Concept · neuroscience
Alzheimer’s and MCI markers
Follow Alzheimer’s and MCI markers — see important new research and changes in evidence.Change log
What changed
Dated edits to this page's evidence: studies added or removed from a claim, claims added or withdrawn, and new explanations tagged here. Rewordings are not listed.
Integrating transcriptomic/proteomic/metabolomic/lipidomic brain data yielded four multimodal profiles, including Knight-C4 with pronounced dysregulation and worse clinical features.
- New claim
- Added a supporting study: Are there molecular subtypes of Alzheimer brains?
- Added a scope qualifier: How safe is clinic lecanemab in practice?
A specialty memory-clinic report on lecanemab treatment adds practice-level evidence beside biomarker/pathology marker papers.
- New claim
- Added a supporting study: How safe is clinic lecanemab in practice?
- Added a scope qualifier: Are there molecular subtypes of Alzheimer brains?
- Concept page published
Alzheimer’s disease and mild cognitive impairment are tracked with overlapping but not identical markers—blood proteins that forecast later dementia, CSF/PET tau that tracks early brain change, and behavioural tasks that separate biomarker-defined MCI from controls. This page states which claim each study supports.
Students meet “Alzheimer’s biomarkers” as one story. In practice, a plasma prognostic marker, a PET decline predictor, a connectome pattern, and a VR navigation task answer different questions. Mixing them invents false agreement.
Evidence
What the evidence shows
Drawn from 7 studies in this library. Each finding starts with a plain-language takeaway, then the denser detail. Supports means evidence for a finding; Challenges means evidence against a stated position; Qualifies marks scope with a short note on each study’s contribution. Challenged positions are labeled — they are not findings.
Blood GFAP and NfL associate with higher later dementia risk years before diagnosis.
In UK Biobank adults with baseline plasma measures, higher GFAP and NfL associated with roughly doubled hazards of incident all-cause dementia over ~13 years, with elevations detectable up to about 15 years before diagnosis and incremental prediction beyond demographic risk scores.
- Using Brain Scans to Predict Cognitive Decline in Alzheimer's— different question — short-horizon decline in diagnosed AD, not biobank prognosis
- Does brain tau protein affect object memory in older adults?— different compartment — CSF tau + fMRI, not plasma GFAP/NfL
Study Role Design N Population Outcome Blood GFAP and NfL predict dementia years ahead Supports CohortUK Biobank prospective plasma GFAP/NfL with registry dementia follow-up N=48542 · 1,312 incident all-cause dementia over ~13 years UK Biobank adults with baseline Olink plasma GFAP and NfL Incident all-cause dementia / ADRD HRs and prediction AUC for GFAP and NfL Using Brain Scans to Predict Cognitive Decline in Alzheimer's Qualifiesdifferent question — short-horizon decline in diagnosed AD, not biobank prognosis CohortBaseline MRI/PET (tau, microglial activation, atrophy) predicting 3-year cognitive decline N=55 · 26 Alzheimer’s-pathology patients (12 dementia, 14 amnestic MCI) and 29 healthy controls Patients with Alzheimer’s pathology and healthy controls with longitudinal cognition Baseline tau and microglial activation predicting decline; atrophy adding no independent value Does brain tau protein affect object memory in older adults? Qualifiesdifferent compartment — CSF tau + fMRI, not plasma GFAP/NfL OtherObservational CSF biomarker–fMRI association study; no assigned exposure N=21 · 21 cognitively unimpaired older adults analysed after spinal tap and mnemonic discrimination fMRI Cognitively unimpaired older adults Object mnemonic discrimination and hippocampal activation vs CSF tau In people with Alzheimer’s, tau and microglial PET can flag who declines faster.
In longitudinal AD imaging, baseline tau PET burden and microglial activation predicted subsequent cognitive decline; when markers were modelled together, posterior cortical tau and inflammation carried prediction while atrophy added little incremental signal.
