SC Anderson, BA; RH Kardon, MD, PhD; YH Kwon, MD, PhD; ML Pereira, MD

University of Iowa Ophthalmology and Visual Sciences and Veterans Administration, Iowa City, IA

INTRODUCTION: To study the effect of unilateral high intraocular pressure (IOP) and its treatment on the asymmetry of optic nerve function over time as measured by the pupil afferent input between the two eyes.

Methods: Eight patients who presented with acute elevation of IOP in one eye were studied over an average of 138 days (range 70-206 days) of treatment using computerized infrared pupillography. A stimulus-response function (log intensity of stimulus vs. pupil contraction of the normotensive eye) was characterized as a function of light input to the affected eye and fellow, normotensive eye using a Naka-Rushton sigmoid curve fit. The difference in the area under each curve, a reflection of the relative afferent pupillary defect (RAPD), was quantified at each test session.

Results: There was a significant linear correlation between the average IOP over the time of the study and the change in the RAPD (R2=0.670, p value=.007) for the 8 patients. The slope of this line indicated that for every 10 mm Hg increase in average IOP, there was a worsening of approximately 4.4 decibels of RAPD. A significant correlation was also found between maximum IOP and worsening in RAPD (R2=0.506, p value=.048). Worsening in RAPD over time occurred from the initial test to the final test over the course of the study in 5 patients. In 4 patients the IOP was not easily controlled at the onset of the study, resulting in spiking of IOP during the study associated with a worsening in the RAPD, followed by an improvement, when the IOP was lowered. The transient change in function was delayed in 2 of the 4 patients with an IOP spike.

DISCUSSION: Elevation of IOP correlated with a proportional amount of optic nerve damage, as measured by the pupil light reflex, over the course of this study. Acute, transient elevations of IOP can cause a transient worsening of function, but it may be delayed. The pupil light reflex, unlike visual field testing, is objective and may be useful in monitoring changes in optic nerve function with time.

Acknowledgements: Veteran’s Administration Merit Review Grant, Zeiss-Humphrey Corp, Lew Wasserman Scholarship (Dr. Kardon), and Research to Prevent Blindness (RPB)





John L Barbur, J A Harlow and Sancho Moro

Applied Vision Research Centre, City University, Tait Building, Northampton Square, London EC1V 0HB.

Introduction: It has long been established electrophysiologically that the pupillomotor nucleus receives input from both transient and sustained neurones. These neurones have different temporal and spatial properties and therefore respond preferentially to different stimulus attributes. In this perimetric study we have investigated how the different spatio-temporal characteristics of visual stimuli can be used to isolate the transient and sustained components of the pupil light reflex response.

Methods: We measured pupil responses, elicited either by sinusoidal or square-wave temporal modulation of luminance contrast. Our pupil perimetry program tests 24 stimulus locations, or any combination of these discrete stimuli in the visual field. Each pupil response measurement is based on the average of 8 traces. Pupil constriction amplitude and latency were computed from the first Fourier harmonic of each averaged trace.

Results: Much larger responses were found for square-wave as compared to sinusoidal modulation. This difference cannot be accounted for simply as a contribution of higher harmonics present in the square-wave modulation or the non-linear response properties of the pupil when sinusoidally modulated stimuli are employed.

Discussion: The results are consistent with two mechanisms driving the pupil light reflex response, a sustained and a transient mechanism. The sinusoidal and square-wave modulations favour the sustained and transient mechanisms. The sustained mechanism has a more linear contrast gain, integrates the input signal over large areas of the visual field and is therefore appropriate to control the steady state size of the pupil in response to changes in ambient illumination. The transient mechanism has a non-linear contrast gain and responds preferentially to rapid, spatially localised luminance contrast changes. This component shows much reduced spatial summation. Isoluminant coloured stimuli, on the other hand, produce the same response for both types of stimulation. These findings can be used to explain why in some cases of retinal or optic nerve disease, the transient component of the pupil light reflex response can be spared in patients that exhibit a large RAPD



Greater Pupillary Escape Differentiates Central From Peripheral Visual Field Loss

Oliver Bergamin M.D. and Randy Kardon M.D. Ph.D.

Department of Ophthalmology and Visual Science, Veterans Administration Hospital and University of Iowa, Iowa City, Iowa.


PURPOSE: To test whether pupil escape observed during a constant light stimulus was greater in eyes with central visual field loss compared to eyes with peripheral visual field loss and normal eyes.

METHODS: 27 normal subjects, 5 patients with central field loss, 11 patients with peripheral field loss, and 8 patients with combined loss (central and peripheral visual field loss) were tested.

A dual-channel infrared pupillograph was used to simultaneously record the right and left pupil diameters at a rate of 60 Hertz in order to characterize the initial, phasic pupil contraction, and the sustained, or prolonged pupil contraction in response to a 5 second duration light stimulus. Full field light stimuli with a diameter of 60 degrees were presented to each eye at 7 different intensities.

The amplitude of the phasic pupil contraction was compared to the amplitude of the sustained pupil contraction at the four brightest intensities (slope of phasic versus sustained contraction) in the normal eyes and in eyes with visual field loss in the center and the periphery.

RESULTS: The sustained pupillary contraction showed a statistically greater reduction in proportion to the phasic component in eyes with central field loss compared to those with peripheral field loss, combination central and peripheral field loss and normal eyes.

CONCLUSIONS: Afferent neurons from the central retina normally contribute a substantial component to the sustained pupil response. Eyes with central field loss can be distinguished from eyes with peripheral field loss by a relatively greater loss of the sustained response, causing greater pupil escape.

This study showed that pupil movements in response to a non-perimetric full-field light stimulus may provide some degree of mapping of retinotopic sensitivity. This information may be of use in interpreting the pupillary movements during the swinging flashlight test and may help predict the pattern of visual field loss.

ACKNOWLEDGEMENT: This work was supported by the Swiss National Science Foundation, the Freiwillige Akademische Gesellschaft, Basel, Switzerland, the Roche Research Foundation, Basel, Switzerland, an unrestricted grant from Research to Prevent Blindness, New York, New York, a grant from Zeiss-Humphrey Corporation, Dublin, California, and a Merit Review and Career Development Award from the Veterans Administration, Washington, D.C. (Dr. Kardon). Dr. Kardon is the recipient of a Lew Wasserman Scholar Award (Research to Prevent Blindness).


Which Part Of The Shape Of The Pupillary Light Reflex Is Most Diagnostic?

Oliver Bergamin M.D. and Randy Kardon M.D. Ph.D.

Department of Ophthalmology and Visual Science, Veterans Administration Hospital and University of Iowa, Iowa City, Iowa.

BACKGROUND: The waveform of the pupil light reflex contains four well-defined landmarks: A) maximal acceleration during contraction (latency time point, or base line pupil size), B) maximal contraction velocity, C) pupil size at time of peak contraction, and D) maximal dilation velocity. Change of neuronal firing rate during specific time windows of the pupil light reflex can be quantified as change of pupil size between these landmarks.

PURPOSE: We evaluated changes in pupil size (neuronal firing) between different time windows of the pupil light reflex in patients and normal subjects to understand which regions were most susceptible to disease.

METHODS: Binocular infrared computerized pupillography was performed on 49 healthy subjects and on 30 patients with known unilateral or bilateral visual field damage. 30 degree diameter light stimuli were presented alternatively to the right and left eye for 200 ms duration every 3 seconds. Ten different stimulus intensities were given and each was repeated 4 times. Right and left pupil diameters were recorded at 60 Hz. The change in pupil size from the start to the end of each of the following time windows was determined: (1) latency time point to peak velocity, (2) latency time point to minimum pupil size, (3) latency time point to peak dilatation, (4) peak velocity to minimum pupil size, (5) peak velocity to peak dilatation, and (6) minimum pupil size to peak dilation. For each time window, the change in pupil size for right eye stimulation was plotted against the result from the left eye stimulation for the 10 intensities. A linear regression of this plot was used to determine slope, offset (y-intercept) and regression coefficient (R2) for each normal subject and each patient. These parameters of the linear regression were used as an indicator of symmetry of input between the two eyes. Each patient was then compared to the normal group for each time window of the pupil light reflex.

RESULTS: The time window of the pupil light reflex that showed the greatest difference from normal subjects was peak velocity to minimum pupil size, followed by latency time point to peak velocity, and latency time point to minimum pupil size. 27 of the 30 patients could be identified as having an abnormal amount of asymmetry between the inputs of the two eyes using this technique. The remaining 3 patients had symmetric input.

CONCLUSIONS: The most diagnostic part of the pupillary light reflex starts at maximal contraction velocity and ends at minimal pupil size. During this window of time, the input of each eye is very symmetric in normal subjects and is very sensitive to the effects of disease over the intensity range tested.

ACKNOWLEDGEMENT: Supported by the Swiss National Science Foundation, Freiwillige Akademische Gesellschaft, Roche Research Foundation, Research to Prevent Blindness, Zeiss-Humphrey Corporation, and a Merit Review and Career Development Award from the Veterans Administration.




