Pupillometer

A pupillometer is an instrument that takes measurements of the pupils of the eyes. The word has two common meanings. In optometry and eyewear dispensing, a pupillometer (also called a PD meter) measures interpupillary distance (pupillary distance, or PD), usually by locating the light reflections on the two corneas. In medicine and research, a pupillometer measures the size of the pupil and the pupillary reflexes, such as constriction in response to light (the pupillary light reflex); this measurement is called pupillometry.

Both meanings matter for virtual reality (VR) and augmented reality (AR). A PD pupillometer is one of the clinical tools used to measure interpupillary distance, which head-mounted display makers such as Meta advise users to match with the headset's lens spacing. The eye tracking cameras built into some headsets can also estimate pupil diameter, and researchers have used headsets such as the HTC Vive Pro Eye and Pico Neo 2 Eye as pupil-size pupillometers.

Reviewed 19 October 2026. Fact-check of every claim against the cited papers (full text or abstract; authors, venues, years and DOIs checked via Crossref and Europe PMC), 21 CFR 886.1700, the AccessGUDID records, the NIDEK, NeurOptics, Varjo and Meta pages and the 20/20 Magazine article. About review dates.

Terminology and regulation

The term covers two different kinds of instrument that share a name. Papers on spectacle fitting use "pupillometer" for a PD-measuring device and note that it is also known as a PD meter. Clinical and psychology papers use the same word for devices that record pupil diameter and pupil reflexes. The spelling "pupilometer", with one "l", also appears in trade and research writing.

In the United States, the federal device classification regulation 21 CFR 886.1700 defines a pupillometer as "an AC-powered or manual device intended to measure by reflected light the width or diameter of the pupil of the eye". Both the AC-powered and the manual device are Class I (general controls) and exempt from premarket notification, subject to the limitations in 21 CFR 886.9. In the FDA's Global Unique Device Identification Database, devices of both kinds are listed under the same product code, HLG ("Pupillometer, Ac-Powered"): the NIDEK PD Meter PM-700, a pupillary distance meter, and the NeurOptics NPi-300, a pupil size and reactivity device that NeurOptics markets for critical care.

Pupillary distance pupillometers

How they work

A ruler measurement of PD records the distance between common points in the two eyes (Gantz and colleagues used the limbus, the edge of the iris), and is often called the anatomical IPD. An automatic pupillometer instead measures the distance between the two principal corneal light reflexes, which is often called the physiological IPD. In the TOPCON PD-5 used in a 2024 comparison study, the operator aligns line markers with the corneal light reflex and needs to look through the device with only one eye.

Writing in 20/20 Magazine, optician Barry Santini explains that corneal reflex pupillometers sight the reflection with a coaxially mounted light source, which avoids the parallax error that occurs when the fixation light is off the operator's line of sight. He notes that the corneal reflex is typically about 1 mm nasal to the pupil center at the spectacle plane, so a binocular PD taken from the reflexes is about 2 mm narrower than one taken from the pupil centers. He also lists sources of error, including variation in angle kappa, slow head rotation during the measurement and uneven contact with the forehead rest, and advises always taking PD readings at least twice. Gantz and colleagues add that the iris and pupil are dynamic structures, so the corneal reflex measurement has an intrinsic imprecision that cannot be removed.

NIDEK states that its PD Meter PM-700 calculates PD for a viewing point distance from 30 cm to infinity with one measurement, displays the result in 0.1 mm increments (or 0.5 mm, as on the earlier PM-600), uses a hairline with 0.25 mm increments for alignment, and has a running time about three times longer than the PM-600. The Essilor Pupillon used by Gantz and colleagues displays the mean of four readings.

Santini writes that pupillometers existed before the 1970s, and that Varilux and other lens companies of the day, American Optical and Titmus, each brought newly designed pupillometers to market, with improvements that included a rest point on the bridge of the nose. He describes the half century from the 1920s to the 1970s as the period when binocular PD measurements were the norm.

Accuracy and repeatability

Several peer-reviewed studies have measured how well PD pupillometers perform, both against physical targets and against ruler methods.

