Autonomic Components of the Human Pupillary Light Reflex

Philip H. Heller,* Franklin Perry,f Don L. Jewerr.f and Jon D. Levinef

To investigate the autonomic components of the pupillary light reflex in humans, we used infrared pupillometry combined with a partial local cholinergic (tropicamide) or alpha-adrenergic (thymoxamine) blockade. The pupillary response curve was analyzed using parameters identical or similar to those employed previously to study the autonomic components of the pupillary light reflex. Tropic- amide increased baseline pupil area and affected five of the eight measured parameters. Thymoxamine lowered baseline pupil area but did not affect any of the parameters. We found the expected cholinergic contribution to the constrictive phase of the pupillary light reflex but no evidence for peripheral alpha-adrenergic activity during redilation. We propose that redilation primarily involves parasympathetic relaxation, modulated by cholinergic inhibition of the dilator muscle and central sympathetic inhibition of the Edinger-Westphal nucleus. Invest Ophthalmol Vis Sci 31:156-162, 1990

While it has been long known that constriction of muscle and an alpha-adrenergic stimulation of the the pupil to a light flash depends on parasympathetic dilator muscle, as required by the Lowenstein and outflow from the Edinger-Westphal (E-W) nu- Loewenfeld model.l>8f9 In vitro study has also demon- cleus,1'2 the autonomic control of the individual strated beta-adrenergic inhibition of both muscles, phases of the human pupillary light reflex is not well alpha-adrenergic inhibition of the sphincter and cho- established. Lowenstein and Loewenfeld employed linergic inhibition of the dilator.10"12 The physiologi- pharmacological agents and surgical ablation to study cal contribution to the light reflex of these latter ef- the pupillary light reflex in cats and monkeys and fects, observed in vitro, is unknown. Other workers described two phases of constriction and primary and have found that the beta-adrenergic blocker timolol secondary redilatory phases.3 They hypothesized that does not affect either baseline pupil area or the pupil- the constrictive phases were due to a parasympathetic lary light reflex.1314 activation modified at its conclusion by a superim- In the present study we measured the pupillary posed central sympathetic inhibition of the E-W nu- light reflex by infrared pupillometry after applying cleus. The primary redilation was thought to be due agents to produce a parasympathetic or sympathetic to parasympathetic relaxation, while the secondary blockade. Our data confirm the expected cholinergic redilation was thought to be due to an increase in contribution to constriction but fail to confirm the peripheral sympathetic activity. Observations in a expected alpha-adrenergic contribution to the redila- small number of patients with neurological lesions4"7 tory phase. Instead we find that the recently described suggest that this model also may apply in humans. cholinergic inhibition of the dilator muscle may con- Information relevant to possible mechanisms un- tribute to the rate of redilation. We present a hypoth- derlying the pupillary light reflex in humans also is esis to explain redilation in the human pupillary light available from pharmacological analysis of human reflex that includes this cholinergic inhibition. intraocular muscle. Several studies have demonstrated a cholinergic stimulation of the constrictor Materials and Methods The pupillary light reflex was studied in 27 normal subjects (16 males and 11 females) between the ages From the *Kaiscr-Permanente Medical Center, Hayward, Cali- fornia, and the tSchools of Medicine and Dentistry, University of of 20 and 40 years. The experimental protocol for this California, San Francisco, California. study was approved by the UCSF Committee on Supported in part by grants AM32634, DE05369 and NS21647 Human Research and informed consent was ob- from the National Institutes of Health and from the Community tained in writing. Service Program of Kaiser Foundation Hospitals. Pupil area was measured every 50 msec by infrared Submitted for publication: November 13, 1987; accepted June 7, 1989. video pupillometry (Micromeasurements, Inc., Reprint requests: Jon D. Lcvine, MD, PhD, Division of Rheu- Berkeley, CA). The light stimulus was a square-wave matology, U-426 (Box 0724), UCSF, San Francisco, CA 94143. pulse of collimated white light, generated by a glow-

Downloaded from iovs.arvojournals.org on 10/02/2021 No. 1 AUTONOMIC COMPONENTS OF THE PUPILLARY LIGHT REFLEX /.Heller er ol 157

Table 1. Pupillary light reflex parameters in normals, drug and comparative normal subgroups"

Group (n) C(%)\ mrci mrdrf d2(%) imrc tmrd, rec (%)