P Berry MD and MD Larson MD

Department of Anesthesia, University of California, San Francisco, CA 94143-0648

INTRODUCTION: Noxious stimulation in awake subjects produces dilation of the pupil with a latency of between 300 and 400 msec. Preliminary studies have shown that during general anesthesia, the latency of pupillary dilation is much longer than 400 msec. The present investigation tests the hypothesis that the delay in pupillary dilation following noxious stimulation during general anesthesia is longer than 400 msec and is caused by slowly conducting afferent nerve (C) fibers that transmit nociceptive information.

METHOD: After Human Research Committee approval and informed consent, 8 volunteers were anesthetized with a propofol induction and maintained with 5% end-tidal desflurane. Vecuronium bromide infusion was used to maintain paralysis. Noxious stimuli were provided by 3 sec 100 Hz 60-80 mAmp electrical current every 5 min, alternating between an upper (C5) and lower (L4) dermatome. Infrared pupillary scans at 20 Hz were taken before and during each stimulus. Latency to pupillary dilation was measured by analyzing these initial pupillary scans to determine the time of onset for the first sustained dilation above the baseline diameter. Estimated additional conduction distances in peripheral nerves arising from the L4 stimulus (leg) compared to the C5 stimulus (shoulder) were estimated from body height. Comparisons between the latency of pupillary dilation between the L4 and C5 stimuli were made with unpaired t tests P<.05 was considered significant.

RESULTS: Mean latency of pupillary dilation following stimulation of the C5 dermatome was 900 msec and from the L4 845 msec P < .05. Estimated additional peripheral nerve conduction length for the 8 volunteers was 72 cm.

DISCUSSION: This report demonstrates that pupillary dilation following noxious stimulation has a long latency, approximately twice the latency that has been previously reported during the awake state. It is unlikely that delays in synaptic conduction brought about by desflurane could account for this effect, and suggests that pupillary dilation during anesthesia is brought about by a different mechanism compared to the dilation observed in the awake condition. C fibers, conducting at a rate of 1 meter/sec would expect to add over 500 msec to the latency of pupillary dilation arising from the L4 dermatome, compared to the C5 dermatome. We observed only a 55 msec delay which would be consistent with the theory that this reflex is generated by the more rapidly conducting A delta pain fibers. The prolonged latency of 800 ~ 900 msec suggests that pupillary dilation is part of the efferent arm, perhaps the motor component of the nociceptive reflex.

Reference: 1. Anesthesiology 87:849-855, 1997



P Bitsios*, E Szabadi and C M Bradshaw

Division of Psychiatry, University of Nottingham; *present address: Department of Psychiatry, University of Heraklion

We have shown previously that the threat posed by the anticipation of an electric shock increases the initial diameter of the pupil and decreases the amplitude of the light reflex response compared to periods when subjects are resting. We termed the threat-induced decrease in light reflex amplitude the "fear-inhibited light reflex" and proposed that it could be a potential laboratory model for human anxiety [1]. It is not clear, however, whether the observed changes in the pupillary measures are specific to the anxiety with which the threat of a shock is presumably associated, or whether they could be caused by the anticipation of any external, not necessarily harmful, event. To address this question, we examined the effects of the anticipation of an electric shock (putative aversive event) and the effects of anticipation of an acoustic stimulus (putative neutral event) on the light reflex.

Twelve healthy volunteers [6 male, 6 female] aged 18-35 years participated in a training session and a few days later in an experimental session. Pupil diameter was monitored in the dark, in dark-adapted eyes by an infra-red binocular television pupillometer. The light stimuli were flashes generated by a light-emitting diode (0.43 mW cm-2, 200 ms, peak wavelength 565 nm). A constant current square pulse (1.5 mA, 50 ms) was delivered twice to the skin overlying the median nerve of the left wrist. Acoustic stimuli (2 kHz, 100 dB, 200 ms) were delivered via headphones over a 70 dB background noise. Mood and feelings were self-rated on visual analogue scales.

The experimental session consisted of two parts, and each part comprised seven identical blocks of three light stimuli of the same intensity and duration. In each part, responses in each block were recorded under either relaxation (R) or anticipation (A) condition; the two conditions alternated regularly. "A" blocks were associated with the anticipation of an electrical stimulus ("shock") in Part 1, and with the anticipation of an acoustic stimulus in Part 2.

The anticipation of the electrical stimulus was associated with significant increases in initial pupil diameter and subjectively rated "anxiety" and "alertness", and a decrease in the amplitude of the pupillary light reflex response, whereas anticipation of the acoustic stimulus was associated with increases in initial pupil diameter and subjective "alertness" only.

These results show that the increase in initial pupil diameter is related to the anticipation of any stimulus, whereas the decrease in the amplitude of the light reflex response is associated with the aversiveness of the anticipated stimulus. Furthermore, the increase in subjective anxiety in association with the anticipation of the electric shock indicates that the anticipation of the aversive stimulus was anxiogenic.

1. Bitsios P, Psychopharmacol, 135, 93-98.



Esther Bulang1, Rolf Fötzsch2 and Holger Lüdtke3

1 Eye Clinic of the District Hospital Radebeul, 2 Department of Neurology at the University Hospital Dresden, 3 University Eye Hospital, Dept. of Pathophysiology of Vision and Neuro-ophthalmology, Tübingen, Germany

PURPOSE: Systemic side effects on the autonomic nervous system induced by botulinum neurotoxin type A (BoNT/A) injections have been reported. The effect of BoNT/A on the pupillary light reaction is discussed controversially.

METHODS: By means of an infrared pupillograph (250 Hz) (CIP Version 8.00 by AMTech Weinheim, Germany) the pupillary light reflex and pupil diameter were recorded during fixation of a target at a distance of 2.50 m in a dimmed room (4 cd/m²) in 11 patients with spasmodic torticollis (age 43 to 63) and 17 normals (age 50 to 64). Stimulation was performed with 78, 157, 627 and 1255 cd/m² light stimuli. In each case 4 pupillograms were recorded. Parameters of evaluation were amplitude, latency, relative amplitude and pupil diameter.

The patients were examined immediately before a new injection, and 7 and 14 days later. All patients received reinjection between 10 and 14 weeks after their last injection.

1. Measurements of the patients before injection therapy were compared with those of normals.

2. Measurements of patients from the moment just before BoNT/A injection were compared with their own values of 7 and 14 days later.

All results were analysed by means of ANOVA.

RESULTS: 1. Amplitudes, latencies, relative amplitudes and initial pupil diameter in patients before BoNT/A reinjections were not significantly different than in normals.

2. After the injection of BoNT/A the amplitudes decreased. The decrease was small but significant (mean difference 0.05 mm, p<0.0001) and also the latencies became longer (mean difference 4 ms, p<0.001). The relative amplitude decreased significantly, too (mean difference 0.87 % of the initial pupil diameter, p< 0.0001). The pupil diameter showed no significant difference.

CONCLUSIONS: In contradiction to another study (1) our investigations show significant subclinical changes in the pupillary light reflex in patients after injections with BoNT/A. Decreased amplitude and increased latency seem to be a consequence of a parasympathetic deficit of the iris sphincter muscle caused by blocked acetylcholine release after botulinum toxin A treatment.

1 Schnider et al.: Pupillometry after botulinum A toxin treatment. Neuro-ophthalmology 17 (1997) 5-7




Eric Granholm, PhD, Shaunna Morris, PhD, Erin Rogers, BA, Boris Vukov, MA

University of California, San Diego, and VA San Diego Healthcare System


INTRODUCTION: The study objective was to advance our ability to diagnose patients with Alzheimer's Disease (AD) early in the disease process using a relatively simple, noninvasive, less-expensive, diagnostic test involving the pupillary light reflex and tropicamide blockade of ocular cholinergic functions. Previous studies have not consistently supported the use of these diagnostic tests.

METHOD: AD patients (n=15), Parkinson's Disease (PD) patients (n=15) and normal control (NC) participants (n=15) were studied. Pupillographic methods were used to measure changes in pupil diameter and the pupillary light reflex during double-blind ocular administration of tropicamide (0.01%) and saline solutions.

RESULTS: All three groups showed similar pupillary dilation with tropicamide. Tropicamide also reduced the amplitude of the pupillary light reflex to a similar extent for all three groups. The amplitude of the pupillary light reflex with saline and tropicamide was abnormally small in both AD and PD, but AD and PD patients did not differ significantly from each other.

DISCUSSION: Tests involving tropicamide blockage of oculomotor functions are not reliable diagnostic tests for AD.

ACKNOWLEDGEMENT: NIMH grant MH56014 to the first author.




Yutaka Hirata1, Kazutsuna Yamaji2, Hiroyuki Sakai3 and Shiro Usui 3

1 Dept. of Electronic Engineering, Chubu Univ. College of Engineering.

1200 Matsumoto-cho Kasugai Aichi 487-8501 Japan.

2 Dept. of Ophthalmology, Wakayama Medical Univ.

3 Dept. of Information and Computer Sciences, Toyohashi Univ. of Technology.

Walter Cannon (1932) and Nobert Wiener (1947) mentioned in their book that the pupil was controlled so as to maintain a constant amount of light impinging on the retina. However, because the movable range of the pupil is about 2 to 8 mm (10^1.2 times in terms of the pupil area) their assumption is not plausible when considering the luminance change in external world which sometimes reaches 10^10 times.