- Blood GFAP and NfL predict dementia years ahead— different population — volunteer biobank prognosis, not clinic AD decline
Study Role Design N Population Outcome Using Brain Scans to Predict Cognitive Decline in Alzheimer's Supports CohortBaseline MRI/PET (tau, microglial activation, atrophy) predicting 3-year cognitive decline N=55 · 26 Alzheimer’s-pathology patients (12 dementia, 14 amnestic MCI) and 29 healthy controls Patients with Alzheimer’s pathology and healthy controls with longitudinal cognition Baseline tau and microglial activation predicting decline; atrophy adding no independent value Blood GFAP and NfL predict dementia years ahead Qualifiesdifferent population — volunteer biobank prognosis, not clinic AD decline CohortUK Biobank prospective plasma GFAP/NfL with registry dementia follow-up N=48542 · 1,312 incident all-cause dementia over ~13 years UK Biobank adults with baseline Olink plasma GFAP and NfL Incident all-cause dementia / ADRD HRs and prediction AUC for GFAP and NfL Tau tends to sit in network hubs and tracks weaker cortical connectivity in AD.
In Alzheimer’s disease, tau PET burden concentrated in highly connected cortical hubs and higher overall tau associated with weakened cortical connection strength; progressive-supranuclear-palsy patterns differed, showing disease-specific mapping of tau onto networks.
- Using Brain Scans to Predict Cognitive Decline in Alzheimer's— related PET tau story, but decline prediction rather than connectome mapping
A VR navigation task can separate biomarker-positive MCI from controls.
An entorhinal-cortex–based virtual navigation task produced larger path errors in MCI than controls, with biomarker-positive MCI performing worse than biomarker-negative MCI—linking a behavioural assay to CSF Alzheimer’s markers.
- Does brain tau protein affect object memory in older adults?— different behavioural assay — object mnemonic discrimination, not VR path error
Study Role Design N Population Outcome Using Virtual Reality to Detect Early Alzheimer's Disease Supports Cross-sectionalImmersive VR navigation test comparing biomarker-defined MCI patients and healthy controls N=86 · 45 MCI patients (12 biomarker-positive, 14 biomarker-negative among CSF-tested) and 41 healthy controls Mild cognitive impairment patients and healthy controls performing entorhinal-based VR navigation Navigation error distinguishing biomarker-positive early Alzheimer’s MCI from biomarker-negative MCI and controls Does brain tau protein affect object memory in older adults? Qualifiesdifferent behavioural assay — object mnemonic discrimination, not VR path error OtherObservational CSF biomarker–fMRI association study; no assigned exposure N=21 · 21 cognitively unimpaired older adults analysed after spinal tap and mnemonic discrimination fMRI Cognitively unimpaired older adults Object mnemonic discrimination and hippocampal activation vs CSF tau Higher CSF tau can track early object-memory and hippocampal activity changes.
In cognitively unimpaired older adults, higher CSF tau correlated with worse object mnemonic discrimination and greater right hippocampal task fMRI activity—evidence that a fluid marker can track early functional change before a dementia diagnosis.
- Blood GFAP and NfL predict dementia years ahead— different scale — biobank plasma prognosis, not small fMRI sample
Study Role Design N Population Outcome Does brain tau protein affect object memory in older adults? Supports OtherObservational CSF biomarker–fMRI association study; no assigned exposure N=21 · 21 cognitively unimpaired older adults analysed after spinal tap and mnemonic discrimination fMRI Cognitively unimpaired older adults Object mnemonic discrimination and hippocampal activation vs CSF tau Blood GFAP and NfL predict dementia years ahead Qualifiesdifferent scale — biobank plasma prognosis, not small fMRI sample CohortUK Biobank prospective plasma GFAP/NfL with registry dementia follow-up N=48542 · 1,312 incident all-cause dementia over ~13 years UK Biobank adults with baseline Olink plasma GFAP and NfL Incident all-cause dementia / ADRD HRs and prediction AUC for GFAP and NfL Brain multi-omics reveals multimodal AD molecular profiles including a severe cluster.
