These are the abstracts submitted to the 23rd Pupil Colloquium, held in Nottingham, U.K. in August 1999. They are ordered alphabetically by first-named author.

The abstracts have been edited to fit a standard format, and the spelling of some words has been corrected, otherwise the abstracts are unaltered.


Peter Howarth

Department of Human Sciences

Loughborough University

August 1999


John L. Barbur and J. Alister Harlow

Applied Vision Research Centre, City University, Northampton Square, London, U.K.


Purpose We investigated how the pupil responds to coloured stimuli that are both photopically and scotopically isoluminant (D-isoluminant). We studied normal subjects and patients with damaged primary visual cortex (V1). Our aim was to test for pupil colour responses in the absence of V1 and to establish how these responses differ from those observed in normal vision.

Methods The subject viewed a uniform background field of luminance 24 cd/m2 and (x, y) - chromaticity co-ordinates 0.288, 0.334. Two directions of chromatic displacement that satisfy the D-isoluminant conditions were possible for the chosen background chromaticity and these resulted in either "red" or "green" hues. A number of chromatic saturations were employed and the stimuli were either presented as a flash of specified duration or modulated sinusoidally at 0.8 Hz for 8 cycles with a minimum of 16 averages. In addition to normal subjects, two hemianopes, both with unilateral damage to the primary visual cortex took part in this investigation. Pupil response amplitude and latency were measured by extracting the signal power and phase shift at the stimulus modulation frequency.

Results Pupil colour responses can be elicited in normal subjects. When large coloured stimuli are employed, significant pupil responses can also be elicited in hemianopes, but only for the "red" stimuli. The "green" stimuli trigger a smaller constriction of the pupil that always follows the offset of the flash. The results obtained with sinusoidal modulation reveal a strong signal at the modulation frequency in the blind hemifield with a latency of 0.52s for the red, and 1.14s for the green. A model for afterimage generation based on changes in the cellular adaptation properties of cone receptors is proposed. The model predicts the properties of chromatic afterimages, including the half-cycle increase in latency observed experimentally.

Conclusions The results show that significant pupil responses can be elicited with either red or green D-isoluminant stimuli in normal vision. Smaller responses remain in the absence of V1, but only for red stimuli. Red afterimages caused by green stimuli can however trigger a significant constriction of the pupil, even in the absence of V1. The new pupil-based technique makes it possible to reveal subcortical processing of chromatic stimuli, when the subjects are completely unconscious of the stimulus.



P. Bitsios, E. Szabadi and C.M. Bradshaw

Division of Psychiatry, University of Nottingham

We have reported that patients suffering from generalised anxiety disorder display attenuated pupillary light reflexes compared to age- and sex-matched healthy controls [1]. Following on from this observation, we examined whether anxiety acutely induced in the laboratory in healthy subjects by the threat of an electric shock has any effect on pupillary function. Our protocol was based on reports in the literature that the threat of an electric shock enhances the acoustic startle reflex response ("fear potentiated startle") [2]. Three experiments were conducted. Data were analyzed with ANOVA and individual comparisons (criterion of significance P<0.05).

In Experiment I, 12 subjects participated in a training session and an experimental session. In the experimental session six SAFE (electric shock is not possible) and six THREAT (electric shock is possible) blocks alternated, a mild electric shock was delivered before the fourth THREAT block. The pupillary light reflex was evoked three times in each block; pupil diameter was continuously recorded. At the end of each block subjects rated their anxiety using visual analogue scales. In the threat condition there was an increase in initial pupil diameter, a decrease in light reflex amplitude and an increase in anxiety ratings.

In Experiment II, 12 volunteers participated in three weekly sessions in which placebo, diazepam 5 mg or diazepam 10 mg were administered orally according to a double-blind balanced design. The design of each session was as for Experiment I, with the exception that four SAFE and four THREAT blocks were applied. Diazepam antagonized the effect of threat on light reflex amplitude in a dose-dependent fashion, and reduced subjectively rated anxiety and alertness. Diazepam had no effect on the pupillary measures in the SAFE condition or on initial pupil diameter in the THREAT condition.

Experiment III was similar to Experiment II, however, in this experiment the effect of diazepam was examined not only on the threat-evoked change in the pupillary light reflex ("fear-inhibited light reflex") but also on the threat-evoked increase in acoustic startle reflex (eye-blink) amplitude ("fear-potentiated startle"). Threat reduced the amplitude of the light reflex response, and increased the amplitude and reduced the latency of the startle reflex response; these changes were accompanied by increases in subjective anxiety. Diazepam antagonized the effect of threat on light reflex and startle response amplitudes and anxiety, in a dose-dependent fashion.

Our results show that the fear-inhibited light reflex, like the fear-potentiated startle reflex, is a valid laboratory paradigm of human anxiety. The neural mechanisms underlying the effects of threat on the two reflexes are likely to involve the amygdala [3] and, in the case of the fear-inhibited light reflex, the locus coeruleus [4].

  1. Bakes A et al 1990. Br J Clin Pharmacol 30, 377-381.
  2. Grillon C et al 1991. Psychophysiology 28, 588-595.
  3. Davis M et al 1993. Behav Brain Res 58, 175-198.
  4. Szabadi E and Bradshaw CM 1996. J Psychopharmacol 10 (Suppl 3), 6-18.



Michael Böttcher

Department of Clinical Pharmacology, Troponwerke GmbH & Co. KG, Neurather Ring 1, D-51063 Cologne, Germany

Introduction: Dynamic light evoked pupillography is a non-invasive method which can easily implemented in the basic phase I program of new CNS active compounds. BAY x 3702 is a full agonist at the 5-hydroxyltrytamine1A (5-HT1A) receptor subtype. The animal experiments revealed antidepressive and neuroprotective properties of BAY x 3702.

Method: Dynamic light evoked pupillographic measurements were used to describe the pharmacodynamic effect of BAY x 3702 in a series of dose escalation studies. In seven double-blind, randomised, placebo controlled studies, BAY x 3702 was administered to healthy male extensive metabolisers. The dosages (oral solutions) administered were 125, 250, 500, 750, 1000, 1500 and 2000 g. Blood samples were taken immediately before pupillographic measurements. The pupillographic target parameters were the initial pupil diameter and the reflex amplitude.

Results: After the administration of 500 g and higher doses of BAY x 3702 the initial pupil diameter showed significant miotic reaction 0.5 to 1.5 hours after drug intake. A maximum reduction of the pupil diameter (about 1.7 mm) was observed 0.5 hours after the administration of 1000, 1500 and 2000 g BAY x 3702. But in this dose range no dose-proportional reduction of initial pupil diameter was found. In contrast to that, a marked reduction of the reflex amplitude ( -0.25 mm) was found 0.5 and 1.5 hours after the administration of 2000 g BAY x 3702.

Concentration/effect-calculations were performed for the individual values of the initial diameter as well as for the reflex amplitude and the individual BAY x 3702 plasma concentrations, respectively. Both concentration/effect-relationships can be fitted to a sigmoidal curve. The concentration/effect-relationship between the initial pupil diameter and the plasma concentration shows a nearly linear decrease in the range between 4 to 35 g/l, whereas the fitted curve decreased significantly for plasma concentrations above than 13 g/l.

Conclusion: The administration of BAY x 3702 induced miotic pupil reaction. This confirmed the effects observed after the administration of other 5-HT1A agonists in humans (e.g. buspirone, ipsapirone) or after the administration of BAY x 3702 in rats. The fact that oral administration of BAY x 3702 affects initial pupil diameter and reflex amplitude with different sensitivity may either indicate the possibility of more than one receptor being involved or reflect an influence of peripheral and central 5-HT1A receptors on the pupil reaction.

Contact: by post


F.D. Bremner and S.E. Smith

Department of Neuro-ophthalmology, National Hospital for Neurology & Neurosurgery, Queen Square, London WC IN 3BG

Introduction Under standard conditions with the pupil dilated by tropicamide, the flash electroretinogram (FERG) shows a series of electrical responses to the light stimulus which are generated by different elements of the retinal circuitry. By leaving the pupil undilated, we have used the same recording set up to detect additional electrical components associated with the pupillary light reflex.

Methods Light stimuli (duration 500 ms, rate 0.2 Hz) were presented in Maxwellian view to one eye. Corneal surface recordings were made from both the stimulated eye and the consensual, non-stimulated eye using JET contact lenses.

Results At our 'standard' stimulus intensity bilateral corneo-positive 'p'-waves were observed in 56% of cases (n = 97 recordings, 18 eyes, 11 normal subjects) time-locked to the light stimulus. The ability to detect p-waves using this recording technique seemed to depend on the resting pupil diameter. When present, the mean p-wave latency was 204 ms (sd = 14 ms) and amplitude 204 m V (sd = 106 m V). This reflex response was abolished by topical 1% tropicamide but not by 10% phenylephrine. It was absent in patients with optic nerve lesions on the side of the stimulus (n = 4) or pupil-involving oculomotor nerve lesions on the side of the recording (n = 2). The amplitude of the p-wave was positively correlated with stimulus intensity but its latency showed an inverse relationship. Using simultaneous videographic pupillography the p-wave was found to preceed the constriction of the pupil by on average 43 ms (sd = l5 ms, n = 20 recordings).