Campbell and Gregory (1960) have demonstrated that the size of the human pupil is controlled to maximize visual acuity (VA) under a given light intensity. VA is determined by multiple factors including quality of the retinal image, and signal processing in retina and the cortex. Of these, the retinal image is closely related to the pupil since spatial frequency properties of the eye optics depend upon the pupil diameter (PD). Our hypothesis was that a function of the pupil is to optimize retinal image quality to maximize VA.

We tested this hypothesis by quantifying the retinal image quality in terms of its information capacity (IC). The IC is a quantity similar to Shannon’s entropy, and it is a logarithm of the number of images that the retinal photoreceptor array can reconstruct or distinguish. The IC is a function of PD, density of the retinal photoreceptor array, and characteristics of visual scene. We demonstrated by computer simulations that 1) there is an optimum PD which maximizes the IC under each light intensity when using the photoreceptor array with anatomically plausible density, and viewing natural scenes, 2) the optimum PDs are quantitatively in good agreement with experimental data of the light intensity–PD curves, 3) but these do not hold when using other photoreceptor arrays (either too sparse or too dense) or/and seeing other visual stimulus such as a white noise image.

These results support our hypothesis and suggest that the human pupil optimizes the IC of the retinal image under natural visual environments. Supported partially by the Japan science society grant, and the Naito research grant (YH).



Yeuk Fai Ho, Corina VandePol, Claudio Privitera, and Lawrence W. Stark

Neurology and Telerobotic Units, University of California, Berkeley


In this paper, a computational model for iris structure is proposed, implemented and visualized in matlab environment. The model employs a network of elements based on the Huxley sliding filament model for striated muscle. A simplified concatenation of these active muscle elements, based on symmetries of the iris structure is arranged radially and circumferentially to represent dilator and sphincter of the iris. Passive elasticities are similarly arranged.

Simulations successfully recreated overall non-linear 'length-tension' curve. It also reproduces experimental results from our lab of pupil deformations due to non-uniform application of pupil dilator drugs. These experiments produced oblately-shaped pupils that were captured on a video pupillometer.

A major extension of the model would be to convert the current version to a complete two-dimensional finite element model for greater homeomorphic correspondence. This actuator model is itself a component of an overall pupil model (consisting of both light and distance stimuli and their sensor operators, central nervous system noise generators, cross-connections to the triadic system) that exhibits non-linear behaviors matching real pupillary responses.




Randy Kardon M.D. Ph.D. and Oliver Bergamin M.D.

Department of Ophthalmology and Visual Science, Veterans Administration Hospital and University of Iowa, Iowa City, Iowa.

BACKGROUND: Variability of the pupil latency time and sensitivity to changes in input are the primary determinants of its clinical usefulness. It is uncertain whether higher sampling frequency during recording of the pupil light reflex would provide any greater accuracy in analyzing the latency time from the tracing of the pupil.

PURPOSE: The variability of the latency time of the pupil light reflex in normal subjects was compared under conditions of changing stimulus intensity and duration by simultaneous recording of the pupil at 60Hz and 1000Hz.

METHODS: Computerized binocular infrared pupillography was performed on 8 eyes of 4 healthy subjects, recording both pupils at 60 Hz and 1000 Hz, simultaneously. Each eye was stimulated 8 times for 50msec duration every 2.5 seconds at 5 stimulus intensities, increasing by 0.5 log units over a 2.0 log unit range. The effect of stimulus duration on latency time was also measured by stimuli of the following lengths: 16.8 msec, 50 msec, 100 msec, and 200 msec. The beginning of the pupil light reflex was defined as the time where maximal acceleration occurred. Three sampling frequencies were compared: 1) 60 Hz (video of bright pupil diameter), 2) 60 Hz velocity interpolated to 300 Hz using a cubic spline method, and 3) 1000 Hz (photodiode of bright pupil image).

RESULTS: the 1000 Hz recording gave a significantly smaller standard deviation (S.D.=4.94 msec) compared to the 60 Hz (S.D.=6.57 msec) and 300 Hz interpolation (S.D.=6.27 msec). Most of the variability in latency time from reaction to reaction appeared to be represented by biological variability of input and not measurement variability or output variability, since the right and left pupil latency times for each pupil light reflex showed less variability. The latency time became shorter in a highly predictable manner with increasing stimulus intensity and duration. The change in latency time was as great as 20 msec/0.5 log unit at the dimmest range of intensity and as short as 10 msec/0.5 log unit at the brightest intensity. The relationship between log intensity and latency time and stimulus duration and intensity time were highly correlated between the input to the right and left eyes of an individual, but varied between subjects.

CONCLUSIONS: The 1000Hz data collection is a promising tool in obtaining more precise latency measurement of the pupillary light reflex. In addition, the four-point interpolation of the 60 Hz velocity (to yield 300 Hz data) might improve the diagnostic power of data collected at 60 Hertz. Most of the variability of latency time appears to be due to biologic variability of the pupil light reflex. The non-linear change in latency with changes in intensity and duration appear to make it more sensitive over the dimmer and shorter duration stimulus range.

ACKNOWLEDGEMENT: This work was supported by the Swiss National Science Foundation, the Freiwillige Akademische Gesellschaft, Basel, Switzerland, the Roche Research Foundation, Basel, Switzerland, Zeiss-Humphrey Corporation, Dublin, California, and a Merit Review Award from the Veterans Administration, Washington, D.C. Dr. Kardon is the recipient of a Lew Wasserman Scholar Award (Research to Prevent Blindness)


Randy H. Kardon MD PhD, Leigh Bassman BS

Department of Ophthalmology & Visual Sciences, University of Iowa, Iowa. Veterans Hospital, Iowa City, Iowa.


PURPOSE: To determine how stimulus brightness can influence the ability to differentiate normal subjects from patients with asymmetric visual loss when contraction amplitude is the parameter being measured.

METHODS; Computerized infrared pupillography was performed on 50 normal subjects with normal visual field testing and 30 patients with asymmetric visual field loss. A light stimulus consisting of a 0.2 second duration light flash subtending a 30 degree radius visual field was alternately presented to the right and left eye every 3 seconds. Ten different stimulus light intensities were used, covering a range of 4.5 log units (3 stimulus cycles per test). The pupil contractions, recorded by computer, were plotted as a function of log stimulus intensity for each of the two eyes to ascertain patterns of damage caused by disease. At each intensity, the asymmetry in pupil contraction (right eye stimulation vs. left eye stimulation) was compared between the normal group and each patient using percentile rank. Patient scores were considered to be outside of normal limits if their inter-eye difference score exceeded the 95th percentile of normals.

RESULTS: There was a highly significant effect of stimulus brightness on the separation of the normal and patient groups. The most common pattern, caused by disease, showed that brighter stimulus intensities were able to more effectively identify a patient as lying outside of the 95th percentile.

CONCLUSION: A brighter stimulus light of short duration allows for greater ease in identifying a patient with unilateral or asymmetric visual dysfunction from normal subjects. This may have implications on how optic nerve and retinal disease affects the ability of ganglion cells to respond to light intensities near threshold (low neuronal firing rate) vs. supra-threshold (high firing rate). The objectivity of contraction amplitude difference between the two eyes, as measured by the pupil light reflex, may prove to have a significant ability to identify "suspect" patients as having truly abnormal optic nerve function better than the subjective measures employed by such tests as the Humphrey Visual Field. Classification of a patient’s pupil response as a percentile score compared to normal subjects provides a useful clinical measure of the extent of disease and change over time.

ACKNOWLEDGEMENT: This work was supported by a Merit Review Award from the Veterans Administration, Washington, D.C and Zeiss-Humphrey Corporation, Dublin, California. Dr. Kardon is the recipient of a Lew Wasserman Scholar Award (Research to Prevent Blindness, New York, NY)



Randy Kardon M.D. Ph.D., Oliver Bergamin M.D., and Susan Anderson B.S.

Department of Ophthalmology and Visual Science, Veterans Administration Hospital and University of Iowa, Iowa City, Iowa.

BACKGROUND: Permanent loss of axons in optic neuropathy causes thinning of the inner layer of the retina containing the ganglion cells. The amount of functional loss of the eye that is associated with a certain degree of thinning of the retinal nerve fiber layer is not known for different causes of optic neuropathy. This information may be useful in understanding how much damage is present and whether recovery of vision is possible in a patient.

PURPOSE: Structural damage of the retinal nerve fiber layer, measured by optical coherence tomography (OCT), functional damage, measured by automated perimetry and the relative afferent pupil defect (RAPD) were compared in patients with different causes of optic neuropathies.

