Remote VPC eye-tracking detects preclinical cognitive decline
PrimaryNeurotrack's eye-tracking approach is based on the claim that gaze behavior during a Visual Paired Comparison recognition-memory task reflects cognitive function relevant to early Alzheimer's-related decline. If device-embedded cameras can accurately capture novelty-preference performance, then webcam-based VPC tasks should provide a scalable way to screen for subtle cognitive impairment before conventional clinical evaluation is widely available. Testable predictions are that webcam-scored VPC performance will correlate with clinical-grade eye-tracker scoring and with established cognitive composites sensitive to preclinical Alzheimer's disease, and that lower novelty-preference performance will associate with poorer cognitive function in domains affected by early decline.
Popperian evaluation
The premise is credible: VPC novelty preference is a recognition-memory behavior, and the supplied studies link it to cognitive composites used in preclinical Alzheimer’s research. The webcam premise also has direct device-validation evidence, including r = 0.92 to 0.95 correlations with clinical-grade eye tracking. The weak point is biological specificity. These data support cognitive signal capture, but they do not show that low webcam VPC performance specifically detects Alzheimer’s pathology before clinical evaluation.
Supporting evidence: In 49 clinically normal older adults, webcam VPC novelty preference correlated with PACC (r = .39, P = .007) and NIHTB-CB (r = .35, P = .03).; Webcam VPC scores correlated strongly with 60 FPS commercial eye-tracker scores: trial-level r = .82 and overall mean r = .92.; A separate 18-subject, three-visit study found mean VPC webcam versus clinical-grade eye-tracker correlation of r = 0.95, with p values below 0.001.
Counter evidence: The cognitive correlations are modest, so VPC performance probably captures only part of the relevant cognitive variance.; The evidence is in clinically normal older adults and does not establish conversion prediction, amyloid or tau association, or diagnostic separation from non-Alzheimer causes of cognitive variation.
The theory explains two observed facts well: webcam gaze scoring can approximate lab eye tracking, and VPC novelty preference tracks some cognitive differences. It explains less well why those differences should indicate preclinical Alzheimer’s decline rather than attention, visual acuity, task engagement, education, sleep, device conditions, or general cognitive reserve. The signal is real enough to take seriously, but it is not yet disease-specific.
Supporting evidence: Lower VPC novelty preference associated with poorer letter fluency (r = .33), category fluency (r = .36), and Trail Making Test A performance (r = -.40).; The strength of VPC relationships with cognitive composites reportedly did not differ between camera devices.; Manual scoring of webcam data had high inter-rater agreement, with kappa = .84.
Counter evidence: Alternative explanations remain plausible because the supplied evidence does not tie VPC performance to Alzheimer’s biomarkers or longitudinal decline.; The main cognitive-composite relationships are modest, which leaves substantial unexplained variance.; Small samples limit how much explanatory weight the theory can carry.
The theory is easy to test and easy to break. It predicts strong agreement between webcam and clinical-grade eye tracking, measurable correlation with Alzheimer’s-sensitive cognitive composites, and lower scores in people with poorer early-decline cognitive profiles. A well-powered study could falsify it if webcam scores fail device agreement, fail cognitive association, or fail longitudinal prediction of decline.
Supporting evidence: The theory names measurable endpoints: novelty-preference performance, webcam versus clinical-grade scoring agreement, PACC, NIHTB-CB, fluency tasks, and Trail Making Test A.; Existing studies already test parts of the claim with reported correlations and p values.; The webcam-device claim has a clear failure mode: poor agreement with clinical-grade eye tracking under real remote conditions.
Counter evidence: The strongest falsification test, prediction of future cognitive decline or Alzheimer’s biomarker status, is not shown in the supplied evidence.; The phrase 'preclinical cognitive decline' can drift unless future studies define thresholds, time horizons, and clinical endpoints upfront.
Reasoning tree
Public endorsements
The public evidence here ties Cecilia Manzanares to Neurotrack as a scientific co-founder, but it does not show her publicly stating, endorsing, or disputing the specific theory that webcam-based VPC eye-tracking can detect preclinical cognitive decline. Company materials show Neurotrack's Alzheimer’s focus and name her as a co-founder, which is indirect context, not a public position from her on this claim.
Evidence publication IDs: 538109e2-11f1-465e-b9a1-1ae31dfb86b1, 4fee2922-5918-4cf8-97c1-9bdfb4bf85fc
The public evidence ties Elizabeth Buffalo to Neurotrack as a co-founder and to learning, memory, and cognitive aging research, but it does not show her explicitly endorsing, describing, or disputing Neurotrack's specific claim that webcam-based Visual Paired Comparison eye-tracking can detect preclinical cognitive decline.
Evidence publication IDs: 538109e2-11f1-465e-b9a1-1ae31dfb86b1
Kaplan publicly talks about brain health checks, early intervention, prevention, and Neurotrack's cognition mission, but the provided evidence does not mention the specific theory here: VPC eye-tracking, novelty-preference scoring, webcam capture, or correlation with clinical cognitive measures. On this theory, the record is silent.
The record ties Jeff Magnusson to Neurotrack as CTO, but none of the provided quotes or publications show him discussing the eye-tracking VPC theory, webcam-based screening, novelty preference, or preclinical Alzheimer's detection. On this theory, he is publicly silent in the supplied evidence.
The public evidence here ties Stewart Zola to memory neuroscience and identifies him as a Neurotrack scientific co-founder, but it does not show him publicly endorsing, describing, or disputing the specific claim that webcam-based Visual Paired Comparison eye-tracking can detect preclinical cognitive decline. On this record, he stays silent on the theory itself.