Conclusions With undilated pupils, the pupillary light reflex can contribute to FERG recordings. The abolition of this response by tropicamide implies that it is generated by depolarisation of iris sphincter muscle fibres. The timing of this response implies an electro -mechanical delay in the iris sphincter muscle in man of around 40ms.

Contact: by post


K.D. Cocker, M.J. Moseley and A.R. Fielder.

Academic Unit of Ophthalmology, Imperial College of Science, Technology and Medicine, London, UK.

In this abstract, we present preliminary data obtained from a pupillometer designed specifically to measure the pupillary light reflex (PLR) in preterm infants. We know of no previous reports in which the dynamics of the PLR have been quantified in this population.

All infants studied, although born prematurely, were neurologically healthy and not in receipt of neuroactive medication. Infants were examined in an isolette above which the recording apparatus was mounted. Diffuse, wide-field, step luminance stimuli (Background Luminance: 6.7 Cd.m -2. Stimulus Luminance: 51 Cd.m -2) were provided by a 93 x 120mm flat-panel LCD screen positioned 33cm above the supine infant. Video-tape recordings of the pupil were obtained with a horizontally mounted infra-red (IR) CCTV camera and a 45 degree IR mirror in front of the stimulus screen. With the eye directly illuminated by an IR source, gentle manual retraction of the eyelids was applied as necessary. Stimulus presentation was initiated by means of a footswitch and a marker signal superimposed onto the video record 250ms before each stimulus. Recordings were later digitally sampled at 24 fps and the pupil diameter measured using a PC-based graphics package for each of the seventy-nine frames of the recording period.

Pupillary responses were recorded from all infants studied (gestational age range: 31 to 38 weeks). Response amplitudes ranged from 1.8 to 28% of pre-response diameter and tended to increase as a function of gestational age. Latencies ranged from 292 to 917ms and tended to decrease as a function of gestational age.

In conclusion, we have successfully demonstrated a technique for eliciting and recording pupillary responses to light from infants born prematurely. Initial data confirm previous qualitative reports that responses to light stimuli are present by at least 31 weeks gestational age (Robinson, J. and Fielder, A. R. Pupillary diameter and reaction to light in preterm neonates. Arch. Dis. Child.1990;65:35-38).

 This research was conducted with the approval of the local (Hammersmith Hospital) Research Ethics Committee and in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to applicable national and local Human-Subjects requirements. All co-authors have read the final, submitted version of the abstract and consented explicitly to having their names included in the list of authors


Alison Finlay, J. Alister Harlow and John L. Barbur

Applied Vision Research Centre, City University, Northampton Square, London, U.K.

Purpose We introduce a new method for the rapid, automatic extraction of pupil latency and response amplitude to sinusoidally modulated stimulus attributes such as light flux, contrast and chromatic saturation. A test based on this method provides information on both the dynamic and steady state relative afferent pupil defect and is suitable for clinical use. The results obtained using this technique have been compared to those previously reported for single flashes.

Method Pupil responses were measured using the P-SCAN system (J. Psychophysiol. 5: 223 - 239, 1991), in which stimuli were presented as a sinusoidal train of 10 cycles, at a temporal frequency of 1Hz, with Hanning windows at onset and offset. Different stimulus parameters were interleaved in the same experiment and pupil response parameters were extracted from the mean of several traces. The technique combines averaging and frequency analysis to increase signal to noise ratio. Three groups of stimuli were investigated, the attributes modulated being: 1. light flux change; 2. contrast at a number of spatial frequencies for grating stimuli; and 3. chromatic saturation of a uniform field that remained photopically and scotopically isoluminant. For all stimuli the target consisted of a 7 deg disc centred on fixation and surrounded by a uniform achromatic background subtending 29ox23o. The background luminance and space and time averaged target luminance remained constant at 16 cd m-2. Discrete Fourier Transform analysis of each mean response trace was carried out automatically. This provides the information needed to compute response amplitude and phase shift (at the modulation frequency), as well as measures of signal to noise ratio and response non-linearity.

Results The experimental findings suggest that the amplitude of pupil response increases monotonically with both contrast and chromatic saturation. Pupil response latencies, on the other hand, did not change with stimulus modulation depth for either luminance or chromatic saturation. For grating stimuli, the response amplitude forms a band-pass response with spatial frequency, peaking at ~ 5 c/o. The pupil response latency decreased at low spatial frequencies and increased at high spatial frequencies, but remained relatively constant between 1 and 10 c/ o. These latencies were of the same order as those elicited by coloured stimuli, but almost 100 ms longer than those elicited by light flux modulation.

Conclusion The new technique provides a rapid and accurate method for the automatic extraction of pupil response amplitudes and latencies to visual stimuli. The pattern of results reported here is consistent with single flash responses (Proc. R. Soc. Lond. B, 265 (1998) 2321 - 2325; Vision Research, 38 (1998) 3353-3358). The phase-lag latencies measured in this study were however longer than the latency to response onset measured in previous studies. Our findings suggest that latency estimates based on phase-shift measurements at low frequencies correlate more closely with the latency to the peak of the pupil constriction, as observed experimentally with single flashes, rather than to its onset response.


J.R. Jennings, M. van der Molen, and S.R. Steinhauer


Psychophysioloigcal "preparatory" responses may or may not depend on a focused expectation of when an expected stimulus will occur. Assuming a unitary preparatory state, positive correlations between psychophysiolgical measures, and between these measures and performance, would be predicted when variable strategies for preparation are induced by a non-aging foreperiod design.

Changes in pupillary diameter, heart rate, and the Contingent Negative Variation (CNV) were examined during trials in which the forperiod of a simple reaction time task was either fixed or unpredictable. In addition, separate trials were recorded in which the imperative stimulus was triggered by psychophysiological changes occurring spontaneously during the foreperiod. 32 male and female college-aged volunteers participated.

When a non-aging foreperiod was used to reduce expectancy, transient pre-stimulus psychophysiological responses were eliminated, but slow changes over the foreperiod -- slowing of heart rate, dilation of the pupil, and cortical surface negativity -- were not eliminated. Triggering of the imperative stimulus by physiological change (either heart rate slowing or decrease in pupil diameter) did not influence reaction time. Correlations among psychophysiological changes in the foreperiod and between these changes and RT were generally low: heart rate deceleration, pupillary dilation, and CNV exhibited independent preparatory changes. The results were more consistent with a multiprocess than a unitary process view of preparation.


Supported by NIMH grants 40418 and 55762. Details of this work are provided in Jennings et al, Psychophysiology, vol 35, pp. 90-98, 1998.


Constantin Kriegbaum, Barbara Wilhelm, Holger Lüdtke and Helmut Wilhelm

Department of Pathophysiology of Vision and Neuro-ophthalmology, Pupil Lab, Schleichstr. 12-16, D-72076 Tübingen,

Purpose The tonic pupil is characterized primarily by pupillary abnormalities. Results of studies investigating accommodation in tonic pupils are contradictive. The aim of this study is to demonstrate the differences between the affected and the normal eye concerning pupillary near response and accommodation.

Methods By means of real time (25 Hz) infrared video retinoscopy and simultaneous pupillography, refraction and pupil diameter were recorded during fixation of a target at distances of 2, 3, 4 and 5 diopters in 11 patients, aged between 15 and 38, with tonic pupil. Measurements were performed with monocular fixation. The diagnosis of tonic pupil was based on the clinical examination and pharmacological testing with pilocarpin (0.1%), i.e. cholinergic hypersensitivity.

Measurements of the affected eye and the normal eye were compared and analysed by paired t-test. Parameters of evaluation were the amplitude of accommodation, the initial pupil size and the pupillary near response.

Results Amplitudes of accommodation were smaller in tonic pupils (mean difference 0.78 dpt, t-ratio 2.54, DF 10, p<0.05). The initial pupil size showed no significant effect whereas the pupillary near response was significantly smaller in tonic pupils (mean difference 0.55 mm, t-ratio 1.48, DF 10, p<0.01).

Conclusions For the first time changes in pupil size and accommodation could be investigated simultaneously in a group of patients with tonic pupils. The expected effects of decreased accommodation and pupillary near response as a consequence of the parasympathetic deficit compared to the healthy eye were shown to be significant.

Certification: This study was conducted in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to requirements of the Ethics Committee of the University of Tübingen

Communicating author: Constantin Kriegbaum MD, Department of Pathophysiology of Vision and Neuro-ophthalmology, Pupil Lab, Schleichstr. 12-16, D-72076 Tübingen.


Merlin D. Larson, MD1, and Peter D. Berry, MB ChB 2

Department of Anesthesiology, University of California, San Francisco CA. USA (1 Professor, 2 Visiting Professor)

Introduction The effect of dopamine on the human pupil is poorly understood. D2 antagonists, such as phenothiazines and butyrophenones, constrict the pupil, whereas high doses of intravenous dopamine produce mydriasis. The mechanism of these effects are unknown, but a further understanding is of interest because these drugs are commonly used in the intensive care setting and can confound the pupillary signs of neurologically unstable patients.