METHODS: 55 patients with different causes of optic neuropathy were studied and included compression, ischemia, optic nerve hypoplasia, optic neuritis (recovered), and glaucoma. Each patient was tested by automated perimetry (Humphrey 24-2 SITA algorithm), and had asymmetry of pupillomotor input measured using neutral density filters to quantify the relative afferent pupil defect (RAPD). The retinal nerve fiber layer thickness was measured in each eye by optical coherence tomography. The inter-ocular asymmetry was compared in all subjects and in each diagnostic category.

RESULTS: A significant linear correlation was found between the interocular difference in nerve fiber layer thickness and the RAPD (correlation coefficient, R2=0.59). A similar linear correlation with the same slope was found comparing the nerve fiber layer asymmetry with the visual field asymmetry (R2=0.58). The correlation between the visual field asymmetry and the RAPD for the entire group of patients was also highly significant (R2=0.70). In compressive optic neuropathy and ischemic optic neuropathy the relationship of nerve fiber layer asymmetry with RAPD and with visual field asymmetry was similar. In recovered optic neuritis, the amount of RAPD per nerve fiber layer asymmetry was almost twice the functional loss measured by visual field asymmetry.

CONCLUSIONS: The overall structural-functional relationship of the retinal nerve fiber layer seems to be well correlated in patients with optic neuropathies in this study. Patients with visual dysfunction exceeding the amount predicted from their nerve fiber layer thickness may be showing a reversible component of visual loss or nerve fiber layer atrophy may not have had enough time to occur. Patients showing better visual function than predicted by the nerve fiber layer thickness, may be exhibiting a compensatory mechanism in the visual pathway. Correlation of structure with function in patients with optic neuropathy gives new information that can help classify disorders of the optic nerve, reveal the extent of reversible damage, and may point to compensatory mechanisms that in some cases help to maintain function in spite of structural loss of nerves.

ACKNOWLEDGEMENT: This work was supported by a Merit Review Award from the Veterans Administration, Washington, D.C. (Dr. Kardon), and a grant from Zeiss-Humphrey Corporation, Dublin, California. Dr. Kardon is the recipient of a Lew Wasserman Scholar Award (Research to Prevent Blindness, New York, NY). Dr. Bergamin is supported by the Swiss National Science Foundation, the Freiwillige Akademische Gesellschaft, Basel, Switzerland, and the Roche Research Foundation, Basel, Switzerland.

In vivo characterization of alpha-adrenoceptor subtypes mediating sympathetically evoked mydriasis in rats

Michael C. Koss, PhD and Yongxin Yu MD

Department of Cell Biology, University of Oklahoma College of Medicine, Oklahoma City, OK 73190


In previous work from this laboratory, we reported that the alpha-adrenoceptors of the cat iris dilator were pharmacologically distinct in that they could not be simply classified as either of the alpha1 or alpha2 subtypes. This study is an extension of that project and was undertaken to determine the alpha-adrenoceptors and subtypes that mediate sympathetically evoked mydriasis in rats.

Reproducible frequency-response curves of pupillary dilation were generated by electrical stimulation of the cervical preganglionic sympathetic nerve (1-32 Hz). Systemic administration of alpha-adrenoceptor antagonists, phenoxybenzamine (0.03-3.0 mg/kg), prazosin (0.01-1.0 mg/kg) and phentolamine (0.3-10 mg/kg), dose-dependently inhibited electrically evoked mydriasis. These drugs also antagonized norepinephrine (1.0-100 ug/kg, i.v.) induced pupillary dilation. The alpha-2 adrenoceptor antagonist, rauwolscine (0.1-1.0 mg/kg), shifted the frequency-response curve leftward, while having no effect on norepinephrine induced pupillary response. Intravenous administration of subtype selective alpha-adrenoceptor antagonists, WB4101 (0.1-1.0 mg/kg) and chlorethylclonidine (1.0-15 mg/kg) dose-dependently inhibited electrically evoked mydriasis. The alpha1D subtype selective antagonist, BMY 7378 (1.0 –3.0 mg/kg) had no effect on sympathetically evoked mydriasis but it significantly shifted the norepinephrine evoked pressor response curve rightward.

These results suggest that both alpha1A and alpha1B subtypes, but not alpha1D adrenoceptors, mediate sympathetically evoked mydriasis in rats. Alpha 2-adrenoceptors do not mediate postsynaptic mydriatic effect, and are located pre-junctionally, inhibiting norepinephrine release during sympathetic nerve stimulation.




Merlin D. Larson, MD

Department of Anesthesiology, University of California, San Francisco


INTRODUCTION: The presence or absence of mydriasis is often used as a rough guide to the adequacy of cardiopulmonary resuscitation following cardiac arrest. The most common explanation for this dilation is that inadequate delivery of oxygen to the midbrain leads to failure of neuronal activity within the pupilloconstrictor nucleus. However, there are scattered reports of successful resuscitation even after prolonged mydriasis. The brain dead organ harvest patient presents a unique opportunity to study pupillary activity in the absence of an intact midbrain and to examine the effects of asphyxia on the pupil. Because the midbrain is dead in these subjects, the pupil has no autonomic control and following aortic cross clamp, no blood can be delivered to the orbit.

METHOD: After institutional review and approval, 7 brain dead patients scheduled for organ harvest were studied. Pupil size, light reactivity, and reflex dilation were measured from the right eye every min for 5 min before and every 1-2 min for 20 ñ 30 min following aortic cross-clamp. Dapiprazole eye drops were instilled into the left eye at least one hour before cross clamp in three cases and pupillary measurements were intermittently taken before and after cross clamp.

RESULTS: Pupil size was stable before cross clamp, even during changes in infusion rates of large doses of dopamine and phenylephrine. Following cross clamp, the pupil dilated in all cases, reaching 8 mm in one case, after a preclamp diameter of 5.2 mm. Mean dilation was 2.2 +- 1.1 mm, Time to peak dilation was 3.9 +- 2.0 min and latency of dilation was 0.8 +- 0.5 min. After the pupil reached maximum size, the pupil gradually decreased in size over the following 20 min. Dapiprazole eye drops prevented the pupillary dilation in contralateral eye of the three cases in which it was used.

DISCUSSION: Because the pupilloconstrictor neurons are already dead in these subjects, the dilation which was observed must be due to factors other than loss of parasympathetic innervation of the pupillary sphincter. Hormonal (epinephrine) or drug delivery to the orbit must also be excluded because immediately following cross clamp, blood flow to the head stops. Because dapiprazole appears to block the dilation, it seems likely that the cause is either a short burst of neuronal activity in the peripheral system innervating the dilator muscle, or release of stored norepinephrine from the presynaptic terminals, as asphyxia intervenes.



Holger Lüdtke, Miriam Schelling, Barbara Wilhelm and Helmut Wilhelm

Pupil Research Group, University Eye Hospital, Dept. of Pathophysiology of Vision and Neuro-ophthalmology, Tübingen, Germany


PURPOSE: For the clinical diagnosis of Horner’s syndrome the cocaine test is commonly used. The aim of this study is to investigate an alternative method to diagnose Horner’s syndrome by measuring the redilation of the pupils with the Procyon pupillometer. Therefore Horner patients were compared with patients with physiological anisocoria.


Subjects. 11 patients with Horner’s syndrome and 7 patients with physiological anisocoria participated in the study. The diagnosis of Horner’s syndrome or physiological anisocoria was defined by the result of the cocaine test.

Pupillography. The Procyon, a binocular pupillograph (CMT International plc, London), was used for the pupillographic measurements. Three different stimuli (1.7, 17, 94 cd/m²) were presented for about 4s. Both pupils were measured for 6.4s with a frequency of 5Hz. The number of measurements per stimulus and patient was varying.

Evaluation of pupillograms. We analyzed two parameters, the maximum difference in absolute anisocoria in mm and the maximum difference in relative anisocoria in % (relative anisocoria means the absolute anisocoria divided by the mean of both pupil sizes). The minimal and maximal anisocoria were determined before and during the redilation by a special computer program.

Statistics. The measurements were averaged for each patient and stimulus condition. A Wilcoxon signed rank test was used to test for differences of absolute and relative anisocoria for the two groups. For diagnostic purposes a ROC curve was calculated for each parameter and stimulus. The area under the ROC curve is given.


The absolute and relative anisocoria show significant differences between the two groups except the absolute anisocoria at 1.7 cd/m² (table 1).

Table 1


1.7 cd/m²

17 cd/m²

94 cd/m²

absolute anisocoria

difference [mm]








relative anisocoria

difference [%]








The more relevant results are the ROC curves, described in table 2.

Table 2: Area under ROC curve


1.7 cd/m²

17 cd/m²

94 cd/m²

absolute anisocoria




relative anisocoria




DISCUSSION: There are clear differences concerning the redilation of pupils with Horner’s syndrome in comparison with physiological anisocoria. The ROC curve of the relative anisocoria at 94 cd/m² shows a sensitivity and specificity of more than 80% for a certain cut-off value. In this study the results of the cocaine test were used as a gold standard, although this standard has known limitations. Pupillography may be used as a substitute for the cocaine test to diagnose Horner’s syndrome. The best pupillographic results were obtained for the brightest stimulus.

