SCIENCE · EVIDENCE LIBRARY

The eye is a measurement surface.
The claims are not interchangeable.

Retina, external-eye photography, pupil dynamics and iris morphology are different evidence layers. This page keeps the published findings attached to the modality that produced them.

WHY THE EYE?

Neural, vascular and visible.

The eye gives non-invasive access to structures tied to the nervous and vascular systems. That makes ocular imaging scientifically interesting—but it does not make every ocular feature a validated biomarker.

≈1.2Mretinal ganglion-cell axons converge to form the adult optic nerveNCBI / StatPearls ↗
01

Ocular imaging ↔ systemic signals

Strongest published evidence currently comes from retinal fundus imaging, OCT and external-eye photography—not from conventional iris organ mapping.

RETINAL FUNDUS · NATURE BIOMEDICAL ENGINEERING · 2018
284,335patients in training

Cardiovascular risk factors from retinal photographs

Deep learning predicted age, sex, smoking status and systolic blood pressure from retinal fundus images. Major adverse cardiac events were classified at AUC 0.70 in the reported validation analysis.

Open paper ↗
EXTERNAL EYE · NATURE BIOMEDICAL ENGINEERING · 2022
145,832patients in development

Poor glucose control from photographs of the front of the eye

The model was validated in 48,644 additional patients. For HbA1c ≥9%, reported AUCs across four validation sets were 0.700, 0.734, 0.693 and 0.676. Elevated lipids were explored separately.

Open paper ↗
RETINAL VESSELS · NATURE BIOMEDICAL ENGINEERING · 2020
70k+retinal images

Retinal-vessel calibre and cardiovascular risk

Automated vessel-calibre measurements showed high agreement with expert graders (ICC 0.82–0.95) and were associated with blood pressure, BMI, total cholesterol, HbA1c and incident cardiovascular disease.

Open paper ↗
OCULOMICS · NATURE CARDIOVASCULAR RESEARCH · 2026
152,918CFP + OCT training images

Ophthalmic imaging linked to cardiovascular and neurological traits

A UK Biobank multi-omic analysis found retinal imaging representations associated with ischemic heart disease, cerebrovascular disease, Parkinson’s disease and dementia, including prospective disease associations.

Open paper ↗
RETINAL VASCULATURE · TROMSØ EYE STUDY · 2013
6,353participants

Retinal vessel calibre tracked established cardiovascular risk factors

Arteriolar calibre was associated with blood pressure, HbA1c and smoking; venular calibre was associated with blood pressure, BMI, HDL/LDL cholesterol and smoking in this population-based study.

Open paper ↗
RETINAL VASCULATURE · MESA · 2006
5,979participants

Microvascular calibre reflected metabolic and inflammatory context

Retinal arteriolar and venular calibre were associated with hypertension, diabetes, obesity and dyslipidaemia. Venular calibre was also associated with inflammatory markers.

Open paper ↗
RETINAL VASCULATURE · EPIC-NORFOLK · 2018
5,947participants

Tortuosity and vessel width tracked cardiometabolic traits

Venular tortuosity was associated with BMI, HbA1c and type 2 diabetes; arteriolar tortuosity was associated with age and systolic blood pressure.

Open paper ↗
RETINAL AGE · NATURE COMMUNICATIONS · 2026
71,343UK Biobank participants

Retinal age gap and systemic aging

A retinal foundation-model age estimate achieved a mean absolute error of 2.85 years. Retinal age gap was associated with cardiometabolic traits, cognition, mortality, dementia and incident cardiovascular disease.

Open paper ↗
02

Pupil response ↔ cognition + autonomic function

Pupillometry measures a dynamic physiological response. Research spans light reflex, autonomic function, attention, cognitive effort and working memory. Clinical use still depends on calibrated acquisition and validated protocols.

PARKINSON’S · SYSTEMATIC REVIEW + META-ANALYSIS · 2025
31papers reviewed

Light-reflex dynamics in Parkinson’s disease

Eleven studies entered meta-analysis. Parkinson’s groups showed significantly reduced maximum constriction velocity and amplitude, with longer constriction latency. The authors stress protocol heterogeneity and the need for standardisation.