Methods After institutional review approval and informed consent, 4 groups of 8 patients each were administered combined epidural-general anesthesia. Noxious stimulation of the C5 dermatome was provided by a tetanic electric current (60-70 mamps, 100 Hz, 3 second duration), after a stable level of epidural analgesia had been established with 3/8% bupivacaine. Pupillary reflex dilation (PRD) was calculated as the area above the baseline for the first 10 seconds following noxious stimulation. PRD was measured at 5 min intervals for 15 min before and 25 min after bolus administration of: saline 5 cc (Group 1), ondansetron 8 mg (Group 2), metoclopramide 20 mg (Group 3), and droperidol 2 mg (Group 4). Repeated analysis of variance was used to test for significant differences for each group compared to the saline control. P< 0.05 was considered significant.

Results Age and weight were not significantly different between the groups. Droperidol and metoclopramide both suppressed reflex dilation. Total change of PRD (mm-sec) following each drug were as follows:

1: Saline: -4.3 (+/-) 6.8,

2. Ondansetron: -2.6 (+/-) 9.4 (NS),

3. Droperidol: -27.0 (+/-) 15.6 (P = .00 1)

4. Metoclopramide: -19.2 (+/-) 15.1 (P< 0.02).

Discussion Both droperidol and metoclopramide are potent D2 antagonists and both interfered with PRD. Because metoclopramide also has 5-HT3 blocking activity, we also tested the effect of ondansetron, another 5-HT3 antagonist. Because this drug was without effect, we conclude that D2 antagonism interferes with reflex dilation in anesthetized patients. In cats it is known that norepinephrine inhibits the pupilloconstrictor nucleus via an alpha 2 adrenergic mechanism. Our studies have not shown a remarkable effect of alpha 1 adrenergic antagonists on reflex dilation and alpha 2 adrenergic agonists do not dilate the human pupil. We therefore propose that, in humans, dopamine may play a role in reflex pupillary dilation (PRD) via the D2 receptor


Holger Lüdtke, Miriam Schelling, Barbara Wilhelm and Helmut Wilhelm

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

Purpose In Horner patients the sympathetic pathway to the dilatator muscle is damaged. To mimick a Horner's syndrome we inhibited the sympathetic in healthy people and tried to show the pupillographic effect of this inhibition


Subjects. In 15 healthy people the sympathetic was inhibited pharmacologically in one eye (Pholedrin on the evening before the measurement, Dapiprazol 1 hour before the measurement in the morning).

Pupillography. The Procyon, a binocular pupillograph, was used for the pupillographic measurements. Three different stimuli (stimulus 1 [1.7 cd/m2], 2 [17 cd/m2], 3 (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 was varying.

Evaluation of pupillograms. The maximum difference in anisocoria (absolute in mm and relative in %) before and during the redilation was defined by a special computer program.

Statistics. An ANOVA with subjects as random factor and stimulus intensity as fixed factor was calculated for both types of anisocoria.

Results Two of 15 subjects were eliminated because of incomplete measurements. The remaining 13 subjects (age: median 24 years, range 23-50 years) showed a highly significant effect between the different light intensities. This effect was seen in absolute and relative anisocoria (absolute difference in anisocoria: F=17.6, DF=12, p<0.0001; relative difference in anisocoria: F=10.9, DF=12, p<0.0001). For larger intensities the difference in anisocoria was larger (absolute difference in anisocoria: stimulus 1: mean 0.67mm, 95% confidence interval (CI) 0.62-0.72mm; stimulus 2 mean 0.95mm, 95%CI 0.91-0.99mm; stimulus 3: mean 1.10mm, 95%CI 1.06-1.14mm; relative difference in anisocoria: stimulus 1: mean 8.5%, 95%CI 7.6-9.4%; stimulus 2 mean 14.3%, 95%CI 13.6-15.1%; stimulus 3: mean 18.3, 95%CI 17.5-19.0%).

Discussion The pharmacological inhibition of the sympathetic could be shown clearly by pupillography. In a further study in patients with Horner's syndrome the results of pupillography will be compared to those of commonly used cocaine testing. Possibilities to replace cocaine testing by pupillography will be proofed.

Certification: This study was conducted in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to requirements of the Ethics Committee of the University of Tübingen.

Communicating author: Holger Lüdtke, Department of Pathophysiology of Vision and Neuro-ophthalmology, Pupil Lab, Schleichstr. 12-16, D-72076 Tübingen.



Sharon L. Merritt, RN, EdD, Harold C. Schnyders, PhD, Minu Patel, MS, Qing Chen, BS and Glenn Clark, BS

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

As people become sleepy, the pupil has been observed to decrease in size. Yoss proposed a method for scaling level of alertness during 15 minutes of alertness level pupillography testing (ALT) by classifying pupil diameter as a percentage of maximum diameter (Stage 1 "awake" = 95-100% to Stage 4 "sleepy" = 64-75%). Intrusions of theta activity into the awake EEG have been associated with fluctuations in level of alertness. The purpose of this study was to determine if objectively defined sleepiness, reflected by changes in theta EEG activity, increased according to pupil stage in untreated narcoleptic (8 F & 8 M, mean age =30.3 +/-8.1) and untreated obstructive sleep apneic (OSA) (7 F & 9 M, mean age =36.9 +/-7.5) subjects diagnosed with PSG and MSLT studies, and normal control subjects (8 F & 8 M, mean age =30.4 +/-7.4) without any clinical evidence of a sleep disorder.

All subjects participated in the ALT during which pupil diameter and EEG recordings were conducted about 12 hours after the midsleep period. Pupil diameter and polysomnography data were collected concurrently at 256 Hz for 15 minutes. 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 4-11, and standardized for each subject 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 4-11 minute period.

ANOVA revealed that there were significant differences in theta power ratios by condition and pupil stage [F(11) = 31.12, p .000]. Post Hoc Least Significant Difference (LSD) testing revealed that there were significant differences (p .05 or <) both within and between subject groups by pupil stage. Within-groups statistically significant differences in theta power ratios were as follows:

narcoleptics - Stage 4 > Stage 3 > Stage 2 > Stage 1;

OSA - Stages 3 & 4 > Stages 2 > Stage 1;

controls - Stages 2 & 3 were > Stage 1.

Between groups the theta power ratio was significantly greater for narcoleptics and OSA subjects across Stages 1-4 compared to controls, and greater for narcoleptics compared to OSA subjects for Stages 3 & 4. The ALT is a measure of increasing sleepiness in awake individuals. Further research is warranted since pupillography is a convenient, objective measure with potential for wide application.

Research supported by NIH, R15 NR04030 and Mr. J. A. Piscopo.

Certification: This research was conducted in accordance with the human subject guidelines provided by the Declaration of Helsinki and approved by the Institutional Review Board of the University of Illinois at Chicago. All co-authors consent to the submission of this abstract.



Minoru Nakayama 1, Dr. Eng. and Yasutaka Shimizu 2, Dr. Eng.

1 The Center for Research and Development of Educational Technology (CRADLE), Tokyo Institute of Technology, Meguro-ku, Tokyo 152-8552, Japan 2 The Graduate School of Decision Science and Technology, Tokyo Institute of Technology, Meguro-ku, Tokyo 152-8552, Japan


It is well known that the pupil of the eye reacts to both brightness and emotional state including interest, enjoyment, and mental workload. A method for extracting emotional change from the pupillary response which is based on an artificial neural network has been developed to evaluate viewer's interest in visual information. This study examines the possibility of estimating viewer's interest and enjoyment of viewing movies by measuring the dynamic pupillary changes, blinking, and subjective interest responses.

The temporal response due to the subject's interest was measured by using an input device that has been developed. The subject was asked to input continuously the degree of his or her own interest while viewing a movie. Sixteen subjects viewed twenty one-minute movies which were edited to include five kinds of T.V. programs: sport, cultural, for children, entertainment and news. Each movie was divided into shots by use of a computer-based video summary system. Pupillary change and other response were evaluated for each temporal shot. The subject's impression for each movie was evaluated using questionnaires which were based on the semantial difference method after viewing.

The three factors, "Evaluation", "Potency", and "Activity", were extracted from the subject's responses on the questionnaires. The results showed, in dynamic evaluation, that there is a relationship between pupillary changes and the subjective interest response. In evaluating pupil size, there is a significant difference in pupil size between the higher and the lower shot for the degree of subject interest response in each kind of movies. However, there is no significant difference in blinking between them. In evaluating impressions from movies, there is a significant difference in pupil size between the higher and the lower group for the factor "Potency," and a significant difference in blinking between the higher and the lower group for the factor "Activity". Also, there is a significant correlation between the subjective interest response and all three of these factors.

Partially supported by the Grant-in-aid for scientific research (B)(2)-10558020 of the Japanese Ministry of Education


M.I.K. Norrish BSc 1 and J.D. Mollon PhD, FRS 2

1 Department of Psychology (HSS), University of Coventry, Priory Street, Coventry, CV1 5FB, UK 2 Department of Experimental Psychology, University of Cambridge, Downing Street, Cambridge, CB2 3EB, UK

The spectral sensitivity of the pupillomotor system has been examined in many studies. In recent years it has been suggested that different components of the pupillary light response depend on different mechanisms and the possibility arises that these mechanisms may have different spectral sensitivities. We used a signal cancellation technique to examine the spectral sensitivity of the transient component of the response.