ACKNOWLEDGEMENT: We thank Dan Taylor from CMT who put the Procyon pupillometer at our disposal.


Sharon L. Merritt, RN, MSN, EdD, Harold C. Schnyders, PhD, Minu Patel, MS and Maureen Smith, MS.

Center for Narcolepsy Research, University of Illinois at Chicago, Chicago, IL

Significant increases in EEG theta power (increases in theta activity in awake EEG) by Yoss pupil stage (percent of maximum diameter, Stage 1 "awake" = 95-100% to Stage 4 "sleepy" = 64-75%) have been found in sleep disorder and normal subjects when pupil diameter and EEG data were collected concurrently at 1400 hours (about 12 hours after the mid-sleep period). The purpose of this study was to determine if theta increased by pupil stage and time of day in a sample of normal subjects (9 F, 6 M, 21-58 years of age) who underwent one day of pupillographic Alertness Level Test (ALT) with concurrent EEG polysomnography (PSG) at 1000, 1200, 1400 and 1600 hours.

Pupil diameter and PSG data were collected concurrently at 256 Hz for 15 minutes with the subjects seated in a quiet, dark room. After cleaning, pupil diameter data were downsampled to 8 Hz and C3/A2 EEG data to 64 Hz. Power spectral density functions (FFT) were calculated on 2 sec epochs for minutes 5-11, and standardized for each individual by dividing each 2 sec epoch by the mean theta power obtained for the first minute of dark recording. The mean pupil diameter for each 2 sec EEG window was calculated, divided by the largest pupil diameter found for each subject during the fourth minute of dark recording, and classified according to the appropriate Yoss pupil stage. The 2 sec theta values for each Yoss stage were aggregated for each subject regardless of when they occurred during the 5-11 minute period.

ANOVA revealed that there were significant differences in theta power ratios by time of day (4 times) and pupil stage (4 stages) [F (15) = 17.66, p .000]. Post Hoc Least Significant Difference testing with Bonferroni correction (critical p .008 or <) revealed there were significant differences in the amount of theta activity both within and between testing times by pupil stage. By time of day statistically significant differences in theta power ratios were as follows: 1000 hr – Stage 4 > Stages 1-3; 1200 hr – Stage 4 > Stages 1-3; 1400 hr– Stage 4 > Stages 1 & 3; 1600 hr – Stage 2 < Stage 1 and > Stage 4. Between testing periods, the theta power ratio tended to be significantly greater for the afternoon testing periods and the sleepier pupil Stages 3 and 4.

These data replicate earlier findings which suggest that the pupillographic ALT, like the EEG-based physiologic sleepiness measures (the Multiple Sleep Latency Test and the Maintenance of Wakefulness Test), is a measure of increasing sleepiness in awake individuals, and is a sensitive measure of a lowered level of alertness at mid-afternoon in normal subjects.

ACKNOWLEDGEMENT: Research supported by Mr. J. A. Piscopo.





Sharon L. Merritt, RN, MSN, EdD, Harold C. Schnyders, PhD, Boris Vern, MD, PhD, Barbara Berger, RN, PhD, Anna Stone, MS, and Fang Fang, MS

Center for Narcolepsy Research, University of Illinois at Chicago, Chicago, IL

Modafinil is a novel Schedule IV wake-promoting compound that is FDA approved to treat the sleepiness of narcolepsy. This add-on study was conducted as part of a Cephalon-sponsored Phase IV open trial designed to test the safety and efficacy of switching narcoleptics from traditional stimulant medications (cylert, methylphenidate, or dextroamphetamine) to 200 or 400 mgs of modafinil [Provigil (TM)] once daily upon arising in the AM. The purpose of this Cephalon-approved add on study was to determine if changes in subjective sleepiness and pupillographic outcome measures could be detected in the sample (6 M, 5 F, 25-68 yrs of age). All subjects met the ICSD criteria for the diagnosis of narcolepsy. Data were collected over a period of 8 weeks when subjects were 1) on their traditional stimulant (screen visit), 2) after stimulant medication wash out that was verified by a urine drug screen (baseline) and 3) after 6 weeks of treatment (1 subject was on 200 mgs and 10 were on 400 mgs of modafinil at the 6 week visit).

On subjective measures the Friedman test with Bonferroni correction (p .017 or <) revealed significant differences in mean rank for ratings on the Epworth Sleepiness Scale (ESS) between the 3 time periods [Chi-Square (2) = 8.5, p .014]. An ESS score of 0-9 is considered normal, 10-15 moderately sleepy and >15 severely sleepy. At screen the mean (SD) rating was 15.5+/-4.4, baseline 17.6+/-5.3 and on modafinil 11.9+/-5.7. On the Functional Outcomes of Sleep Questionnaire (FOSQ) Friedman results approached significance [Chi-Square (2) = 6.7, p .035]. On the FOSQ a lower rating on items associated with the 5 scales denote more difficulty with performing everyday activities (1 = extreme difficulty to 4 = no difficulty, scores are obtained by multiplying the average rating for each scale by 5). The mean (SD) FOSQ scores for the respective time periods were 13.3+/-2.4, 12.5 +/-3.6 and 14.6+/-2.9.

The 15 minute pupillographic Alertness Level Test (ALT) was conducted at midafternoon with the subjects seated in a quiet, dark room at all three time periods. After cleaning, pupil diameter data were downsampled to 8 Hz and the pupillary unrest index (PUI) was calculated for the 4-11 minute time period of the ALT. Higher scores represent a greater amount of pupil unrest and increasing sleepiness in awake individuals. A mean of 4.9+/-1.98 has been reported by Wilhelm et al. for normally sleeping, alert individuals. Friedman PUI results revealed no significant differences between the 3 time periods [Chi-Square (2) = 2.2, p .34]. The respective mean (SD) for the 3 time periods were 8.5+/-4.9, 11+/-6.1 and 11.1+/-5.1. Finally, on ratings of medication effectiveness, subjects rated modafinil as equally effective as their prior stimulant in controlling sleepiness (Wilcoxon Z = -1.34, p .18), but indicated greater satisfaction with their prior stimulant (Wilcoxon Z = -2.05, p .04).

Various reasons that need further investigation but might explain the inconsistencies found between subjective ratings and pupillography findings include 1) the placebo effect, 2) subjective difficulties in assessing their level of sleepiness, 3) residual rebound sleepiness after taking traditional stimulants and 4) inadequate dosing with modafinil.

ACKNOWLEDGEMENT: Research supported by Mr. J. A. Piscopo.




Sancho Moro, Alister Harlow, Byron Lam*, M. Liu* and John L Barbur

Applied Vision Research Centre, City University, London, UK.

* Bascom Palmer Eye Institute1, 900 NW 17th Street, Miami, FL 33136;

Introduction: A new pupil perimetry test has been developed to extract small pupil response amplitudes and latencies in response to periodic modulation of stimulus luminance or chromatic contrast. The aim of this investigation was to establish how pupil responses to colour and light flux modulation alter in patients with acute, unilateral optic neuritis (ON) and the extent to which stimulation of the unaffected eye also yields abnormal responses.

Methods: A radial pattern was employed to stimulate one of 24 locations in the visual field. The stimulus contrast was modulated sinusoidally at 1 Hz for 10 cycles with a minimum of 8 averages. The coloured stimulus was generated by modulating its chromatic saturation against the neutral background. Each stimulus remained photopically isoluminant throughout and had zero rod contrast. 15 normal subjects and 17 patients with optic neuritis were investigated in this study.

Results: Pupil responses to light flux modulation yield significantly shorter latencies in the normal subject group. The difference in latency between colour and luminance triggered pupil responses can be as large as 80 ms with small inter-subject variability. Optic neuritis patients show significantly longer latencies and much reduced response amplitudes in the affected eye for both luminance and chromatic modulation.

Discussion: Both light flux and colour responses are affected in patients with optic neuritis with much increased latencies. Pupil colour responses are affected most. Some patients also show significantly smaller pupil response amplitudes and increased latencies in the unaffected eye. Recovery of visual acuity correlates well with recovery of the pupil colour response.





Claus Niemann, MD and Merlin D. Larson, MD

Department of Anesthesiology, University of California, San Francisco


INTRODUCTION: The human pupil constricts with low doses of intravenous opioids. A recent report suggests that pupillary constriction can be observed even in the absence of subjective symptoms such as sedation or euphoria. We hypothesized that a low dose (7.5 mcg/kg) of intravenous alfentanil would constrict the pupil but not produce any alteration of mood.

METHOD: After institutional approval and informed consent, patients scheduled for elective surgery were administered an intravenous bolus of alfentanil 7.5 mcg/kg. Pupil size and light reactivity were measured before and for every min for 10 min following injection. Each subject was instructed to report any subjective feelings of dizziness, sedation, euphoria, nausea, or light-headedness.