Integrating transcriptomic/proteomic/metabolomic/lipidomic brain data yielded four multimodal profiles, including Knight-C4 with pronounced dysregulation and worse clinical features.
- How safe is clinic lecanemab in practice?— clinic anti-amyloid treatment experience — not a multi-omics subtype discovery
Study Role Design N Population Outcome Are there molecular subtypes of Alzheimer brains? Supports Computational / modellingMulti-cohort brain multi-omics ML clustering of AD molecular profiles N=42 · Knight-C4 high-dysregulation cluster n=42 within four multimodal profiles Human AD brain cohorts with multi-omic + clinical/neuropath data Multimodal molecular clusters linked to cognition and progression How safe is clinic lecanemab in practice? Qualifiesclinic anti-amyloid treatment experience — not a multi-omics subtype discovery CohortRetrospective consecutive case series of specialty-clinic lecanemab initiations N=234 · Patients with early symptomatic AD receiving ≥1 infusion Early symptomatic Alzheimer disease patients at Washington University Memory Diagnostic Center Infusion reactions, ARIA (including symptomatic), and treatment withdrawal Specialty clinics are delivering lecanemab with real-world workflow constraints.
A specialty memory-clinic report on lecanemab treatment adds practice-level evidence beside biomarker/pathology marker papers.
- Are there molecular subtypes of Alzheimer brains?— molecular subtypes ≠ treatment eligibility/outcomes
Study Role Design N Population Outcome How safe is clinic lecanemab in practice? Supports CohortRetrospective consecutive case series of specialty-clinic lecanemab initiations N=234 · Patients with early symptomatic AD receiving ≥1 infusion Early symptomatic Alzheimer disease patients at Washington University Memory Diagnostic Center Infusion reactions, ARIA (including symptomatic), and treatment withdrawal Are there molecular subtypes of Alzheimer brains? Qualifiesmolecular subtypes ≠ treatment eligibility/outcomes Computational / modellingMulti-cohort brain multi-omics ML clustering of AD molecular profiles N=42 · Knight-C4 high-dysregulation cluster n=42 within four multimodal profiles Human AD brain cohorts with multi-omic + clinical/neuropath data Multimodal molecular clusters linked to cognition and progression
Open questions
Tensions and limits
Some items are genuine disagreements on the same question. Others mark different assays, populations, or outcomes — limits on how far one study travels — not a forced fight between papers.
Plasma GFAP/NfL forecast registry dementia years later in a volunteer cohort; tau/inflammation PET predict short-horizon decline in people already diagnosed with AD. Both can be valid—the limit is not treating prognosis and clinic decline as one assay.
- Blood GFAP and NfL predict dementia years ahead
- Using Brain Scans to Predict Cognitive Decline in Alzheimer's
Study Role Design N Population Outcome Blood GFAP and NfL predict dementia years ahead Supports CohortUK Biobank prospective plasma GFAP/NfL with registry dementia follow-up N=48542 · 1,312 incident all-cause dementia over ~13 years UK Biobank adults with baseline Olink plasma GFAP and NfL Incident all-cause dementia / ADRD HRs and prediction AUC for GFAP and NfL Using Brain Scans to Predict Cognitive Decline in Alzheimer's Supports CohortBaseline MRI/PET (tau, microglial activation, atrophy) predicting 3-year cognitive decline N=55 · 26 Alzheimer’s-pathology patients (12 dementia, 14 amnestic MCI) and 29 healthy controls Patients with Alzheimer’s pathology and healthy controls with longitudinal cognition Baseline tau and microglial activation predicting decline; atrophy adding no independent value VR navigation separates biomarker-defined MCI behaviourally; CSF tau in unimpaired adults tracks hippocampal hyperactivity and object memory. Shared theme: early Alzheimer’s biology can show up before frank dementia—different tasks, samples, and endpoints.