A maxwellian-view optical system was used to present an adapting background field, a test stimulus and a reference stimulus. The short-wavelength background was used to limit rod intrusion. The test and reference stimuli were modulated sinusoidally in counter-phase at 2.35 Hz. The wavelength of the reference stimulus was 550nm throughout the experiment. The pupillary responses were recorded for eight test stimulus intensities at each test wavelength. There will be a response null and phase reversal when the potencies of the two stimuli are the same. The action spectrum can be calculated from the radiance of the test stimulus required at each wavelength to produce a response null. Psychophysical low-frequency heterochromatic flicker photometry settings were also made on the same apparatus to provide a comparison with the pupillary responses. Three colour normals and four dichromats, two protanopes and two deuteranopes, were tested.

It was found that there are significant differences between the psychophysical and pupillary action spectra for colour normals, but not for the dichromats. In colour normals the sensitivity is lower for the pupil at wavelengths above 580nm. There are no statistical differences between the pupillary action spectra of the colour normals and of the protanopes, although for the psychophysical responses the expected differences remain. The pupillary action spectrum for colour normals is well modelled not by the photopic luminosity function but by the Stockman, MacLeod and Johnson m-cone template. The possibility that the m-cones alone can mediate the transient component of the pupillary light response must therefore be considered, so too must the implications of this for the labelling of M and L cones in the retina.

This abstract complies with the ethical guidelines for human subjects provided by the declaration of Helsinki. Both co-authors have seen and agreed to the above submitted abstract.


M.A. Phillips, E. Szabadi and C.M. Bradshaw

Division of Psychiatry, University of Nottingham, UK

Central 5-hydroxytryptaminergic (5HTergic, serotonergic) neurones have been implicated in autonomic regulation, although their exact role is not fully understood [1]. Drugs interacting with 5-HT receptors have been reported to affect pupil diameter: both antagonists of 5HT2 receptors [2] and agonists of 5HT1A receptors [3] cause miosis.

We examined the effects of two 5HT1A receptor agonists on pupil diameter at two luminance levels (darkness and 32 Cd m-2). The drugs used were buspirone, an anxiolytic drug with partial agonistic activity at 5HT1A receptors and lesopitron, a recently developed anxiolytic drug with full agonistic activity at 5HT1A receptors. Pupil diameter was monitored with a binocular infrared television pupillometer before and after the administration of treatment for 4 h at 20 min intervals. Two experiments were conducted.

In Experiment I, fourteen healthy male volunteers participated in seven weekly sessions, each associated with the ingestion of one capsule (buspirone 5, 10 and 20 mg, lesopitron 10, 20 and 40 mg, and placebo), according to a double-blind balanced, cross-over design. Both buspirone and lesopitron tended to decrease pupil diameter. In darkness, only the highest dose of buspirone (20 mg) caused a miosis that was statistically significant. However, at the luminance level of 32 Cd m-2 buspirone 10 and 20 mg evoked statistically significant miotic effects, as did the highest dose of lesopitron (40 mg). The miotic effect was significantly greater at 32 Cd m-2 than in darkness after each dose of buspirone and the highest dose (40 mg) of lesopitron.

In Experiment II, pupil diameter and oral temperature were monitored with an electronic thermometer at 40 min intervals. 20 healthy male volunteers participated in two weekly sessions, each associated with the sublingual application of 100 l hydroalcoholic solution (lesopitron 20 mg, placebo), according to a double-blind balanced cross-over design. Lesopitron caused a significant miosis both in darkness and at the luminance level of 32 Cd m-2 ; the miosis was greater at 32 Cd m-2 than in darkness. Lesopitron tended to decrease oral temperature; this effect however was not statistically significant. The greater effectiveness on the pupil of lesopitron administered sublingually in a solution indicates the importance of first-pass metabolism in reducing the effectiveness of the drug when administered by the mouth. The miosis observed in both experiments may be due to either a sympatholytic or a parasympathomimetic effect of the drugs, or both. The light-dependence of the miosis indicates that the 5-HT1A receptor agonists can modulate the light reflex, possibly via the noradrenergic control of central cholinergic neurones in the Edinger-Westphal nucleus.

  1. Ramage AG and Fozard JR 1987. Eur J Pharmac 138, 179-191
  2. Fanciullacci M et al 1995. Clin Pharmac Exp Ther 57, 349-355
  3. Millson DS et al. 1991. Br J Clin Pharmacol 32, 447-454
  4. Cleare AJ et al 1998 Int Clin Psychopharmacol 13, 23-32



M.A. Phillips, E. Szabadi and C.M. Bradshaw

Division of Psychiatry, University of Nottingham, UK

Spontaneous fluctuations of pupil size in darkness are characteristic of a person's level of vigilance: these fluctuations are accentuated when subjects are drowsy [1,2]. Recently the "pupillographic sleepiness test" (PST) has been developed for the recording and quantification of "pupillary fatigue waves" [3]. This test has already been used to detect daytime sleepiness both in healthy sleep-deprived subjects [4] and in patients suffering from sleep disorders [5]. In this study, we used the PST to compare the sedative effects of three antidepressant drugs: amitriptyline, a tricyclic antidepressant, fluvoxamine, a selective serotonin reuptake inhibitor and reboxetine, a selective noradrenaline reuptake inhibitor. Whilst amitriptyline is known to be sedative, both fluvoxamine and reboxetine are novel antidepressants with little clinical sedative potential [6].

Thirteen healthy males (19-30 years) participated in 4 weekly sessions, each associated with one oral drug condition (placebo, fluvoxamine 100 mg, reboxetine 4 mg and amitriptyline 100 mg), according to a balanced double-blind design. Pupil diameter was recorded continuously over 11 min in darkness using a dedicated monocular television pupillometer (AMTech Weinheim Germany). Average pupil diameter, power of pupil diameter fluctuations (obtained by Fast Fourier transformation), and the pupillary unrest index (PUI), a measure of cumulative changes in pupil size, were computed [3]. Subjective "alertness", "anxiety" and "contentedness" were rated using visual analogue scales [7]. Measurements were carried out 30 min after arrival in the laboratory and 3 h after drug ingestion. Data were analysed by ANOVA followed by Dunnett's corrected t-test (criterion of significance P< 0.05).

Pupil diameter was increased by reboxetine but was unaffected by the two other antidepressants. Amitriptyline increased pupillary fatigue waves evidenced by an increase in the power of fluctuations, consistent with the sedative property of the drug. Neither reboxetine nor fluvoxamine had any effect on the power of fluctuations or PUI. The effect of amitriptyline upon pupillary fatigue waves was parallelled by decreases in "alertness" and "contentedness" self-ratings by the subjects; both amitriptyline and reboxetine caused a reduction in "anxiety" ratings. The PST seems to be a suitable method for the detection of drug-induced changes in the level of arousal.

  1. Lowenstein O et al 1963. Invest. Ophthalmol 2, 138-157
  2. Yoss R E et al 1970. Neurology 20, 545-554

3. Lüdtke H et al. 1998. Vision Res 38, 2889-2896

  1. Wilhelm B et al 1998a. Sleep 21, 258-265
  2. Wilhelm H et al 1998b. Graefe's Arch Clin Exp Ophthalmol 236, 725-729
  3. Szabadi E & Bradshaw C M 1995. In Seminars in Clinical Psychopharmacology ed. King, D J London: Gaskell
  4. Bond A J and Lader M H 1974. Br J Med Psychol 47, 211-218


F.G. Priemer1 and A. Gappmaier2

1University of Munich, Germany, Institute of Legal Medicine 2University of Salzburg, Austria, Institute of Legal Medicine

The effect of THC (Cannabis) to the pupillary reaction is discussed controversially. In some police training manuals wide and sluggish pupils are described as valuable signs of the influence of THC. On the other hand some authors had measured the pupilsize during THC-impairment and had observed reduced pupillary diameters.

13 healthy volunteers inhaled a joint containing 40 mg THC within 10 min. The experiments were carried out before the THC consumption, as well as 40 min and 80 min after smoking. Pupil size, and the pupillary light reflex after a short light-stimulus (4 different intensities), were registered by the IR-videopupillography-device CIP 8.00 (AMTech, Weinheim).

Latency of pupillary constriction remained unchanged by THC. The pupil diameter tended to decrease during THC-impairment as well as the amplitude of pupillary constriction after the light stimulus. In accordance with Hepler et al. (Am. J. Ophth. 74: 1185-90, 1972) THC caused effects like reduced amplitude of pupillary constriction, increased heart-rate and decreased lacrimation as well as salivation can be characterised as atropine-like or parasympatholytic effects.