RESULTS: Thirteen patients were studied. All patients except one reported feeling some subjective sensation. The pupil constricted in all cases. Peak effect on pupillary constriction was at 3 min. At this time pupil size was 54% +- of preinjection size (P <.05). Recovery of initial size was observed in only one subject 10 min after injection. Because we did not exclude any subject because of disease condition or use of concurrent medication, some unusual pupillary responses were observed. Unusual pupillary responses following alfentanil were observed in 2 patients on dopamine 2 antagonists, 1 patient with advanced diabetes, and in one patient taking methadone for heroin addiction. One subject exhibited an exaggerated response to alfentanil and subsequently became apneic following a second small dose of opioid.

CONCLUSION: There is clearly a use for an objective measure of opioid effect that is not dependent upon a subjective report by the person receiving the drug. Other measures of opioid effect such as end tidal CO2 and CO2 response curves are relatively insensitive and are cumbersome to measure. The large mean pupillary constriction that we observed suggests that even smaller doses of alfentanil might have pupillary effects with no subjective symptoms. The unusual responses we observed in this small sample may have relevance during studies examining the pharmacodynamic properties of mu active opioids.





M.L. Rosenberg and R. Clarke

New Jersey Neuroscience Institute. Edison, New Jersey.

PURPOSE: We have observed evidence of interocular inhibition of the pupillary response in humans after alternating monocular stimulation. Data from sequential hemifield stimulation studies suggested that this inhibition takes place in the pretectal olivary nuclei, where inputs from each eye converge. This study evaluates the temporal characteristics of this inhibition by systematically changing the time of the interstimulus intervals.

METHODS: Six normal subjects were tested using monocular hemifield stimulations. The hemifield stimulations were given every two seconds and the interstimulus interval varied from 1.95 to 0.016 seconds. Such a cycle was repeated from 4-6 times per test and each test repeated 2-4 times for each subject at each interstimulus interval. Each hemifield was stimulated twice during each cycle, first preceded by stimulation of the contralateral nonhomonymous hemifield (control) and then preceded by stimulation of the contralateral homonymous hemifield (test). The degree of inhibition was taken as the amplitude of the test response divided by that of the control response. We then modelled the pattern of inhibition as a decreasing exponential.

RESULTS: Each hemifield of every subject showed greater inhibition for the smallest interstimulus interval compared to the longest. All but three of the 24 hemifields had data that fit well with a model using a second order exponential decay (each p < 0.05). The initial time constant ranged from 10 to 748 msec. with a median of 116 msec. (std 195 msec.).

CONCLUSION: All subjects showed evidence of inhibitory interactions between homonymous hemifields of the two eyes. In most hemifields the temporal characteristics of the inhibitory interactions could be represented with a second order decaying exponential function.




Hiroyuki Sakai1, Yasuo Ishigure2, Kazutsuna Yamaji3, Shiro Usui1

1. Department of Information & Computer Sciences, Toyohasi University of Technology, Toyohashi 441-8580, Japan

2. NTT Cyber Space Laboratories, NTT Corporation, Musashino 180-8585, Japan

3. Department of Ophthalmology, Wakayama Medical University, Wakayama 641-0012, Japan

INTRODUCTION: Three ocular functions, namely pupillary reflex, lens accommodation and eye movement, control the retinal image by functioning co-operatively. To investigate the mechanism of the co-operative behaviour, we developed a video-based recording system which can measure them simultaneously.

METHODS: In the system developed, the subject sits with his head in a chin-rest, and observes various visual stimulus through the dichroic mirror. The images of right and left eye are taken by CCD cameras respectively. Each eye image consists of two fields: one for a black pupil image (BPI) obtained by illuminating the iris, the other for a white pupil image (WPI) produced by the knife edge optics (KEO). They are captured by a personal computer and then processed in real-time. The BPI is binarized to detect the pupil area using an oval fitting. The pupil diameter and the eye position are measured from its major axis and center co-ordinate respectively. The torsional eye movement is calculated as the angle that takes the maximum value of correlation function of the iris pattern extracted. In the WPI, the fundus reflection is inclined in proportion to the accommodative state by the KEO. The lens accommodation is measured as a slope of the reflection, provided that the slope is normalized by square of pupil diameter. Because of the KEO, it is necessary to keep one eye on the line of sight for measuring accommodation. Therefore the convergence is measured from the other eye position.

RESULTS: The precision of the system is evaluated using schematic eyes.

The results are as follows: 1) 0.08mm for pupil diameter, 2) 0.3 and 0.2deg for horizontal and vertical eye position, 3) 0.4deg for torsion, and 4) 0.1D for accommodation.

By using this recording system, we are now studying the characteristics of triadic ocular function induced by an apparent depth cue based on linear perspective.

ACKNOWLEDGEMENT: A part of this study has been supported by NTT Cyber Space Laboratories, Japan.



Greg Siegle, PhD, 1,2 Stuart Steinhauer, PhD, 1, 2 Michael Thase, MD, 1 Cameron Carter, MD 1

1University of Pittsburgh Medical School,

2VA Pittsburgh

INTRODUCTION: Cognitive clinical research suggests that depressed individuals engage in elaborative processing of emotional information long after it is presented. Affective neuroscience suggests that there are distinct physiological and brain correlates of such emotional information processing styles, potentially involving increased activity in brain structures responsible for recognizing emotional information (e.g., the amygdala) and decreased activity in areas responsible for shutting them off (e.g., dorsolateral prefrontal cortex) in response to emotional information. We examined whether pupil dilation could be used as an on-line index of cognitive and physiological aspects of sustained processing of emotional information in depressed individuals, and whether this index was related to self-reported rumination.

METHOD: Depressed and control participants responded to a task comprised of alternating emotional processing (valence identification) trials and non-emotional processing (Sternberg memory) trials during two testing sessions. At the first session pupil dilation was measured during the tasks. The second involved functional magnetic resonance imaging (fMRI) assessment on a 1.5T GE scanner. Predictions for pupil dilation and fMRI were generated using a computational neural network model of attention to emotional information.

RESULTS: As predicted, compared to controls, depressed individuals displayed sustained pupil dilation in response to trials requiring emotional processing up to 15 seconds after an emotional word was presented, especially when it was personally relevant. fMRI data revealed that depressed individuals displayed sustained bilateral amygdala activity, particularly in response to negative words, as well as decreased dorsolateral prefrontal cortex (DLPFC) activity in response to negative information up to 30 seconds after words were presented. Together, sustained amygdala and DLPFC activity to negative vs. positive words predicted 52% of the variation in sustained pupil dilation to negative vs. positive words, after controlling for depression. Moreover, differences in both sustained pupil dilation and amygdala activity to negative vs. positive words were moderately correlated with self-reported depressive rumination.

DISCUSSION: Together, these data suggest that depressed individuals engage in sustained processing of emotional information, characterized by increased amygdala activity and decreased prefrontal cortical activity. This sustained processing is related to self-reported rumination. This research supports the use of pupil dilation as a non-invasive measure for detecting these cognitive and brain disruptions in emotional information processing.


ACKNOWLEDGEMENT: Supported by MH55762, MH30914, MH01306-05, MH16804, & the Veterans Administration



Lawrence W. Stark

Neurology and Telerobotics Unit

University of California, Berkeley

The pupil, because of its relative simplicity – single degree of freedom, non-invasive and easily controlled light stimulus, only a transparent surface between the iris muscle and the instrumentation, and classical reflex to light – has served as a paradigm for the introduction of quantitative control theory into physiology.

Electrical engineering contributed much. Instrumentation ranged from infra-red reflection photo-cells to video cameras. Very early on-line computers with special analog-to-digital and digital-to-analog ports enabled real-time data acquisition. The bioengineering instrumentation enabled both optical and electronic manoeuvres to produce open-loop and altered gain operational modes.

Mathematical analysis concentrated on use of the Fourier series, and thus structured the nature of the input-output experiments. These were small, steady-state signals amenable for averaging and continuous registration as opposed to more usual transient stimuli and responses. Following Norbert Wiener’s correlation and spectral analysis, early studies of both transfer function spectra and of noise outputs were accomplished. Later, these were extended to non-linear functional analysis, with white noise inputs and Volterra-Wiener series outputs.

The conceptual changes in approach were most important. Control theory required considerations of feedback, open-loop conditions, evaluations of gain and stability. Control theory also enabled one to understand the nature of oscillations as produced by instability conditions. Modelling may be considered as important conceptually. Linear lumped parameter models derived from transfer functions were structured so as to be "homeomorphic" models. This means that the elements of the model had one-to-one relationships with the anatomical and physiological elements of the system being studied. Eventually these models were run on digital computers using rapid matrix multiplication techniques.

Thus we were able to not only determine the pupil transfer function, but also to demonstrate how operational changes of gain produced oscillations. A series of noise studies also pioneered many physiological endeavours.



Abstract for After-Dinner Talk at 24th Pupil Colloquium


Professor Lawrence W. Stark

Neurology and Telerobotics Units

University of California at Berkeley


The "spirit of the times", swept through the 50's and 60's.

A belief that the new paradigm of cybernetics would inform the intellectual world received reinforcement by easy, personal contact with the brilliant pioneers of cybernetics: Wiener, Shannon, McCulloch.

This helped solidify our understanding of their most important contributions, and to inspire our own scientific work.





