Four decades of memory science, in a three-minute test.
MemTrax is a continuous-recognition memory test built on Dr. J. Wesson Ashford's 40+ years of Alzheimer's research at Stanford and the VA Palo Alto. More than 600,000 tests, taken by over 340,000 people worldwide.
Runs in any browser · ~3 minutes · No installation · Language-independent
Memory loss is usually found years too late. Our mission is to make memory measurable — for everyone, everywhere.
Alzheimer's begins damaging memory long before a diagnosis is made. A test that is fast, precise, and available in any browser means the earliest signal no longer has to wait for the clinic.
Science first
Every scoring decision traces to peer-reviewed research and a normative base of hundreds of thousands — never to convenience.
A screen anyone can take
Three minutes, any device, no installation, language-independent. Early detection only matters if it reaches people at scale.
Support, never replace
MemTrax informs clinical judgment — it doesn't substitute for it. Results are built to be shared with, and acted on by, clinicians.
One task. A reproducible signal of memory function.
You view a sequence of images; half of them repeat. For each one, you mark whether you've seen it before. The test records accuracy and reaction time on every response — a precise, language-independent measure of recognition memory, the function most specifically disrupted in early Alzheimer's disease.
Present
50 images appear in sequence, half of them repeats at varied intervals.
Respond
You mark each image as seen before or new, as quickly as you can.
Capture
Percent correct and response time are recorded to the thousandth of a second.
Score
A composite is normed against a reference dataset of hundreds of thousands.
The science behind MemTrax didn't start with a website.
It grew out of more than forty years of Alzheimer's research — from foundational neuroscience to a memory test now taken hundreds of thousands of times worldwide.
First double-blind cholinesterase trial for AD
Dr. Ashford publishes the first double-blind study of a cholinesterase drug to treat Alzheimer's disease — the class of medication still most widely prescribed for the condition today.
Ashford et al. · Am J Psychiatry, 1981
The neuroscience foundation
His PhD dissertation is the first to demonstrate massive reciprocal processing by neurons in the brain — a finalist for the Society for Neuroscience's Lindsley Prize, and the groundwork for understanding how Alzheimer's attacks neurons.
The Neuroplasticity Hypothesis
He proposes the neuroplasticity hypothesis of Alzheimer's disease — still a leading model for the pathology today. At its basis, AD is a disorder of neuroplasticity that specifically disrupts episodic memory.
Brief Alzheimer Screen
Work on practical screening accelerates, including a Brief Alzheimer Screen derived from the CERAD dataset — part of a long focus on detecting impairment early, and at scale.
Mendiondo, Ashford, Kryscio & Schmitt · Stat Med
The continuous-recognition test, in front of audiences
A 50-image continuous-recognition test is administered to 1,018 participants across 25 sites. Recognition memory (d′) declines measurably with age — establishing the paradigm that becomes MemTrax.
Ashford, Gere & Bayley · J Alzheimers Dis, 2011
Validated against the MoCA
In a clinical sample, every MemTrax measure tracks with the MoCA, and cutoffs for detecting MCI are established — moving the test from concept to validated screen.
van der Hoek, Nieuwenhuizen, Keijer & Ashford · J Alzheimers Dis, 2019
Machine-learning classification of MCI
Across 259 patients at two hospitals in China, MemTrax metrics combined with machine learning classify cognitive health against the MoCA — establishing robust, model-based screening.
Bergeron et al. · J Alzheimers Dis, 2020
Outperforming the MoCA
In a 150-participant Chinese cohort, MemTrax detects MCI-AD at AUC 0.839 versus the MoCA's 0.740 — better accuracy than the standard, in a head-to-head comparison.
Liu et al. · J Alzheimers Dis, 2021
Normative data at population scale
Analysis of 344,165 unique individuals across 602,272 tests defines acceptable performance ranges and a reverse-exponential response-time model; a Chinese normative study of 26,633 subjects builds percentile curves by age and education; and the paradigm extends to post-stroke cognitive impairment.
Ashford, Clifford, Bergeron et al. · Frontiers in Aging Neuroscience & Aging, 2022–2023
The evidence base, in the numbers.