Dr. F.G. Priemer

Institute of Legal Medicine, University of Munich, Frauenlobstrasse 7a,

D- 80337 Munich,



Peter Roessger

Computer Aided Animation (CAA) GmbH, Germany

Driver fatigue is one of the major research topics in the human factors departments of the car industry. The idea that driving itself fatigues the driver and the idea of predicting the moment a driver starts to sleep, have occurred in many variations during the last years. The major problem is to find a useful, valid parameter of measuring fatigue. Pupillomotorical fatigue waves could serve as such a parameter in two different ways. On-line measurements inside the car to predict the moment the driver falls asleep, and off line to measure fatigue changes induced by driving seem to be possible.

CAA GmbH plans, evolves, and visualises in car equipment for major German car manufacturers like Audi, BMW, Opel and Volkswagen. The integration of ergonomics specialists into the developmental process makes it possible to optimise the interface concepts under human factors criteria. The reduction of driver fatigue by using human factors knowledge in all consequences as well as the measurement of driver fatigue are major focuses of the developmental work at CAA.

In an experiment pupillomotorical fatigue waves were used to measure changes in fatigue induced by driving all day. All subjects made three journeys with three different luxury sedans. Each of the cars had a different character (big & luxury, high-tech and sportive).

Major results are:

Implications from these findings on theories of driver fatigue will be discussed.

In a second part of the presentation problems of the application of fatigue waves on the road will be discussed. An assessment of the chances to use fatigue waves to measure driver fatigue on-line will be given.



M. Rosenberg, M. Sobol, R. Clarke and N. Borden

N.J. Neuroscience Institute, at Seton Hall University

Purpose To evaluate the effect on the pupillary light reflex in patients with homonymous hemianopias.

Methods Pupil responses to a dim hemianopic were measured in 28 normal and 13 hemianopic patients with occipital lesions. An algebraic model of the pupil light reflex pathway was used to calculate the percentage of crossed and uncrossed fibers from each eye in the chiasm and from each pretectal nucleus in the midbrain. We derived a measure of the hemianopic pupillary defect (HPD) by taking the ratio of the percentage of crossed fibers from the contralateral eye to that of the ipsilateral eye. HPD's in symmetric crossings would be 1 and in a total hemianopic defect would be zero. Patients were divided into two groups based on MRI's. Group 1 consisted of 6 patients with additional midbrain or optic tract involvement while the 7 patients in group 1 appeared to have isolated occipital cortex involvement. Most defects were total hemianopias and were similar in the two groups.

Results The mean HPD was 1.0 (std = .11) in normals. This differed significantly from both group 1 (0.4 std .25 p < 0.003 ) and group 2 (0.72 std = 0.26, p < 0.05), but group 1 and 2 were not different from one another (p = .052). Midbrain calculations suggested abnormalities in 4/6 patients of group1 but only 1/7 in group 2. Dividing the patients into acute and chronic lesions showed a much more striking difference between the HPD's of the two groups ( means 0.39 & 0.85 p = 0.005).

Conclusions The HPD decreased as the chance of an optic tract or midbrain problem increased, independent of visual field changes. This suggests that associated optic tract/midbrain damage is common, especially in acute lesions, and explains most of the pupillary changes seen in hemianopic patients. A mild pupillary effect from isolated cortical damage cannot be ruled out.


Rüdiger Schmid, Per Ceurremans, Holger Lüdtke, Barbara Wilhelm and Helmut Wilhelm

University Eye Hospital, Dept of Pathophysiology of Vision and Neuro-ophthalmology D-72076 Tübingen, Schleichstr. 12-16, Germany

Purpose The pupil reaction to focal light stimuli can be used for an objective evaluation of the visual field. Obtaining reliable PLR for all subjects may be dependent on conditions of stimulation and measurement. We tried to evaluate the best stimulus conditions to obtain reproducible pupil reactions for each stimulus location for each subject tested.

Methods We performed pupil perimetry in 26 young normals. An infrared video-pupillography device was used for PLR registration and the stimuli were generated on a computer screen at a distance of 30 cm to the subject's eye. We presented 25 stimuli within the central 21 degree visual field and ran four trials of a fixed sequence of stimuli (stimulus duration 200 ms, interval 2.5 sec). Three different stimulus sizes were used (2, 3 and 4 degrees of diameter) with two different luminances (26 and 54 cd/m2) on a background of 1 cd/m2 (i.e. 3 dB difference between both luminances).

Results Pupil responses increased with increasing stimulus size and luminance. Stimulus size and luminance amplified the effect of each other. Within the central 15 degrees, a PLR could be evoked with most stimulus intensities and sizes, whereas at 20 degrees of eccentricity, in several individuals, the dim and small stimuli did not elicit a response. The slope of retinal sensitivity from the centre to the periphery was steeper with larger stimuli than with smaller stimuli, and with brighter stimuli compared to dimmer stimuli. The amount of PLR amplitudes was very similar for the 3 degree, 26 cd/m2 stimuli and the 2 degree, 54 cd/m2 stimuli.

Conclusions In pupil perimetry, light stimuli need to be more intense than in visual threshold perimetry. Stimulation over threshold often risks causing stray light and thus blur the focal retinal response. In our study, we obtained focal pupil responses with all stimuli used. The very small and dim stimuli however proved not to be suitable to obtain a PLR at every testing location for all subjects tested.

 Certification: This study was conducted in accordance with the guidelines provided by the declaration of Helsinki and in adherence to requirements of the Ethics Committee of the University of Tübingen. Communicating Author: Rüdiger Schmid, University Eye Hospital, Dept of Pathophysiology of Vision and Neuro-ophthalmology, D-72076 Tübingen, Germany.

Granted by fortuene F. 1222350.



S.E. Smith and F.D. Bremner

Department of Neuro-ophthalmology, National Hospital for Neurology & Neurosurgery, Queen Square, London WC IN 3BG

Introduction Pupil involvement is common in some autonomic neuropathies (e.g. diabetes mellitus) but rare in others (e.g. multiple system atrophy). We present a case of idiopathic pure sub-acute autonomic failure in which both the sympathetic and the parasympathetic innervation of the pupil were affected.

Case report A 29 year old emmetropic man presented with a 3.5 year history of sequential pupil enlargement and accommodative paresis making it impossible for him to read without glasses. Over the same period he had developed a dry mouth, patchy anhidrosis, severe postural hypotension, difficulty emptying his bladder and impotence. His past history included eczema, psoriasis, lactose intolerance and ulcerative colitis (treated by total colectomy). The pupils were round, central and of normal appearance. Infra-red video-pupillography findings were as follows:



normal range


Diameter (dark)

5. 1 mm

5.6 - 8.4 mm

5.2 mm

Diameter (light)

4.6 mm


4.8 mm

Light reflex ampl.

0.4 mm

>1.5 mm

0.3 mm

Near response

0 mm

>1.0 mm

0 mm

0. 1 % Pilocarpine*

2.3 mm

<0.5 mm

2.6 mm

1% Phenylephrine

1.3 mm

<0.5 mm

1.2 mm

*0.1% Pilocarpine also induced 3 dioptres of myopia

No cause was found for his sub-acute pan-autonomic failure, but there has been some spontaneous improvement in autonomic function and two years later he is able to read without glasses. Pupillography confirms a gradual recovery with increase of the light reflex amplitudes to 0.65mm/0.95mm in the right/left eyes respectively.

Comment The pupillographic findings indicate both sympathetic and parasympathetic deficit in the context of a widespread autonomic neuropathy. The aetiology of his illness remains unknown, but may have an auto-immune basis.

Contact: By Post


Stuart R. Steinhauer, PhD, Ruth Condray, PhD, Annette Kasparek, BS, and George G. Dougherty, Jr., MD.

Biometrics Research Program, Dept. of Psychiatry, University of Pittsburgh School of Medicine, Dept. of Veterans Affairs Medical Center, and Carnegie-Mellon University.

Pupillary dilation can be elicited by complex salient events, or by simple motor activity. We previously hypothesized that motor components contribute to an early inhibition of the parasympathetic system via the oculomotor nucleus, while later and more extensive dilations are associated with hypothalamically-mediated sympathetic activity. Increasing ambient light decreases pupil diameter by stimulating the parasympathetic pathway; consequently, the contribution of parasympathetic inhibition, as indicated by dilation, would be greater in light as compared darkness.

Pupil diameter was recorded in 29 normal volunteers to simple movements. Subjects made uncued, voluntary finger presses over 60 sec, with instructions to pace responses approximately 5 sec apart. Separately, subjects made the same response when cued by single tones (10 stimuli, ISI 5-7 sec). Each condition was recorded in darkness and under room illumination, with 4 sessions involving baseline and placebo conditions, or during pharmacological blockade of the sphincter or dilator musculature. Response onset was used to time-lock averages for uncued responses. Both stimulus and response onset were used separately to time-lock averages for cued responses.

For the baseline and placebo sessions, recording in light always produced an earlier-onset dilation than recording in darkness. Uncued responses produced significantly smaller and earlier peak dilations than cued responses. During blockade of the dilator by dipaprizole, the early components were present, but later dilations were attenuated. During blockade of the sphincter by tropicamide, the early components of the response were eliminated, but later dilation was observed. Principal components analysis, and sequential difference analysis, were used to provide converging evidence for a very early effect (300 msec) of parasympathetic inhibition.

