# Developing a blood fatty acid-based algorithm as an early predictor of cognitive decline and dementia: Applying machine learning to harmonized data from prospective cohort studies

> **NIH NIH R41** · OMEGAQUANT ANALYTICS, LLC · 2024 · $506,500

## Abstract

Alzheimer’s disease (AD), the most common type of dementia, imposes a substantial, global, socioeconomic
burden. An estimated 6.5 million Americans aged 65 and older are living with AD, the most prevalent form of
dementia. In the US, estimated health-care payments in 2022 for all patients with AD or related dementias
(ADRD) amount to $345 billion. Without the means to identify high risk individuals, many will find care too little
and too late: more than half of individuals with dementia or cognitive decline have not been diagnosed. There
is a need for accessible and inexpensive early predictive biomarkers of memory loss and/or incident ADRDs to
facilitate the early identification of high-risk individuals, providing the time necessary to make meaningful
lifestyle changes to slow or prevent disease progression. This is especially important since markers like tau or
beta-amyloid are primarily markers of existing, not impending disease. Emerging evidence suggests that
erythrocyte (RBC) omega-3 fatty acid (FA) levels may serve as an early signal of impending disease up to 5
years before AD/ADRD develops. As a clinical laboratory that specializes in providing FA measurements,
interpretation and customized behavioral interventions, OmegaQuant Analytics (OQA) supports a large and
growing customer base of researchers, clinicians, businesses, and individuals. Through a partnership with the
Fatty Acid Research Institute (FA expertise; biostatistical support; data access), we propose to develop a
highly predictive FA-based profile using an innovative approach leveraging existing prospective cohort data. To
do this, we will determine the extent to which it is possible to predict memory loss and/or incident all-cause
dementia from an RBC FA signature. We will harmonize data from 19,922 individuals with assessment of
incident all-cause dementia or an assessment of memory (e.g., Wechsler Memory Scale), with complete FA
profile data and with an average of 10+ years of follow-up. We will then apply statistical / machine learning
algorithms to determine the extent to which we can predict incident ADRD or a change in memory from FAs,
with separate models for high-risk subgroups, including racial/ethnic groups [Blacks, Hispanics]. Results will be
used to create a Fatty Acid Memory Index (FAMI) and Fatty Acid Dementia Index (FADI). We will create
consumer-friendly interpretative reports for FAMI and FADI including actionable steps to change dietary FA
behaviors to potentially modify memory loss/ dementia risk. We will determine if other clinical laboratories or
clinicians are willing to pay at least $30/test (wholesale price) for each test [Profitability pathway #1 (PP#1)].
We will also determine individual consumers’ willingness to pay OQA directly $50/test (retail price per test) for
either FAMI or FADI (PP#2). Proof of concept feasibility will set us up for a larger-scale prospective study and
improved machine-learning/modelling in Phase II. Ultimately, we hope to ...

## Key facts

- **NIH application ID:** 10820645
- **Project number:** 1R41AG085816-01
- **Recipient organization:** OMEGAQUANT ANALYTICS, LLC
- **Principal Investigator:** Bill Harris
- **Activity code:** R41 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2024
- **Award amount:** $506,500
- **Award type:** 1
- **Project period:** 2023-12-01 → 2025-11-30

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10820645

## Citation

> US National Institutes of Health, RePORTER application 10820645, Developing a blood fatty acid-based algorithm as an early predictor of cognitive decline and dementia: Applying machine learning to harmonized data from prospective cohort studies (1R41AG085816-01). Retrieved via AI Analytics 2026-05-22 from https://api.ai-analytics.org/grant/nih/10820645. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*