AID-AD

About Our Group

Alzheimer’s disease and AD-related dementias (AD/ADRD) are complex multifactorial conditions affecting millions of people around the world. The number of people with AD dementia, prodromal AD, and preclinical AD is estimated to exceed 400 million worldwide. ADRD such as Lewy body dementia, Limbic-predominant Age-related TDP-43 Encephalopathy (LATE), and vascular cognitive impairment/post-stroke dementia substantially increase the prevalence and economic burden. In the U.S., the costs to all payers for the care of people living with AD and other dementias will total an estimated $384 billion in 2025. Clinically, signs and symptoms of AD include cognitive impairments, especially memory, as well as emotional disturbances, though clinical presentations of AD vary substantially. Age and family history, and APOE4 genotype, are among the known major risk factors. Therapeutic options are limited, with recent trials targeting AD’s hallmark proteinopathies delivering only modest average slowing of clinical decline, so progression typically continues at roughly 65–80% of the placebo rate.

​

A wealth of AD/ADRD datasets and resources now support multimodal research across genetics, imaging, biomarkers, clinical cohorts, and biospecimens. AI is transforming our lives and can have a significant impact on AD/ADRD research and care by enabling the integration and analysis of complex, multimodal data spanning genetics, imaging, proteomics, clinical assessments, and digital health measures. AI methods are being applied to improve early and accurate diagnosis, predict disease onset and progression, and uncover novel biomarkers and molecular subtypes that capture the heterogeneity of AD/ADRD; however, most of them have been applied to a single modality at a time.

AID-AD: Artificial Intelligence for Discovery, Diagnosis & Delivery of Care in Alzheimer’s Disease and Related Dementias will serve as a national hub to advance AI methods, including multimodal approaches, and their applications in aging and AD/ADRD with the significant goal of democratizing AI-ready datasets and AI models and connecting researchers to enable innovation by bridging together the fields of AI and AD/ADRD discovery science. AID-AD proposes a comprehensive strategy for multimodal AI in AD/ADRD that unites imaging, genomics, electronic health records, and other clinical and digital health data into integrated, scalable tools. Multimodal AI is defined to include both single-modality models applied across diverse biomedical data types and integrative models that explicitly fuse heterogeneous modalities. Unlike prior approaches that analyze modalities in isolation, our vision emphasizes the novel power of cross-modal integration, enabling the discovery of hidden disease signatures, more accurate diagnosis, and personalized therapeutic strategies. Looking ahead, we recognize that true innovation will arise from combining the strengths of human intelligence clinicians, caregivers, and patients with the computational capabilities of emerging technologies which promise to expand the limits of pattern recognition and predictive modeling.

​​​