Study	Method	Main findings
Obstfeld and Chou (1998), Ophthalmic and Physiological Optics	Nine corneal reflection pupillometers from three makers, tested on two steel balls spaced to simulate PDs of 55, 60, 65 and 70 mm; two observers	Average error of the mean reading across all PDs 2.3 mm; average standard deviation 0.26 mm; most accurate at 65 mm; poor eye relief and scale friction were significant error sources; the authors advise checking accuracy periodically, especially for very small and very large PDs
Holland and Siderov (1999), Ophthalmic and Physiological Optics	Viktorin's method, corneal reflection and a pupillometer; one experienced optometrist measured 49 people twice, and 35 optometrists measured one person	Limits of agreement were narrowest for the pupillometer and widest for Viktorin's method; inter-examiner repeatability was slightly poorer than intra-examiner repeatability
Murray, Hunfalvay and Bolte (2017), Translational Vision Science & Technology	IPD from a remote infrared eye tracker compared with a PL850 pupilometer and an Essilor Digital CRP	Correlation with the PL850 r = 0.971 (mean difference 0.407 mm) and with the Essilor device r = 0.985 (659 measurements each)
Gantz, Shneor and Doron (2021), Journal of Optometry	Millimeter ruler versus Essilor Pupillon in 199 adults; repeatability retest of 30 people after 1-2 weeks	Statistically but not clinically different; mean differences (manual minus automatic) 0.21-1.25 mm depending on condition; 63-74% of monocular and 77% of binocular distance differences within the clinical range; both methods repeatable
Jung and Chu (2024), Clinical Optometry	TOPCON PD-5 pupillometer as reference, versus PD ruler, autorefractor and a phone app in 40 adults	Ruler gave larger and the app smaller distance PDs than the pupillometer; resulting prism errors were within the ISO 16034:2002 tolerance for all methods

Pupil size pupillometers

Development

In 1958 Otto Lowenstein and Irene Loewenfeld described a new electronic pupillography instrument and some of its clinical applications in the A.M.A. Archives of Ophthalmology. According to Merlin Larson and Vineeta Singh, writing in 2016, infrared pupillometry was introduced in 1962, but portable instruments using the technology had only recently become available in hospitals. Larson and Matthias Behrends date portable infrared pupillometry to 1989 and describe the principle: the instrument floods the eye with infrared light and measures the reflected image on an infrared sensor, then calculates pupil size, light reflex variables and reflex dilation and shows them on a screen after each time-stamped measurement. Larson and Singh note an early concern that eye movements of the subject or hand movements of the examiner could introduce artifacts resembling a light reflex, and state that algorithms built into the instruments maintain a steady image of the pupil so that such movement does not produce false readings.

Clinical use and the Neurological Pupil index

Automated pupillometers report measures such as baseline pupil diameter, minimum diameter after the light stimulus, percentage constriction, latency, constriction velocity and dilation velocity; Citerio and Taccone note that subtle pupil changes are hard to quantify accurately by visual inspection alone. The Neurological Pupil index (NPi), a proprietary composite score from NeurOptics devices, was introduced in a 2011 paper by Jeff W. Chen and colleagues. In 134 intensive care patients at eight sites, examined with a portable handheld pupillometer, patients with abnormal pupil reactivity had an average peak intracranial pressure of 30.5 mmHg versus 19.6 mmHg for those with normal reactivity, and the first pupil abnormality appeared on average 15.9 hours before the pressure peak.

In a 2026 review, Giuseppe Citerio and Fabio Taccone (both on NeurOptics' scientific advisory board) explain that the NPi compares several features of the light reflex with a manufacturer-derived normative database and reports a value from 0.1 to 4.9 when a reaction is detected; values of 3 or more are generally treated as normal, and 0 means no measurable response. They also note that there is no randomized evidence that NPi-guided management improves outcomes. A secondary analysis of the multicenter ORANGE study, covering 318 patients with acute brain injury, found that NPi values overall were not significantly associated with invasively measured intracranial pressure. NeurOptics, based in Irvine, California, lists the NPi-300 and NPi-200 pupillometers for critical care and the VIP-400 and PLR-4000 for applied research, and describes the NPi-300 as a handheld device that provides pupil size and reactivity data.