Stuart R. Steinhauer, PhD, Annette Kasparek, MS, Ruth Condray, PhD, Greg J.Siegle, PhD

Biometrics Research Program, VA Pittsburgh Healthcare System and WPIC, Dept. of Psychiatry, University of Pittsburgh School of Medicine;

Recording of pupil diameter during performance of demanding tasks offers a window on cognitive function. Central mechanisms invoked during continuing processing demands must involve brain regions having the ability to sustain attention and information. Frontal cortical regions have been implicated as subserving such functions , as indicated by neuropsychological deficits in patients with frontal damage, and by neuroimaging studies. Activation of frontal cortical regions has inhibitory effects on the parasympathetic center of the pupil, resulting in relaxation of the pupil sphincter. Change in pupil diameter was used to examine effects of sustained processing related to direct sympathetic stimulation of the dilator muscles, compared with dilation due to inhibition of the parasympathetic pathway. Pupil diameter was recorded in 22 (11 female) normal volunteers who performed arithmetic tasks which were difficult (subtract 7) and easy (add 1). Recordings were obtained both in darkness (to maximize the sympathetic component) and in light (to maximize dilation due to parasympathetic inhibition). Baseline diameter was marginally increased prior to verbalization during the difficult compared to easy task across lighting conditions. There was a significant interaction of task by lighting condition during the period of verbalization: pupil diameter was increased more for the hard relative to easy task in the light than in darkness. For a subgroup of 12 volunteers, recordings were repeated at three additional sessions at which temporary blockade or either the dilator or sphincter of the pupil was accomplished pharmacologically (using dipaprizole or tropicamide, respectively), or placebo was instilled. This allowed activity in the final autonomic pathways to be isolated. The smallest task effect was observed after blockade of the parasympathetic pathway (sphincter blockade), while an enhanced effect of task was observed after blockade of the sympathetic (dilator) pathway, especially for recordings in light rather than darkness. The findings indicate that the greatest contribution to pupillary change for the demanding task was associated with inhibition of the parasympathetic system. This is consistent with the notion of a significant active inhibitory input at the midbrain from descending cortical pathways.

ACKNOWLEDGEMENT: Supported by MH-55762 and the Medical Research Service of the Dept. of Veterans Affairs




E Szabadi, C M Bradshaw and P Bitsios

Division of Psychiatry, University of Nottingham

The acoustic startle response consists of the contraction of the skeletal and facial musculature in response to a sudden loud sound, and this response can be conveniently recorded in man by measuring the electromyographic (EMG) response of the orbicularis oculi muscle (startle eyeblink response). The EMG response can be enhanced by exposure to ‘threatening’ stimuli: for example, cues signalling the imminent delivery of an electric shock (‘fear-potentiated acoustic startle response’) [1]. We have shown recently that the pupillary light reflex is attenuated by exposure to cues signalling the threat of an electric shock (‘fear-inhibited light reflex’). We have also reported that the inhibition of the light reflex by threat can be antagonised by the anxiolytic drug diazepam [2]. In the experiments reported here, we compared the effects of the threat of electric shock and a chemical anxiogen, pentagastrin, on the two laboratory models of human anxiety.

In Experiment I, 12 male volunteers participated in three weekly sessions in which they received oral treatment with placebo, diazepam 5 mg and diazepam 10 mg, according to a balanced crossover double-blind design. One hour after ingestion of the treatments, miotic responses to light pulses and EMG responses of the orbicularis oculi muscle to sound pulses were elicited during alternating periods in which the threat of an electric shock (electrodes attached to the subject’s wrist) was present (THREAT) and absent (SAFE). The THREAT condition was associated with a significant increase in the amplitude of the EMG response, a significant reduction of the miotic response amplitude, and an increase in self-rated anxiety. Diazepam attenuated all these effects of THREAT. Diazepam did not affect the amplitude of the miotic response under the SAFE condition, but did suppress the EMG response under this condition.

In Experiment II, the effects of pentagastrin were compared with those of the cold pressor test, a procedure known to elicit sympathetic activation in 12 male volunteers. The experiment consisted of two parts, each comprising two sessions in which they received (1) pentagastrin (0.3 m g/kg, i.v.) and a control infusion (saline), and (2) cold pressor test (90 s hand immersion at 4 C) and a control immersion (37 C), using a balanced single-blind protocol. EMG responses of the orbicularis oculi to sound pulses (Part 1) and miotic responses to light pulses (Part 2) were recorded before, during and after the infusions and hand immersions. The startle response was not significantly affected by pentagastrin, but was reduced during the cold pressor test. Resting pupil diameter increased during both pentagastrin infusion and the cold pressor test, but neither procedure altered the amplitude of the light reflex. The anxiety induced by these treatments, unlike anxiety induced by the threat of electric shock, was not accompanied by potentiation of the startle response or reduction of the miotic response.

1. Grillon C et al., Psychophysiol, 28, 588-595.

2. Bitsios P et al., Psychopharmacol, 135, 93-98.



F. Thoss, B. Bartsch, H. Luedtke and D. Sandner

Carl Ludwig Institute of Physiology of the University Leipzig, Liebigstr.27, D-04103 Leipzig; Eye Clinics of the Universities Dresden and Tübingen

The light sensitivity of the pupil shows spontaneous oscillations. These oscillations could be due to the oscillations of the visual sensitivity we have described earlier. On the other hand, fluctuations of the sympathicovagal balance could contribute. As a first step of a discrimination between these possibilities, we investigated the influence of forced oscillations of the sympathicovagal balance on light sensitivity and pupil diameter. We generated these oscillations by a weak periodic work load and measured them by the heart rate. We expected to determine the temporal relations between connected parameters by means of a correlation analysis with the heart rate. In the analysis presented in 1999, we related the different parameters to the time shape of the workload itself. Following discussion with Harry Wyatt we realized that the individual physical and psychological load are very difficult to describe by a common mathematical relation and that therefore our analysis had a weak basis.

As the workload, our subjects had to pull the handle of a hand ergometer twenty times during 20 seconds in intervals of two minutes. The total duration of each experiment was 660 s. During this time the ECG, the pupil diameter, and the reactions of the pupil on constant, threshold near light stimuli were recorded. As well as the current diameter, the light reflex amplitudes and the heart rate were calculated from the records.

The results show that the investigated parameters are influenced in a characteristic manner. The heart rate and the pupil diameter increase, the light reflex amplitude decreases periodically as a result of the increase of the sympathetic and decrease of the parasympathetic activity due to the workload. The time shifts between heart rate and pupil diameter respectively reflex amplitude were determined by cross correlation. The medians for 11 subjects show a significant earlier decrease of light sensitivity (2.4 s) related to the increase of the heart rate. Between pupil dilatation and increase of the heart rate no significant time shift results.



Pupil studies in a patient with Parinaud’s Syndrome – NEW INSIGHTS INTO the organisation of central inhibition

Barbara Wilhelm1 and John L Barbur2

1 University Eye Hospital, Department of Physiology of Vision & Neuro-Ophthalmology, D-72076 Tübingen, Germany

2Applied Vision Research Centre, City University, Tait Building, Northampton Square, London EC1, England


INTRODUCTION: Parinaud’s Syndrome (PS) is associated with lesions to the dorsal midbrain/pretectal area. Classically the pupillary near reaction is preserved, while the pupil light reflex (PLR) is largely absent. Studies in patients with PS syndrome may provide answers to a number of questions concerning the organisation of pupil pathways. In this study we test for:

A. The existence of pupil response components that may be elicited directly through the Edinger-Westphal (EW) nuclei in the absence of signal processing in the pretectum, and

B. The hypothesis that the central sympathetic inhibition of the pupil is mediated directly via the EW nuclei and does not require the pretectum.

METHODS: Clinical and pupillographic examination of a 23 years old patient with Parinaud’s Syndrome who in childhood underwent several operations for a large pretectal tumor. The patient was examined at the Applied Vision Research Centre, City University London and at the Department of Pathophysiology of Vision and Neuro-ophthalmology, University Eye Hospital Tübingen. Different pupillographic set-ups were used with modification of stimulus intensity, duration, colour and pattern. In addition, the pupillographic sleepiness test was also applied during sleep deprivation to test for pupil sleepiness waves as a sign of preserved central sympathetic inhibition.


CONCLUSIONS: Transient weakening of the steady-state inhibition that controls the response of the pupil may provide the mechanism for the generation of the pupil response to gratings and chromatic stimuli. The sparing of these responses and the preservation of sleepiness waves in our PS patient suggest that the central inhibition of the pupil response is mediated directly via the EW nuclei and does not require the pretectum.

ACKNOWLEDGEMENT: This study was supported by The Royal Society, London




Helmut Wilhelm

Pupil Research Group, University Eye Hospital, Dept. of Pathophysiology of Vision and Neuro-ophthalmology, Tübingen, Germany


introduction: The tonic pupil or Adie’s syndrome is a frequent and mysterious disorder. In many patients it occurs as an idiopathic phenomenon, in about half of the cases associated with hypo- or areflexia. This presentation comprises two cases that might add some new information.

