Disease Research PlaybookAlzheimer's & DementiafMRI

Alzheimer's Disease (AD) Open Data Guide: Integrating Neuroimaging (BIDS), Plasma Biomarkers & Multi-Omics

A practical methodology for combining amyloid/tau PET, structural MRI, plasma p-tau217, and single-nucleus transcriptomics across open repositories.

CNS Open Data Editorial & Neuroinformatics TeamUpdated: 2026-10-1512 min read

Executive Summary

Modern Alzheimer's disease research has shifted from purely syndromic clinical definitions toward a biological staging framework defined by Amyloid (A), Tau (T), and Neurodegeneration (N), alongside Neuroinflammation (I) and Vascular (V) co-pathologies. This guide explains how researchers can navigate open repositories—specifically OpenNeuro (BIDS neuroimaging), Agora AMP-AD (multi-omics target validation), and Alzforum (therapeutic & biomarker pipelines)—to design reproducible cross-cohort studies.

SECTION 01

1. Mapping Open Repositories to the Biological ATN(IV) Framework

No single repository captures the entire continuum of Alzheimer's disease from preclinical amyloid accumulation to late-stage neurofibrillary tangle propagation and synaptic loss. To build robust computational models, investigators must understand the complementary strengths of each open archive. OpenNeuro hosts BIDS-standardized structural T1w MRI, resting-state fMRI, and amyloid/tau PET datasets under CC0 public domain terms, ideal for macroscale network atrophy and Centiloid standardization. Meanwhile, Agora (Accelerating Medicines Partnership in AD) aggregates bulk and single-nucleus transcriptomics, proteomics, and metabolomics from dorsolateral prefrontal cortex (DLPFC), parahippocampal gyrus, and temporal cortex cohorts (ROSMAP, Mayo Clinic, Mount Sinai Brain Bank).

Key Practical Takeaways

Use OpenNeuro BIDS datasets for in vivo volumetric MRI, cortical thickness (FreeSurfer), and functional connectomics.
Use Agora (AMP-AD) and SEA-AD for post-mortem cell-type-specific gene expression, Biodomain enrichment, and Target Risk Scores.
Use Alzforum Therapeutics & AlzBiomarker to benchmark clinical trial endpoints (CDR-SB, plasma p-tau217, ARIA rates).

SECTION 02

2. Controlling for APOE ε4, Age, Sex, and Vascular Co-Pathologies

A frequent cause of non-replicable findings in secondary Alzheimer's analyses is inadequate control of demographic and genetic confounders. APOE ε4 dosage (0, 1, or 2 alleles) profoundly shifts the age of amyloid onset by up to 10–15 years and modifies microglial lipid metabolism and blood-brain barrier integrity. Furthermore, women exhibit higher tau tangle burden per unit of amyloid pathology in temporoparietal cortices. In post-mortem omics studies, Post-Mortem Interval (PMI), RNA Integrity Number (RIN), and brain bank batch effects must be explicitly modeled in linear mixed-effects regressions.

SECTION 03

3. Bridging Macroscale Neuroimaging with Single-Cell Vulnerability

One of the most impactful secondary research paradigms combines spatial maps of cortical atrophy or tau PET uptake with spatial gene expression atlases (Allen Human Brain Atlas / SEA-AD MERFISH). By parcellating both neuroimaging SUVR volumes and transcriptomic expression matrices into a common anatomical atlas (such as the Desikan-Killiany 68-region or Schaefer 400-region parcellation), researchers can test whether regional tau propagation follows intrinsic functional connectivity gradients or regional enrichment of specific vulnerable neuronal subpopulations (e.g., RORB+ excitatory neurons).

SECTION 04

4. Translational Alignment: Anti-Amyloid Therapies & Biomarker Stratification

With the regulatory approval of anti-amyloid monoclonal antibodies (Lecanemab, Donanemab), translational research increasingly focuses on identifying non-responders, predicting Amyloid-Related Imaging Abnormalities (ARIA-E edema and ARIA-H microhemorrhages), and discovering combination targets for tau and neuroinflammation (TREM2, MS4A, NLRP3).