The data indicate that the primary contribution of the motor response is associated with early inhibition of the parasympathetic pathway. In contrast, the increased dilation under the cued condition is reflected primarily in the sympathetic component. Greater amplitudes were elicited in the stimulus locked than response locked averages of the cued condition, suggesting that stimulus relevance was a more critical contribution to the dilation response than was the performance of the motor response alone.

ACKNOWLEDGEMENT: Supported by the US Public Health Service: National Institute of Mental Health Grant #PHS-MH55762, and by the Medical Research Service of the Dept. of Veterans Affairs.

CERTIFICATION: These procedures were reviewed and approved by the Institutional Review Boards of the University of Pittsburgh School of Medicine, and Pittsburgh VA Medical Center, consistent with government guidelines.



Stuart R. Steinhauer, PhD, Ruth Condray, PhD, and Annette Kasparek, BS.

Biometrics Research Program, Dept. of Psychiatry, University of Pittsburgh School of Medicine and Dept. of Veterans Affairs Medical Center

It has been clearly demonstrated that the pupillary reaction to light can be inhibited by a variety of psychosensory operations, such as intense (non-visual) stimuli, by anticipation of fearful events, or in states of anxiety. It was hypothesized that cognitive operations could also be demonstrated to show inhibition of the light reaction. Inhibitory inputs to the Edinger-Wesphal complex of the oculomotor nucleus include descending cortical and ascending reticular inputs. Ongoing cognitive activation was predicted to increase inhibition of this pathway, resulting in reduced pupillary constriction.

Pupillary constriction to light was elicited during a sustained processing task. 33 normal volunteers participated in this study. Light reactions were recorded during 4 sessions conducted on separate days: a baseline session, followed by randomized placebo, tropicamide, or dipaprizole instillation at separate sessions. All subjects had been screened to exclude narrow angle. At each session, after dark adaptation, light reactions to a 1 s light stimulus (4 s ISI) were recorded using infra-red pupillography under three conditions: with no separate task, as subjects performed a serial-7 subtraction task (subtract 7 from a given value, say out loud, and repeat), or as subjects performed a simple adding task (add+1) to control for verbalization.

Averaged response curves were computed following the first stimulus in each series. In the baseline and placebo conditions, performance of the serial-7 task produced a significant decrease in the amplitude of constriction, and earlier time to maximum constriction, as well as significant increase in overall diameter. The add+1 task produced no decrease in constriction amplitude, but overall diameter was increased to a level intermediate between the no-task and subtract-7 conditions. Latencies to onset of constriction and time of maximum constriction were not affected. The same findings were observed even for the first trial alone. Temporary blockade of the dilator muscle by dipaprizole produced no change in the effects seen under placebo, with the exception that overall diameter was similarly decreased for all task conditions. In contrast, blockade of the constrictor muscle by tropicamide produced nearly flat responses, and higher overall diameters, with only a slight increase in overall diameter for the subtract-7 condition compared to no task or add+1 conditions.

Control of these parameters emanates almost entirely from the parasympathetic pathway rather than reciprocal sympathetic activation. Thus, task performance produces both tonic inhibition (decreased parasympathetic outflow leading to increased initial diameter) as well as selective inhibition of active phasic responses. The findings are interpreted as representing primary influences of descending inputs from cortical regions during high cognitive loads. Measurement of light reflex inhibition parameters under different task conditions may provide a sensitive, quantifiable assessment of relative "mental effort" associated with varying cognitive activations.

ACKNOWLEDGEMENT: Supported by the US Public Health Service: National Institute of Mental Health Grant #PHS-MH55762, and by the Medical Research Service of the Dept. of Veterans Affairs.

CERTIFICATION: These procedures were reviewed and approved by the Institutional Review Boards of the University of Pittsburgh School of Medicine, and Pittsburgh VA Medical Center, consistent with government guidelines.



C.A. Sullivan, R. Lawrence and A.J. Britten*.

Neurodegeneration Research Group, St George's Hospital Medical School and * Department of Medical Physics, St George's Hospital.

The pupillary response to dilute (0.01%) Tropicamide solution has been proposed as a diagnostic test of Alzheimers disease (AD), with poor confirmation from other centres. Review of the reported work showed a wide variability in experience, possibly reflecting differences in either the drug stimulus, the patient group or the measurement procedure. Monocular pupillometry had been used by several of the groups, and since this may be sub-optimal when looking for small anisocoria we chose to use a binocular video pupillometer (Procyon UK), allowing variability reduction by multi-frame capture and analysis. A feature of several negative studies was the large dilation produced in controls with 0.01% Tropicamide, whereas the original positive study had minimal response in controls, and this lead us to investigate the Tropicamide dose response.

Aims To optimise the measurement conditions, with respect to Tropicamide dose and pupillometry technique, for the Tropicamide test in a group of patients with a diagnosis of a dementia of the Alzheimer's type, and controls.

Methods Pupil diameters were measured in 8 controls (age 28-68, mean 48 years) at 30 minutes following Tropicamide eye drops at concentrations of 0.01% and 0.001%, prepared within 1 week and with no preservative. Video binocular pupillometry was carried out in 6 patients with a provisional diagnosis of AD (aged 65-87 years) and 6 elderly controls (aged 57-81 years). A concentration of 0.001% Tropicamide was used, measured with the treated eye in darkness at 30 minutes post -Tropicamide.

Results Mean percentage pupil diameter increase in was 34 % (range 9.9% to 66 %) and 6.9 % (range -4.7% to 12.1 %) with Tropicamide concentrations of 0.01% and 0.001% respectively. Mean change in anisocoria was 3.4% in elderly controls and 5.0 % in AD in response to 0.001% Tropicamide.

Conclusion Tropicamide eye drops at a standard 0.01% concentration produced a large mydriasis in controls. With reduced Tropicamide concentration of 0.001% there was no difference in anisocoria change between controls and subjects with a provisional diagnosis of Alzheimer's disease. Further dose response studies may be warranted.



E. Szabadi, P. Bitsios, S. Tavernor and C.M. Bradshaw

Division of Psychiatry, University of Nottingham

The intracellular enzyme monoamine oxidase (MAO) plays a crucial role in terminating the action of the neurotransmitter noradrenaline in sympathetically innervated tissues [1]. Tyramine is an indirectly acting sympathomimetic amine which exerts its pharmacological effects by releasing noradrenaline [2]. It is known that MAO inhibitors can enhance pharmacological responses to tyramine [3]. MAO has two forms, A and B, and the two forms of enzymes show separate distributions both in the periphery and the brain [1]. There is little evidence regarding the form of enzyme which is involved in terminating sympathetically mediated responses, although MAO-A has been implicated.

In this experiment we examined the effects of two selective inhibitors of MAO (moclobemide: selective MAO-A inhibitor; selegiline: selective MAO-B inhibitor) on tyramine-evoked mydriasis in man. Twelve healthy male volunteers participated in three monthly sessions, each associated with ingestion of one capsule (moclobemide 450 mg, selegiline 10 mg or placebo) according to a double-blind balanced cross-over design. Tyramine hydrochloride eye drops (75 mM, 2 x 10 ? l) were instilled three times in the left conjunctival sac at 40-min intervals. Pupil diameter was monitored with a binocular infrared television pupillometer before, and for 4.5 h after, ingestion of the capsule. The pupillary response to tyramine was expressed as the area under the pupil diameter x time curve. A blood sample was taken before and 2 h after ingestion of the capsule for the assay of platelet MAO-B activity and plasma 3,4-dihydroxyphenylglycol (DHPG) concentration, an index of MAO-A activity. Data were analyzed by ANOVA, followed by Bonferroni's corrected t-test, using a significance criterion of P< 0.05.

Both moclobemide and selegiline caused significant miosis in the untreated eye. The mydriatic response to tyramine was potentiated by moclobemide. Selegiline reduced platelet MAO activity and mocolobemide reduced DHPG, consistent with the known profiles of these drugs on the two forms of MAO. The potentiation of tyramine-evoked mydriasis by moclobemide is likely to reflect the inhibition MAO activity in the iris, whereas the lack of effect of selegiline argues against the involvement of MAO-B. The miosis caused by the two MAO inhibitors is likely to be due to a central sympatholytic action of the drugs.

 Berry M D et al 1994. Progr Neurobiol 42, 375-391.

  1. Trendelenburg U 1972. In: Handbook of Experimental Pharmacology vol 33: Catecholamines, eds Blaschko H and Muscholl E. Springer, Berlin, pp. 336-362.
  2. Rand et al 1968. Br J Pharmacol 33, 287-303.


E. Szabadi, M.A. Phillips and C.M. Bradshaw

Division of Psychiatry, University of Nottingham, UK

Spontaneous fluctuations of pupil size in darkness are characteristic of a person's level of vigilance: these fluctuations are accentuated when subjects are drowsy [1]. Recently the "pupillographic sleepiness test" (PST) has been developed for the recording and quantification of "pupillary fatigue waves" [3]. This test has already been used to detect daytime sleepiness both in healthy sleep-deprived subjects [4] and in patients suffering from sleep disorders [5]. Central noradrenergic neurones are located in small nuclei in the lower brainstem, and they project to most parts of the neuraxis including the cerebral cortex. The central noradrenergic system has been implicated in arousal mechanisms [6]. In this experiment we examined the hypothesis that drugs which selectively activate or deactivate central noradrenergic neurones (e.g. yohimbine and clonidine, respectively), will modify the level of arousal in a predictable fashion, and that these changes will be detectable using PST.