Case 1: The first case is a 20-year old man with metastatic osteosarcoma thigh diagnosed at the age of 18 and treated by amputation and chemotherapy. After development of metastasis after chemotherapy he was treated with Gd2 antibodies. Those are combined from humans and mice and directed against gangliosides. During the first iv treatment this patient realised worsening of his near vision and the nurse saw his pupils maximally dilated and unreactive to light. Because of additional renal problems the ophthalmological examination had to been delayed until two days after this first treatment. He showed maximally dilated pupils (8 mm), not reacting to light and near. Accommodation was completely paretic. His pupils constricted strongly to 0.1% pilocarpine (2.75 mm) and dilated slightly to pholedrine (8.75 mm) indicating a functional sympathetic innervation. One year later he showed partially restored accommodation and pupillary near reaction and did not longer need reading glasses.

This case gives a hint that auto-antibodies may play a role in the development of a tonic pupil.

Case 2: The second case deals with areflexia and tonic pupil. Adie summarised under his eponym not only the combination of a tonic pupil and areflexia but also isolated tonic pupil and areflexia. The latter was later put into question. Usually the pupillary abnormality is detected at the same time as the reflex abnormality. We report a 55 year old man who presented with the typical symptoms of a recently acquired tonic pupil (dilated pupil not reactive to light or near vision but reactive to 0.1 % pilocarpine). He reported that he had been hospitalised in the neurology department because complete loss of deep tendon reflexes had been found during a routine examination. There were no other pathologic findings. Pupils were described as being normal at this time. Adie may have been right: Areflexia can precede pupillotonia for many years.




Helmut Wilhelm

Pupil Research Group, University Eye Hospital, Dept. of Pathophysiology of Vision and Neuro-ophthalmology, Tübingen, Germany


INTRODUCTION: Intoxication by the toxin of Clostridium botulinum is rare and difficult to diagnose if occurring as a single case.

CASE REPORT: A 30 years old male patient noticed dilated pupils and blurred vision. At examination the pupils reacted sluggish to light, not at all to near and accommodation was completely absent. Additionally, after request, he reported gastrointestinal pain and diarrhoea after having eaten vacuumed ham and conserved meat. Lab testing proved botulinus intoxication. The patient survived without further complications. We were able to obtain complete pupillary and accommodation testing including pupillography, pilocarpine 0.1% test and infrared video retinoscopy.

RESULTS: Accommodation paralysis in combination with an absent near reaction of the pupil was the dominating sign. Pupils were somewhat dilated but light reaction was preserved. Both pupils constricted clearly to pilocarpine 0.1%.

DISCUSSION: Botulism is a rare occasion where near reaction of the pupil and accommodation are more impaired than light reaction, probably because of specific high affinity of the toxic protein to the ciliary ganglia cells serving accommodation. The "Near-light dissociation" of the pupil is hardly to understand with our knowledge of the pupillary innervation. There is little doubt that the site of the lesion has to be the ciliary ganglion. However, if it is possible to block the near reaction completely with a partial preserved light reaction, the final common pathway of the light and near reaction has to be reconsidered.



A Brief Guide for Pupillographers.

Helmut Wilhelm

Some of the information that may be required when reporting studies of pupillary movements; being notes from a discussion held on 13.8.99, at the 23rd Pupil Colloquium, Nottingham, England


Subjects Record any medication the subject is taking (including smoking, caffeine, and, if relevant, alcohol, and especially topical medications). Also note the subject’s age, gender and iris colour.

Equipment Describe the methods used to document the pupils: the type of pupillometer used, how it works, and its spatial and temporal resolution.

Light Stimulus Describe the light stimuli and their sequencing with care. Include duration, intensity, size (angular) including distances, colour or other relevant characteristics, location in the visual field, interstimulus interval, etc

Other Stimuli Provide a similarly complete description of any other physical stimulus (eg: "an isolated 100 msec burst of white noise at 90 dB to both ears via earphones")

Settings Record the time of day, degree of sleep deprivation if relevant, recording situation: background luminance in lab studies, adaptation time, how was fixation maintained ? was the fellow eye covered?,

Data Evaluation Be explicit about data handling: use of filtering if applied (artefact rejection), the number of measurements made and how they were evaluated (mean, averaging, algorithm of baseline, latency, amplitude, artefact determination, absolute values in mm or mm2, if possible.





Harry J. Wyatt, PhD

Schnurmacher Institute for Vision Research

SUNY State College of Optometry, 33 W. 42nd St., New York, NY 10036, USA


INTRODUCTION: Earlier work suggested an important role for aqueous convectional flow in anterior segment transport, including pharmacokinetics (J Oc Pharm Therapeutics, 1996; ARVO 1997; ARVO 2000; J Oc Pharm Therapeutics, in press). In the context of such flow, the present work considers drug entry into the anterior chamber, and the subsequent pathway to an anterior segment target.

METHODS: Convectional flow was modelled in a simplified anterior chamber. The model employed principles of fluid flow, and drew on experimental results and on a theoretical analysis of anterior chamber convection by Canning, Dewynne, Fitt & Greaney. Drugs were "applied" at a single corneal location in the model, and subsequent drug movement was treated as a sequence of steps, each step either by diffusion or by transport in the convectional flow ("bulk flow"). Model results were compared with experimental data from studies using focal application of mydriatics.

RESULTS: From "three-dimensional" pupil data (shape as a function of time), it is possible to determine a "primary latency" (earliest effect, occurs on the same side of the iris as the application) and a "secondary latency" (effect appearing on the side of the iris opposite to the application). Estimates of initial latency were in reasonable accord with experimental observations, but secondary latencies showed dramatic discrepancies. For example, with the head upright, modelling predicts that a drug should reach superior iris a few minutes after it reaches inferior iris; however, experimental focal applications to inferior limbus often produced near-zero effect in superior iris (phenylephrine) or an effect with much longer latency than predicted (tropicamide). This discrepancy generalized to some cases with head tilt; e.g., if an eye was tilted temporal-side-down, and phenylephrine was applied focally at the temporal limbus, little effect was observed in nasal iris. A related phenomenon was observed in the data for pupil elongation and center-displacement: "With-gravity" applications (e.g. at inferior limbus with the head upright) caused pupil elongation and pupil-center displacement; however, "against-gravity" applications (e.g. at superior limbus with the head upright) caused pupil elongation but not center displacement. In keeping with the latency results, this seemingly-paradoxical finding is due to drug arriving on both sides of the iris after "against-gravity" applications, but primarily on one side after "with-gravity" applications.

DISCUSSION: Inclusion of secretory flow in the pattern of anterior chamber fluid flow may help to reconcile modelling and experiment. Although secretory flow is slow compared to convectional flow, for some locations — in particular, near the pupil and near the angle — it can be significant. One implication of these considerations is that localized application of a drug (e.g. with a sustained release device) might be surprisingly ineffective at reaching certain target locations.





Kazutsuna Yamaji1, Takeshi Yoshitomi1, Shiro Usui2, Yoshitaka Ohnishi1

  1. Dept of Ophthalmology, Wakayama Medical University, Wakayama, Japan.
  2. Dept of Information & Computer Sciences, Toyohashi University of Technology, Aichi, Japan

INTRODUCTION: After the first work done by Lawrence Stark (1957), several mathematical models of the pupil have been developed. However, due to lack of the physiological evidence, each model uses assumptions for determining its structure or function. In this study, mechanical properties of the iris sphincter and dilator muscles were experimentally examined in order to develop bio-physiologically plausible model.

METHODS: Iris sphincter and dilator muscle specimens were prepared from male albino rabbits. Two kinds of experiments were performed: Exp.1: isometric contraction experiment and Exp.2: isotonic quick release experiment. In Exp.1, passive and active tensions at several steps of muscle length were recorded under the isometric condition. Active tension of the sphincter was induced by acetylcholine 10mM and of the dilator was induced by phenylephrine 0.1mM. In Exp.2, first, the muscle was fully contracted by drugs previously described, and then quick release was performed under different load conditions. Tension and displacement of the muscle were simultaneously recorded.

RESULTS: The tension-length relationship was obtained from the Exp.1. The result indicate that, for example, passive tension of the sphincteUse of uninitialized value in concatenation (.) or string at E:\listplex\SYSTEM\SCRIPTS\filearea.cgi line 455, line 777. r increases as a function of the length at which the active tension is maximized, however, passive tension of the dilator increases as a function of the relatively shorter muscle length. Comparison of these properties allows us to understand the contribution of each muscle in determining the pupil size at steady state. In Exp.2, load-extension and load-velocity relationships were obtained. The load-velocity curve could be described by Hill’s equation. Moreover, it was shown that viscous-like force, which is obtained from the load-velocity curve, is a function of not only the velocity but also the contractile force.

CONCLUSIONS: Tension-length, load-extension and load-velocity properties provide the information for parallel elastic, serial elastic and viscous component of the model, respectively. Our next step is to combine the each component to complete the model, and compare the model response with the pupillary response in vivo to confirm the accuracy of the model.