2. Cardiovascular Health and Artificial Intelligence (CHAI)
  3. HEAL - Health Enhancement through AI and Location
  4. HEAL Research

HEAL Research

Research Focus Areas

Our research integrates geospatial concepts, advanced data processing, harmonization, analysis, and machine learning, centering on several key themes:

  1. AI-Enhanced Environmental Risk Assessment: 
    AI-based spatial health analysis graphic from HEAL (Health Enhancement through AI and Location) research uses satellite images and deep learning to predict how common diseases like heart disease and stroke are across neighborhoods. A scatter plot shows strong alignment between predicted and actual rates. City maps compare observed and estimated disease levels. Highlighted image areas show what influenced the AI's predictions鈥攔oads (linked to higher disease) and green spaces (linked to lower disease).
    Utilizing deep learning and computer vision techniques on satellite and street view imagery to extract granular features of the built environment (e.g., road quality, green space, building conditions) and quantify their association with cardiovascular disease risk and health disparities across diverse geographic settings.
  2. Geographic and Environmental Determinants of Health: Investigating how dynamic environmental exposures (like air pollution, neighborhood walkability), mobility patterns, and healthcare accessibility interact to shape cardiovascular health outcomes and contribute to health inequities. This includes analyzing factors like proximity to clinical trial sites and the environmental footprint of healthcare delivery.
  3. Foundation Models and AI for Integrated Spatial Health Intelligence: We are actively exploring and adapting cutting-edge Artificial Intelligence, including Large Language Models (LLMs) and multimodal Foundation Models, to create a deeper, more integrated understanding of spatial health dynamics.

Research Objectives

The core objectives of HEAL are to:

  • Develop and validate innovative AI and machine learning models using diverse geospatial data sources (satellite, street view, drone and sensor data) for precision environmental exposure assessment relevant to health.
  • Quantify the impact of built, natural, and social environments on cardiometabolic health and health disparities using advanced spatial and statistical methods.
  • Identify cardiometabolic health risk and resilience in geographic areas and populations.
  • Analyze how dynamic factors like environmental change, mobility, and access to resources influence population health.
  • Generate actionable, evidence-based insights to inform public health strategies, clinical decision-making, urban planning policies, and resource allocation for risk mitigation, health promotion and healing.