MemTrax sits within Dr. Ashford's broader body of work — hundreds of publications across four decades of Alzheimer's and memory research. The findings below are specific to the test.
vs. MoCA 0.740
| Study | Cohort | Key finding | Metric |
|---|---|---|---|
| Liu et al., 2021 J Alzheimers Dis | 150 (CON / MCI-AD / AD), China | Detected MCI-AD with higher accuracy than the MoCA (72% sensitivity, 84% specificity) | AUC 0.839 vs MoCA 0.740 |
| Bergeron et al., 2020 J Alzheimers Dis | 259 patients, two hospitals, China | Machine-learning models on MemTrax metrics classified MCI vs. normal cognition | 10-fold CV AUC |
| van der Hoek et al., 2019 J Alzheimers Dis | 82 (45 normal / 37 MCI) | All MemTrax measures associated with the MoCA; MCI cutoffs established | MTx vs MoCA |
| Ashford et al., 2022 Front Aging Neurosci | 344,165 individuals · 602,272 tests | Defined normal performance ranges; reverse-exponential response-time model | Normative |
| GAMLSS norms, 2023 Alzheimers Dement (DADM) | 26,633 subjects, China | Age- and education-specific percentile curves for mass screening | Percentile norms |
| Zhao et al., 2023 Front Hum Neurosci | Ischemic-stroke patients | MemTrax screens for post-stroke cognitive impairment in ~1.5–2 minutes | PSCI screen |
Representative selection. Full citations and DOIs in the Publications section below.
The peer-reviewed record.
A selection of the published work on MemTrax and the continuous-recognition paradigm. These sit within Dr. Ashford's broader bibliography of hundreds of papers on Alzheimer's disease and memory.
- 2023 10.18632/aging.204828
- 2022 10.3389/fnagi.2022.1005298
- 2021 10.3233/JAD-200936
- 2020 10.3233/JAD-191340
- 2019 10.3233/JAD-181003
- 2019 10.3233/JAD-181198
- 2011 10.3233/JAD-2011-110950
A fuller bibliography — including the post-stroke (Zhao et al., 2023), GAMLSS normative (2023), and early screening publications — is maintained at medafile.com and on PubMed. Citations available on request.
A career spent making memory measurable.
Dr. Ashford graduated from UC Berkeley in 1970 and earned his MD and PhD at UCLA, where his dissertation was the first to demonstrate massive reciprocal neuronal processing in the brain. In 1981 he published the first double-blind study of a cholinesterase drug for Alzheimer's — the drug class still most widely prescribed today — and in 1985 proposed the Neuroplasticity Hypothesis of Alzheimer's, still a leading model of the disease.
He was vice-chair for research in psychiatry at the University of Kentucky, and since 2003 has been a senior research scientist at the Stanford / VA Aging Clinical Research and Alzheimer's Disease Centers. He chairs the Memory Screening Advisory Board of the Alzheimer's Foundation of America, coordinating its National Memory Screening Initiative.
"Cognitive tests that can rapidly, sensitively, frequently, inexpensively, and precisely measure memory are principally needed."
MemTrax is the applied result of that conviction — and of the company built and run by his son, Curtis B. Ashford, to bring the test to the world.
From a research instrument to a worldwide screen.
The continuous-recognition test has been deployed across research registries, international clinics, and large-scale screening programs.
Brain Health Registry
Integrated into the Brain Health Registry by Dr. Michael Weiner, principal investigator of the Alzheimer's Disease Neuroimaging Initiative (ADNI).
China & Europe
A dedicated Chinese-language deployment supports large validation cohorts; an early European version was developed with HAPPYneuron in France.
AFA Memory Screening
Aligned with the Alzheimer's Foundation of America's national initiative and proposed for the Medicare Annual Wellness Visit.
A clear, objective readout — the moment the test ends.
One short recognition task produces a normative summary you, or your clinician, can act on immediately.
Not a diagnosis. MemTrax measures recognition memory and is intended to support, not replace, clinical evaluation.
One test is a snapshot. Track your memory over time, against your own baseline.
A library producing 600+ unique test variants means you can retest as often as you like with minimal learning effect — and watch your trend line, not a single score.
- Your personal baseline. Repeated tests establish the range that's normal for you — the reference that matters most.
- Change you can see early. A drift below your own band is a more sensitive signal than any one-time cutoff.
- A record worth bringing to the clinic. Share a longitudinal chart with your physician instead of a memory of how you felt.
Pursuing FDA qualification as a Drug Development Tool.
MemTrax is advancing toward Clinical Outcome Assessment qualification, with MemTrax as primary Requestor.
Validation
Peer-reviewed studies vs. the MoCA.
CompleteNormative base
600K+ tests; published norms.
CompleteDDT / COA
Qualification package in progress.
In progressQualification
FDA review and determination.
PlannedForty years of science. Three minutes of your time.
Take the test yourself, or bring a validated recognition assessment into your practice.
Not a diagnosis · Intended to support clinical evaluation