Open paper ↗
MCI · PROSPECTIVE COHORT · 2024
727participants

Task-evoked pupil response in mild cognitive impairment

The study included 213 participants with MCI and 514 cognitively normal controls. MCI participants showed reduced pupil dilation during cue and target-onset periods in oculomotor cognitive tasks.

Open paper ↗
DEMENTIA / MCI · SCOPING REVIEW · 2023
8included studies

Pupil response during cognitive tasks

Across the included literature, task-evoked pupillary responses differed between cognitively impaired and healthy groups. Reduced dilation was reported in Alzheimer’s disease; evidence in MCI remained less consistent.

Open paper ↗
WORKING MEMORY · REVIEW · 2024
3memory operations

Encoding, maintenance and prioritisation

A review of visual working-memory research describes pupil dynamics across encoding, maintenance and prioritisation, and discusses relationships between pupil size and memory performance.

Open paper ↗
ALZHEIMER’S DISEASE · PROOF-OF-CONCEPT · 2024
254participants

Quantitative light-reflex pupillometry in dementia

A university memory-clinic study reported AUROC values of 0.74–0.81 for models distinguishing Alzheimer’s disease from healthy controls and other dementias. It remains proof-of-concept rather than a stand-alone diagnostic test.

Open paper ↗
COGNITIVE PUPILLOMETRY · METHODS REVIEW · 2023
1major constraint

Measurement is powerful; methodology still varies

Cognitive pupillometry is used to study attention, mental effort and working memory, but experimental design and preprocessing are not yet standardised across research groups.

Open paper ↗
LC–NE / CORTICAL PLASTICITY · REVIEW · 2023
LC–NEphysiology

Pupil dynamics as a window into arousal and neuromodulation

A review links pupil-size variation with arousal, attention and mental effort, discussing the locus-coeruleus noradrenergic system as a key physiological pathway behind these effects.

Open paper ↗
COGNITIVE LOAD · EXPERIMENTAL STUDY · 2026
6 / 9digit-load conditions

Pupil dilation changed with increasing cognitive load

Under higher-load conditions, older adults showed reduced pupillary reactivity; pupil dilation also correlated with age and Stroop interference in the reported sample.

Open paper ↗
03

Iris structure ↔ measurable morphology

Here the evidence is about anatomy, morphology, genetics and ocular geometry. It is a different evidence class from retinal systemic biomarkers and from cognitive pupillometry.

IRIS MORPHOLOGY · GWAS · 2011
2,121discovery sample

Crypts, furrows and pigmentation have measurable genetic associations

Replicated associations linked crypt frequency with SEMA3A, contraction furrows with TRAF3IP1 and the peripupillary pigmented ring with SLC24A4. These are morphology/genetics findings—not systemic-disease maps.

Open paper ↗
IRIS SURFACE · SSOCT · 2016
1,704participants

Surface features correlated with iris volume and curvature

Higher crypt grade was associated with smaller iris volume, while darker irides and higher furrow grade were associated with larger iris volume in Asian eyes.

Open paper ↗
IRIS SURFACE · STANDARDISED GRADING · 2014
κ .775crypt agreement

Crypts and furrows can be graded with reproducible agreement

The study reported strong intragrader agreement and intergrader weighted κ values of 0.775 for crypts and 0.836 for furrows, alongside associations with iris thickness.

Open paper ↗
IRIS MORPHOLOGY · POPULATION STUDY · 2016
3ancestry groups

Iris surface-feature frequency differs across populations

Fuchs’ crypts, Wolfflin nodules, pigment spots, contraction furrows and conjunctival melanosis showed population-dependent frequency differences across East Asian, European and South Asian groups.

Open paper ↗
IRIS / CAD · EXPLORATORY ML · 2023
198participants

A small iris–coronary-disease study reported high classification accuracy

The proof-of-concept study included 94 participants with CAD and 104 controls and reported 93% SVM accuracy. It is a single small study and does not establish clinical validity or validate iridology.

Open paper ↗
04 · EVIDENCE BOUNDARY

Iris morphology stays a research question.

IRISPECTRA can measure visible structure from calibrated iris photographs. Published retinal, external-eye and pupillometry findings do not validate disease, organ or personality inference from iris zones. Any iris-specific systemic association must be tested independently against clinical ground truth.