Python Example: Stratified Linear Modeling of Cortical Thickness / Biomarkers by APOE Genotype

Standard Python template ready for integration into local Jupyter notebooks or automated pipelines.

import pandas as pd
import statsmodels.formula.api as smf

# Load BIDS participants.tsv merged with FreeSurfer cortical thickness & plasma p-tau217
df = pd.read_csv("./data/ad_cohort_harmonized.tsv", sep="\t")

# Encode APOE e4 allele dosage (0, 1, or 2 copies) and standardize continuous covariates
df["apoe4_dosage"] = df["apoe_genotype"].map({"e3/e3": 0, "e2/e3": 0, "e3/e4": 1, "e4/e4": 2})
df["age_z"] = (df["age"] - df["age"].mean()) / df["age"].std()

# Fit Ordinary Least Squares / Mixed-Effects model controlling for age, sex, and scanner site
model = smf.ols(
    formula="entorhinal_thickness_mm ~ ptau217_pg_ml * C(apoe4_dosage) + age_z + C(sex) + C(site_id)",
    data=df.dropna(subset=["entorhinal_thickness_mm", "ptau217_pg_ml", "apoe4_dosage"])
).fit()

print(model.summary())

Research Governance & Reproducibility Checklist

Stratify or adjust all statistical models for APOE ε4 allele count, biological sex, and chronological age.
Harmonize multi-site MRI/PET measurements using ComBat or neuroHarmonize before pooling cohorts.
Verify whether post-mortem transcriptomic signatures are driven by shifts in cell-type proportions (neuronal loss / gliosis) vs. intrinsic per-cell regulation.

Key Biomarkers & Standards Referenced in This Guide

View Full Glossary

APOE

Apolipoprotein E (APOE ε2 / ε3 / ε4)

Major lipid carrier in the central nervous system; the APOE ε4 allele is the strongest genetic risk factor for late-onset Alzheimer's disease.

Read entry

p-tau217

Phosphorylated Tau at Threonine 217 (Plasma / CSF p-tau217)

Highly sensitive fluid biomarker reflecting both cerebral amyloid-beta plaque burden and early tau tangle phosphorylation in Alzheimer's disease.

Read entry

PET

Positron Emission Tomography (Amyloid / Tau / FDG / SV2A PET)

Molecular neuroimaging technique using radiolabeled tracers to quantify in vivo protein pathology (amyloid plaques, tau tangles), glucose metabolism, or synaptic density.

Read entry

fMRI

Functional Magnetic Resonance Imaging (BOLD fMRI)

Non-invasive neuroimaging modality measuring blood-oxygen-level-dependent (BOLD) hemodynamic signals to map brain network activity and functional connectivity.

Read entry

scRNA-seq

Single-Cell & Single-Nucleus RNA Sequencing (scRNA-seq / snRNA-seq)

High-throughput transcriptomic profiling at individual cell or nucleus resolution, enabling identification of disease-associated microglia (DAM), astrocytes, and vulnerable neuronal subtypes.

The neural mechanisms underlying spontaneous perceptual switches during bistable perception, particularly the dynamic interplay between large-scale brain networks, remain incompletely understood. Here, using functional magnetic resonance imaging (fMRI) in six participants and a bistable kinetic dept...

Abstract / Summary

Preprint ArticleOpen Science#bioRxiv Preprint#neuroscience#ALS

Oct 1, 2026

Ponzi, A. +2

Visit Source DOI Data

View Details

More Research Playbooks & Tutorials

Disease Research Playbook

Parkinson's Disease (PD) Open Data Guide: Alpha-Synuclein Assays, Basal Ganglia LFP & Cohort Harmonization

11 min read

Disease Research Playbook

ALS & Motor Neuron Disease Open Data Guide: TDP-43 Cryptic Exons, NfL Trajectories & ENCALS Models

11 min read

Disease Research Playbook

Multiple Sclerosis (MS) Open Data Guide: Federated Real-World Data (OMOP CDM) & MRI Lesion Analysis

10 min read

Back to Research Playbooks & Guides

Disease Research PlaybookAlzheimer's & DementiafMRI

Alzheimer's Disease (AD) Open Data Guide: Integrating Neuroimaging (BIDS), Plasma Biomarkers & Multi-Omics

A practical methodology for combining amyloid/tau PET, structural MRI, plasma p-tau217, and single-nucleus transcriptomics across open repositories.