Psychology research

In cognitive science, pupillometry is used as an indirect measure of mental activity. In a 2018 review, Sebastiaan Mathôt describes three kinds of pupil response: constriction to brightness (the pupil light response), constriction during near fixation (the pupil near response), and dilation with arousal and mental effort (the psychosensory pupil response). He writes that a series of studies by Hess and Polt (1960 and 1964) and later by Kahneman and Beatty (1966) brought pupillometry to the attention of psychologists: in Hess and Polt's 1964 study the pupil grew larger as mental calculations became harder, and in Kahneman and Beatty's study pupil size reflected the number of digits being memorized. Pupil diameter can also be recorded with general-purpose video eye trackers; Murray and colleagues, for example, recorded pupil diameter with a 120 Hz remote eye tracker that uses infrared corneal reflection.

Use in VR and AR

Headset fit

Meta's support documentation for its Quest headsets states that lens spacing should line up with the wearer's IPD as much as possible for the best image clarity, and recommends an optometrist for the most accurate IPD measurement, with a mirror and millimeter ruler as a home alternative. The 2017 study by Murray and colleagues showed that an infrared eye tracker can measure IPD in close agreement with two corneal reflection pupilometers. As with spectacles, the result depends on the reference point: a corneal reflex pupillometer gives the physiological IPD, while a ruler gives the anatomical IPD, and the two differ slightly.

Headset eye trackers as pupillometers

Some headset eye tracking interfaces report pupil size directly. Varjo's developer documentation lists per-eye pupil diameter estimates in millimeters, iris diameter estimates and a pupil-to-iris ratio, and states that these fields "are experimental features and provided in this release without any accuracy guarantees".

Peer-reviewed studies that used headsets this way include:

Study	Headset or device	Purpose and findings
Pielage et al. (2025), Ear and Hearing	HTC Vive Pro Eye built-in eye tracker, both eyes at roughly 90 Hz	Listening effort in a VR speech-in-noise task with 33 normal-hearing adults; task load changed baseline pupil size and peak pupil dilation, while the presence of virtual people did not affect any outcome measure
Negi et al. (2024), Frontiers in Ophthalmology	The authors' Pupil+ pupillometer rebuilt on a Pico Neo 2 Eye headset with built-in Tobii eye tracking (90 frames per second)	Detection of relative afferent pupillary defect in 77 patients with unilateral optic neuropathy and 77 controls; sensitivity 78.5%, rising to 85.1% after correcting for physiological asymmetry; specificity and area under the ROC curve between 81% and 96.3%
Tang et al. (2025), Ophthalmology	VR head-mounted display used for "virtual reality-based infrared pupillometry" (VIP)	Light responses of 185 participants; constriction time after the brightest stimulus was significantly associated with long COVID status
Karaer et al. (2026), Scientific Reports	BulbiCAM, a VR-headset eye tracker from BulbiTech AS (Oslo), compared with Pupil Labs Neon glasses	In 39 healthy adults, pupil tests were reproducible between visits (ICC 0.76-0.88 for pupil and pursuit tests); bias against the Neon glasses was -0.48 mm for baseline diameter and -0.56 mm for peak constriction
Courtney et al. (2024), JMIR Serious Games	VR nicotine cue paradigm with eye tracking	In 108 nicotine and tobacco product users, pupil diameter was larger in response to product cues than to control cues

Scene brightness complicates this use. In a 2026 paper, Yilmaz and colleagues write that luminance-driven pupillary light reflexes can mask the subtler task-evoked pupil responses in VR. They proposed an iso-luminant VR protocol that holds visual luminance constant and varies cognitive load only through auditory tasks; in 22 sessions, the light reflex had a mean amplitude of 1.52 mm, while high load produced a mean extra dilation of 0.38 mm over low load.

See also

• Interpupillary distance
• Eye tracking
• Eye
• Human visual system
• Infrared