Sixteen healthy males (18-25 years) participated in 4 weekly sessions, each associated with one oral drug condition (placebo, clonidine 0.2 mg, yohimbine 22 mg, clonidine 0.2 mg + yohimbine 22 mg), according to a balanced double-blind design. Pupil diameter was recorded continuously over 11 min in darkness using a dedicated monocular television pupillometer (AMTech Weinheim Germany). Average pupil diameter, power of pupil diameter fluctuations (obtained by Fast Fourier transformation), and the pupillary unrest index (PUI), a measure of cumulative changes in pupil size, were computed [3]. Autonomic functions known to be sensitive to centrally acting noradrenergic drugs (blood pressure, heart rate and salivary output) were recorded. Subjective "alertness", "anxiety" and "contentedness" were rated using visual analogue scales. Measurements were carried out 30 min after arrival in the laboratory and 2 h after drug ingestion. Data were analyzed by ANOVA followed by Dunnett's corrected t-test (criterion of significance P< 0.05).

Pupil diameter was increased by yohimbine and clonidine + yohimbine and tended to be reduced by clonidine. The power of pupillary fluctuations tended to be increased by clonidine and reduced by yohimbine, and the combination of the two drugs cancelled out each other's effects. PUI tended to be increased by clonidine and was significantly reduced by yohimbine and the combination of the two drugs cancelled out each other's effects. Clonidine decreased and yohimbine increased systolic and diastolic blood pressure and salivation, and the two drugs antagonised each other's effects. Clonidine reduced subjectively rated alertness. These results confirm the arousal increasing effect of the centrally acting noradrenergic activating drug yohimbine and the opposite effect of clonidine, and the suitability of PST to detect these effects.

  1. Lowenstein O et al 1963. Invest. Ophthalmol 2, 138-157
  2. Lüdtke H et al. 1998. Vision Res 38, 2889-2896

  1. Wilhelm B et al 1998a. Sleep 21, 258-265
  2. Wilhelm H et al 1998b. Graefe's Arch Clin Exp Ophthalmol 236, 725-729
  3. Robbins TW and Everitt BJ (1995) Arousal systems and attention. In: Gazzaniga MS (ed) The cognitive neurosciences. MIT Press, Cambridge Mass, pp 703-720


A. Tales, PhD*, S.R. Butler, PhD**, T. Troscianko, PhD* and G. Wilcock DM, FRCP***

* School of Experimental Psychology, University of Bristol, Bristol BS8 1TN ** Burden Neurological Institute, Stoke Park, Bristol ***Department of care of the Elderly, Frenchay Hospital, Bristol

The finding of neurofibrillary degeneration in neurons of the Edinger Westphal nucleus in Alzheimer's disease reported by Hunter ( Acta Neuropatholologica, 1985, 68: 53-58) prompted us to ask whether the reflex pupillary constriction to light (which includes the Edinger Westphal nucleus in its anatomical pathway) would display functional abnormality in Alzheimer's disease.

Three groups of individuals were tested: twelve healthy young people, twelve people with probable Alzheimer's disease and twelve healthy older adults. The pupillary light reflex was measured using a head-mounted infra-red pupillometer. As the Edinger Westphal nucleus is particularly involved with pupillary constriction we predicted that the amplitude of pupillary constriction would be reduced in Alzheimer's disease compared to normal ageing.

Our results confirmed this prediction. As the Edinger Westphal nucleus is cholinergic in nature, the measurement of pupillary constriction may provide a way of indicating change in central cholinergic status in Alzheimer's disease over time, or in response to drug treatments.

 This research was conducted in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to applicable national and local Human-Subjects requirements.



S.J. Tavernor, C.M. Bradshaw and E. Szabadi

Division of Psychiatry, University of Nottingham

Experimental anxiety in healthy subjects, induced by the threat of an electric shock, is associated with a reduction of the amplitude of the pupillary light reflex, which can be reversed by diazepam (see Bitsios et al, this Colloquium). In this experiment we examined the effects of pentagastrin, a peptide, which has been proposed as a model anxiogen in man [1] on the pupillary light reflex in normal subjects. For comparison we also examined the effects of the cold pressor test, a procedure which has been reported to activate the sympathetic nervous system without markedly increasing subjective anxiety [2].

Twelve healthy males (18-35 years) participated in two sessions in which they received (1) pentagastrin intravenously and a control infusion (saline), and (2) cold pressor test (immersion of one hand in water at 4o C for 90 seconds) and a control immersion (37o C for 90 seconds), in a balanced single-blind design. In each session the two treatments were administered in 12-min recording periods separated by 60 min. Resting pupil diameter and constrictor pupillary responses to 15 200-ms light pulses (565 nm, 0.43 mW cm-2) delivered to one eye at 30-s intervals were recorded in darkness by infrared television pupillometry; treatments were initiated after the 7th stimulus. Heart rate and blood pressure were recorded by finger plethysmography. After each recording subjects rated their anxiety retrospectively on visual analogue scales [3] and the Panic Symptom Scale [4]. Data were analyzed by ANOVA followed by paired t-test (criterion of significance P < 0.05). Both pentagastrin and the cold pressor test increased resting pupil diameter without any significant effect on the parameters of the light reflex response.

Both pentagastrin and the cold pressor test caused increases in heart rate, systolic and diastolic blood pressure, subjective ratings of anxiety and panic symptoms score. Pentagastrin and the cold pressor test had qualitatively similar effects on the pupil and cardiovascular measures, consistent with sympathetic activation. Although both pentagastrin and the cold pressor test increased subjectively rated anxiety this was not reflected in a reduction in the amplitude of the pupillary light reflex response. Therefore, we can conclude that the psychological anxiogen (ie. threat of an electric shock) and the chemical anxiogen (ie. pentagastrin) evoke anxiety via different mechanisms.

Lines C et al 1995. Br J Clin Pharmacol 93, 235-242.

Marriott I et al 1990. Clin Sci 79, 43-50.

Bond A J and Lader M H 1974. Br J Med Psychol 47, 211-218.

Van Megan H J G M et al 1994. Psychopharmacol 114, 449-455.



Franz Thoss, Holger Lüdtke, and Dirk Sandner

Carl Ludwig Institute of Physiology of the University Leipzig University Eye Hospital Tübingen, University Eye Hospital Dresden

In 1997 we demonstrated the influence of periodic workload on heart rate, pupil diameter and light reflex amplitude of the pupil in some cases. The aim of this study was to analyse the time relations between variations of those parameters. The present paper continues these investigations and presents the results of 15 subjects.

By a weak workload the activity of the sympathetic system was raised periodically. In intervals of two minutes our subjects had to pull the handle of a hand ergometer twenty times during 20 seconds. At this work the force was constant (10 N) and the distance was adjusted between 20 and 40 mm depending on the constitution of the subject. The total duration of each experiment was 660 s. In intervals of 10s a weak, threshold-near 0.1s light stimulus was given to the left eye. The influence of the pupil diameter of this eye on the reflex amplitude was eliminated pharmacologically. The pupil diameter of the right eye and the ECG were recorded continuously. From the records the pupil diameter, the light reflex amplitude, the heart rate and their power spectra and cross correlation functions were calculated.

The results of this calculation show that the three investigated parameters are to be 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 activity by the work load. The results concerning the analysis of the time co-ordination between the parameters did not meet our expectations. To our astonishment the time shift between work load and the investigated functions shows strong inter-individual differences.

 The investigation was conducted in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to national and local Human-Subjects requirements.



Max Warga, Barbara Wilhelm, Holger Lüdtke and Helmut Wilhelm

University Eye Hospital Tübingen, Department of Pathophysiology of Vision and Neuro-ophthalmology

Purpose Darkness is one of the major conditions for the pupillographic sleepiness test (PST by AMTech, Weinheim, Germany) which measures sleepiness-related oscillations in pupil diameter. The aim of the present study was to investigate the interference of light-induced oscillations into sleepiness waves. In a further step we wanted to isolate merely light-induced oscillations under constantly high alertness level.

Methods The PST based on infrared pupillography has been described in detail elsewhere 1. Additionally another pupillographic system developed in our department and video-tapes were used for the recordings in the present study. Healthy young subjects (n=15) were measured every two hours during a night of sleep deprivation in a constant light condition of 2 cd/m2, 40 cd/m2 and 400 cd/m2 . Time interval was 10 days minimum to avoid accumulated sleep-debt. In a second experiment subjects (n=13) were tested at high alertness level and the time of testing was reduced to 5.5 minutes. Light intensities applied with an interval of 30 minutes were 0 cd/m2, 0.5 cd/m2, 2 cd/m2, 40 cd/m2 .