- Using Virtual Reality to Detect Early Alzheimer's Disease
- Does brain tau protein affect object memory in older adults?
Study Role Design N Population Outcome Using Virtual Reality to Detect Early Alzheimer's Disease Supports Cross-sectionalImmersive VR navigation test comparing biomarker-defined MCI patients and healthy controls N=86 · 45 MCI patients (12 biomarker-positive, 14 biomarker-negative among CSF-tested) and 41 healthy controls Mild cognitive impairment patients and healthy controls performing entorhinal-based VR navigation Navigation error distinguishing biomarker-positive early Alzheimer’s MCI from biomarker-negative MCI and controls Does brain tau protein affect object memory in older adults? Supports OtherObservational CSF biomarker–fMRI association study; no assigned exposure N=21 · 21 cognitively unimpaired older adults analysed after spinal tap and mnemonic discrimination fMRI Cognitively unimpaired older adults Object mnemonic discrimination and hippocampal activation vs CSF tau
Common misconceptions
A raised blood NfL or GFAP level diagnoses Alzheimer’s disease on its own.
These proteins are nonspecific injury/inflammation signals that associate with later dementia risk in cohorts; they are not a same-day AD rule-in test and rise in other neurological conditions.
All MCI is early Alzheimer’s, so any MCI marker equals an AD diagnosis.
Biomarker-positive and biomarker-negative MCI can look different on the same VR task; MCI is a clinical stage that needs biomarker and clinical context, not a single label.
Exam-style questions
Short-answer questions that ask you to explain or compare, not recall.
Why can plasma NfL/GFAP prognosis and tau PET decline prediction both be true without being the same claim?
They answer different questions in different populations: biobank adults years before dementia versus people already diagnosed with AD over a few years. Shared biology (neurodegeneration signals) does not make the assays interchangeable.
A clinic wants one “Alzheimer’s accuracy number” pooling VR MCI discrimination with biobank AUCs. What goes wrong?
Those metrics come from different outcomes (task error vs incident dementia), prevalences, and time horizons. Pooling invents a head-to-head that the designs do not support.
The studies
7 studies in this library bear on Alzheimer’s and MCI markers, ordered by citations.
- How tau buildup shapes brain networks in dementia
In Alzheimer's disease, tau builds up in highly connected brain hubs and weakens their connections, while in progressive supranuclear palsy, tau accumulates in deep brain structures and forces the cortex to use less efficient, indirect pathways.
- Using Brain Scans to Predict Cognitive Decline in Alzheimer's
Baseline PET scans tracking tau protein and brain inflammation predict future cognitive decline over three years, while brain shrinkage measurements add no extra predictive power.
- Using Virtual Reality to Detect Early Alzheimer's Disease
A virtual reality walking test assessing entorhinal cortex function accurately identifies patients with early-stage, biomarker-proven Alzheimer's disease.
- Blood GFAP and NfL predict dementia years ahead
In 48,542 UK Biobank participants followed ~13 years, elevated plasma GFAP and NfL preceded dementia up to 15 years and improved prediction beyond CAIDE/DRS risk scores (AUC up to ~0.89).
- Does brain tau protein affect object memory in older adults?
Elevated levels of tau protein in the brain fluid of healthy older adults are linked to overactivity in the hippocampus and worse memory for distinguishing highly similar objects.
- How safe is clinic lecanemab in practice?
Among 234 memory-clinic patients on lecanemab, infusion reactions (37%) and ARIA (22%) were common but generally manageable; mild dementia had much higher symptomatic ARIA.
- Are there molecular subtypes of Alzheimer brains?
Brain multi-omics integration found four multimodal AD profiles, including severe Knight-C4 with worse cognition and heavy molecular dysregulation.
Learn alongside