CNS Open Data Editorial & Neuroinformatics TeamUpdated: 2026-10-1512 min read

Executive Summary

SECTION 01

1. Mapping Open Repositories to the Biological ATN(IV) Framework

Key Practical Takeaways

Use OpenNeuro BIDS datasets for in vivo volumetric MRI, cortical thickness (FreeSurfer), and functional connectomics.
Use Agora (AMP-AD) and SEA-AD for post-mortem cell-type-specific gene expression, Biodomain enrichment, and Target Risk Scores.
Use Alzforum Therapeutics & AlzBiomarker to benchmark clinical trial endpoints (CDR-SB, plasma p-tau217, ARIA rates).

SECTION 02

2. Controlling for APOE ε4, Age, Sex, and Vascular Co-Pathologies

SECTION 03

3. Bridging Macroscale Neuroimaging with Single-Cell Vulnerability

SECTION 04

4. Translational Alignment: Anti-Amyloid Therapies & Biomarker Stratification

Python Example: Stratified Linear Modeling of Cortical Thickness / Biomarkers by APOE Genotype

Standard Python template ready for integration into local Jupyter notebooks or automated pipelines.

import pandas as pd
import statsmodels.formula.api as smf

# Load BIDS participants.tsv merged with FreeSurfer cortical thickness & plasma p-tau217
df = pd.read_csv("./data/ad_cohort_harmonized.tsv", sep="\t")

# Encode APOE e4 allele dosage (0, 1, or 2 copies) and standardize continuous covariates
df["apoe4_dosage"] = df["apoe_genotype"].map({"e3/e3": 0, "e2/e3": 0, "e3/e4": 1, "e4/e4": 2})
df["age_z"] = (df["age"] - df["age"].mean()) / df["age"].std()

# Fit Ordinary Least Squares / Mixed-Effects model controlling for age, sex, and scanner site
model = smf.ols(
    formula="entorhinal_thickness_mm ~ ptau217_pg_ml * C(apoe4_dosage) + age_z + C(sex) + C(site_id)",
    data=df.dropna(subset=["entorhinal_thickness_mm", "ptau217_pg_ml", "apoe4_dosage"])
).fit()

print(model.summary())

Research Governance & Reproducibility Checklist

Stratify or adjust all statistical models for APOE ε4 allele count, biological sex, and chronological age.
Harmonize multi-site MRI/PET measurements using ComBat or neuroHarmonize before pooling cohorts.
Verify whether post-mortem transcriptomic signatures are driven by shifts in cell-type proportions (neuronal loss / gliosis) vs. intrinsic per-cell regulation.

Key Biomarkers & Standards Referenced in This Guide

View Full Glossary

APOE

Apolipoprotein E (APOE ε2 / ε3 / ε4)

Major lipid carrier in the central nervous system; the APOE ε4 allele is the strongest genetic risk factor for late-onset Alzheimer's disease.

Read entry

p-tau217

Phosphorylated Tau at Threonine 217 (Plasma / CSF p-tau217)

Highly sensitive fluid biomarker reflecting both cerebral amyloid-beta plaque burden and early tau tangle phosphorylation in Alzheimer's disease.

Read entry

PET

Positron Emission Tomography (Amyloid / Tau / FDG / SV2A PET)

Molecular neuroimaging technique using radiolabeled tracers to quantify in vivo protein pathology (amyloid plaques, tau tangles), glucose metabolism, or synaptic density.

Read entry

fMRI

Functional Magnetic Resonance Imaging (BOLD fMRI)

Non-invasive neuroimaging modality measuring blood-oxygen-level-dependent (BOLD) hemodynamic signals to map brain network activity and functional connectivity.

Read entry

scRNA-seq

Single-Cell & Single-Nucleus RNA Sequencing (scRNA-seq / snRNA-seq)

High-throughput transcriptomic profiling at individual cell or nucleus resolution, enabling identification of disease-associated microglia (DAM), astrocytes, and vulnerable neuronal subtypes.

Read entry

CDR-SB

Clinical Dementia Rating — Sum of Boxes

Primary composite clinical endpoint in early Alzheimer's disease trials, summing 6 cognitive and functional domains (Memory, Orientation, Judgment, Community, Home, Personal Care; range 0–18).

Read entry