Results Amplitude spectrum 0.8 Hz increased significantly during sleep-deprivation at 40 cd/m2 (ANOVA p<0.001) and 400 cd/m2 (ANOVA p<0.001), while at 2 cd/m2 the increase was non-significant. Compared to measurement in darkness the changes at 40 cd/m2 and 400 cd/m2 were much less pronounced. At constant alertness level both the amplitude spectrum 0.8 Hz and the mean frequency showed increment with increasing light intensity.

Conclusion Effects of sleep-deprivation are easier to detect when recording pupillary oscillations in darkness or brighter light than in dim light conditions. Light-induced oscillations under constantly high alertness level occur inconstantly during the recording time within the same subject and show large inter-subject differences. Light-induced oscillations differ from sleepiness waves predominantly by their regularity. The relationships between light intensity, amplitude spectrum and frequency of spontaneous pupillary oscillations raise some questions demanding further experimental work and alternative approaches in analysis.

1Lüdtke H, Wilhelm B, Adler M, Schaeffel F, Wilhelm H. Mathematical procedures in data recording and processing of pupillary fatigue waves. Vision Research 1998; 38:2889 - 2896.

Certification: This study was conducted in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to requirements of the Ethics Committee of the University of Tübingen.

Acknowledgement: B. Wilhelm and H. Lüdtke worked on the present study with financial support of the Steinbeis Center for Biomedical Optics and Functional Testing Tübingen (Prof. Dr. E. Zrenner).



Barbara Wilhelm, Andreas Koerner, Holger Lüdtke, Helmut Wilhelm and Klaus Ederle*

University Eye Hospital Tübingen, Department of Pathophysiology of Vision and Neuro-ophthalmology *Klinik Loewenstein, Department of Pneumology, D-74245 Loewenstein

Purpose The pupillographic sleepiness test (PST) detects sleepiness in normals and patients with any kind of dyssomnia as well1,2. The parameters quantifying pupillary unrest differ significantly between groups of patients with sleep apnea syndrome and healthy subjects. The present study should test whether or not the PST could be used as a screening method for the detection of SAS. The background is characterized by 1. the fact that many patients are not diagnosed before any complicating disease or accident caused by excessive daytime sleepiness occurs and 2. a present lack of practicable and time-saving screening methods.


Subjects. 140 normal male subjects were compared to 54 patients with polysomnographically diagnosed SAS, all subject were 30 to 70 years old.

PST. All measurements were performed between 8 am and noon without any preceding consumption of coffein, nicotine or any vigilance-influencing drugs. The PUI3 served as the parameter of evaluation, all analyses were carried out on the natural logarithm of the PUI because of its normal distribution.

Results The PUI results of the normal subjects and the SAS-patients were significantly different. At a PUI of 5.7 mm/min the sensitivity and specificity of the PST were 73%. The positive predictive value for a PUI of 10.0 mm/min was 40%. The test reliability was studied in PST-repetition after four months in 38 normal subjects. The correlation (Pearson rank coefficient r=0.64) between both PST-measurements was highly significant (p<0.0001).

Conclusions Though the PST identifies acute and chronical sleepiness it further also showed practical qualities for the use as a screening method for the detection of obstructive sleep apnea syndrom. The specificity, sensitivity and reliability data combined with the objective, noninvasive and time-saving properties of the PST offer the possiblity of a reasonable screening method for this serious sleep disorder.

1 Wilhelm B, Wilhelm H, Lüdtke H, Streicher P, Adler M. Pupillographic assessment of sleepiness in sleep-deprived healthy subjects. Sleep 1998; 21:258-265.

2 Wilhelm H, Lüdtke H, Wilhelm B. Pupillographic sleepiness test applied in hypersomniacs and normals. Graefe's Arch Clin Exp Ophthalmol 1998; 236:725 - 729.

3 Lüdtke H, Wilhelm B, Adler M, Schaeffel F, Wilhelm H. Mathematical procedures in data recording and processing of pupillary fatigue waves. Vision Research 1998; 38:2889 - 2896.

 Certification: This study was conducted in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to requirements of the Ethics Committee of the University of Tübingen

Acknowledgements: The studies of this paper were granted by fortuene F. 1222074.1 and (DFG) Deutsche Forschungsgemeinschaft WI 1066/3-1.



Helmut Wilhelm, Tobias Peters, Holger Lüdtke, Cyriaque Barry and Barbara Wilhelm

University of Tübingen, Eye Hospital, Dept. of Pathophysiology of Vision and Neuro-ophthalmology; D 72076 Tübingen

Purpose Objective testing of visual function is a strong demand not only in ophthalmology. Because the pupillary light reflex is usually equal in both eyes, interocular differences in the light reflex can be used to detect even subtle damage of the prechiasmal visual pathways. The swinging flashlight test is clinically widely used for this purpose. An interocular difference in the direct light reflex or a difference between the direct and consensual reaction of the same eye indicates relative afferent pupillary defect (RAPD) if no efferent defect or marked contraction anisocoria is present. A RAPD is only found in neuroretinal damage, never in disturbacies of the optic media. Therefore a RAPD is an important sign in differential diagnosis of an unexplained visual loss. It may be measured by balancing the light reflexes with neutral density filters. Although a RAPD is a very strong indicator of neuro-retinal pathology, it is occasionally seen in subjects without any obvious disease. Therefore we studied the occurrence of a RAPD in normal subjects.

Methods 101 subjects were first examined clinically and then by means of pupillography. The swinging flashlight test was carried out using neutral density filters in 0.15 log-unit steps for measurement. Pupillographically, light emitting diodes were placed in front of each eye, alternately flashing for 2 s and a binocular realtime pupillometer recorded the direct and consensual pupillary response. After artifact detection and removal, the amplitudes of pupillary responses were determined. The mean of the direct and the consensual response was used for evaluation.

Results Pupillographically, in 55 subjects (55%) the RAPD was 0.05 log-units or less, in 64 (64%) 0.1 log unit or less. Only 15 persons (15%) exceeded 0.15 log units and only 2 exceeded 0.3 log-units (maximum 0.39). Clinically no RAPD was found in 83 of 102 subjects (81%), in 11 subjects a RAPD of 0.15 log units and in 8 subjects one of 0.3 log units was seen. No RAPD larger than 0.45 was found clinically. All subjects with a RAPD of 0.3 log units and more underwent perimetry, but no field abnormality was detected.

Discussion A RAPD may be present even in normal subjects. However, this RAPD is very small, both clinically and pupillographically. A RAPD exceeding 0.39 log-units has not been found, even when pupillographical measures were used. The cause of this RAPD may be an imbalance of the afferent pupillary system including the connections between visual pathways and pretectal nuclei in the midbrain. In conclusion: A small RAPD may be found in normals, a RAPD exceeding 0.3 log-units is very unlikely in normal subjects.

 Acknowledgement: The development of the automated swinging flashlight test was granted by TOMEY.

Certification: This study was conducted in accordance with the guidelines provided by the Declaration of Helsinki and in adherence to requirements of the Ethics Committee of the University of Tübingen.



Harry J. Wyatt, PhD

Schnurmacher Institute for Vision Research, State University of New York, State College of Optometry, New York, NY. USA.

Purpose The iris is highly mobile, even though relatively inextensible collagen fibers form the strongest part of its structure. Rohen (1951) observed a pattern of iris collagen fibers forming arcs from iris root to pupil, thereby avoiding the problems of circumferential fibers (which would hinder pupil dilation) or radial fibers (which would hinder pupil contraction). He described two sets of arcs with opposite handedness, forming a meshwork which was also interwoven with blood vessels. However, constructing such a mesh to permit pupil mobility is not simple because (a) fiber extensibility is limited, and (b) sliding of meshwork components - during pupil size changes - would lead to wear and tear.

Methods A computational model was used to determine the optimal properties of such a mesh. The form of the fiber arcs was optimized across a set of pupil sizes by: (i) minimizing strain (i.e. stretch) of the fibers, and (ii) constraining each point on a fiber to move only in the radial direction (minimizing sliding). A logarithmic spiral was used as an initial solution, and a twenty term polynomial was added to it. A single optimization run was performed using one value of fiber angular extent (the angle of the iris "pie-slice" over which a fiber arc extends). Initially, the assumption was made that overall iris strain is linear; i.e., a point halfway between root and pupil margin remains halfway between them for all pupil sizes. Later, this constraint was relaxed.

Results Solutions converged to a repeatable spiral form. As the angular extent increased up to about 100 deg, the optimization improved, but further increases did not lead to further improvement. Removing the constraint of linear overall iris strain led to modest further improvements in the optimization.

Comparison with data and conclusions One way of characterizing the meshwork is the "lattice angle" (the angle between fibers with left- or right-handed arcs), which is a function of both radial position and pupil size. Some lattice angle data is available from Rohen's work, and this was compared with results of the optimization. The comparison suggests that the collagen iris "skeleton" approximates a minimum-wear-and-tear meshwork.




Pupil Email list


 If you do not currently belong to the Pupil mailing list, you can join by sending the following message to

 join pupil firstname lastname

Note: Type your own personal names instead of firstname and lastname ! line 493. /P>

The Mailbase computer will then send you a message asking you to confirm your subscription by posting back a code. This is a security measure. Once you have done this you can send and receive messages from the list.