NLS (27) mean 43.9 43.4 13.2 7.9 0.296 0.515 1.29 83.4 SE 1.2 2.4 0.7 0.6 0.007 0.007 0.02 1.5 Tropic (4) mean 9-7* 19.8+ 8.1§ 7.2 0.339* 0.543 1.195 95.0§ SE 2.3 3.7 2.2 1.0 0.010 0.017 0.075 2.8 lgNLs(5) mean 38.5 52.4 15.7 6.1 0.255 0.511 1.283 81.8 SE 2.7 5.1 1.5 0.7 0.016 0.007 0.054 3.7 Thymox (5) mean 50.9 34.7 11.2 6.3 0.298 0.493 1.228 86.7 SE 1.5 5.7 0.8 1.3 0.009 0.014 0.036 2.9 smNLs(ll) mean 46.3 32.7 10.2 8.8 0.306 0.500 1.269 82.8 SE 1.7 2.3 0.8 1.2 0.010 0.010 0.035 3.0

* Tropicamide compared to large normals (IgNLs), thymoxaminc com- t These parameters arc significantly correlated with baseline pupil area. pared to small normals (smNLs). For parameters independent of pupil area, X P < 0.01. comparison with the entire control group did not alter findings of statistical § P < 0.05. significance.

modulator tube (Sylvania R-l 131C; Mountainview, nate accommodation effects. The subject pressed a CA), focused in the plane of the pupil and contained button to begin each recording period, which contin- within the pupil (ie, a Maxwellian-view stimulus).1516 ued until 5 sec after the stimulus was delivered. The duration of the stimulus (200 msec) was chosen Length of the prestimulus interval was randomly set to be shorter than the latency to the light reflex4 (Table 1) to further ensure elimination of variation in stimulus intensity. Microcomputer-controlled track- ing was used to compensate for minor head and eye movements. Pupil area data were stored in a PDP-8 computer. Pupil area was found to be stable after 5 min of adaptation to a constant dim level of illumination (0.34 ft-candles, measured by a Tektronix J-16 pho- tometer probe located on the wall one meter in front of the subject). Starting shortly before a trial, the subject fixated on a red light-emitting diode, coaxial with the stimulus and placed at optical infinity to elimi-

50 r

-60 0 12 3 4 5 TIME (SEC)

-10 1 2 3 4 5 Fig. 2. A single response to light (top) and its time differential (below): BPA, mean baseline pupil area over 1 sec prior to flash; tc, TIME (SEC) time to constriction (latency); C (%), size of constriction (as % of BPA); d2 (%), % secondary redilation (% redilation occurring be- Fig. 1. A series of pupillary light reflex responses from a typical tween 2.2-2.6 sec after the stimulus (S); rcc (%), % redilation at 5 subject. Responses to six 200 msec 1200 cd/m2 stimuli (S) pre- sec after the stimulus; mrc, tmrc, maximum rate of constriction sented at 1 min intervals at time t = 0. Pupil area (PA) measured at and its time after S and; mrdi, tmrdi, maximum rate of primary 20 Hz. redilation and its time after S.

Downloaded from iovs.arvojournals.org on 10/02/2021 158 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / Jonuory 1990 Vol. 31

100 90 Q 9.7 +/- 2.3 **

80 Xlg 38.5 +/- 2.7 70 ^ 60

^J 50 "" A V v v )O X X X XX jX X X 40 x X ; *, X X 30 X A 50.9 +/- 1.5 Fig. 3. (A, B) Five analy- 20 Xsin 46.3 +/- 1.7 sis parameters as a function D 10 a of baseline pupil area (BPA) in normal (X) and after par- - I 1 0 1 . 1 tial parasympathetic (tropicamidc, D) or sympa- 10 20 30 40 50 60 70 thetic (thymoxaminc, A) block. Dotted lines indicate BPA classification of normals into small, medium and 100 large BPA groups. Small BPA normals (Xsm) were 90 compared to subjects A 34.7 +/- 5.7 a 19.8 +/- 3.7 **. treated with thymoxaminc 80 Xsm 32.7 +/- 2.3 Xlg 52.4 +/- 5.1 Large BPA normals (Xlg) were compared to subjects 70 treated with tropicamidc (see Table 1). H 60 I X £ 50 Al X * 40 X Xy IX 30 X* a 20 a 10 0 10 20 30 40 50 60 70 B BPA

at either 3 or 4 sec in order to reduce effects of flash examined by tangential light beam to detect the pres- anticipation on the pupillary response. Up to 12 re- ence of a narrow anterior chamber, before adminis- sponses were obtained at 1 min intervals. Baseline tration of tropicamide. The only adverse reaction pupil area was determined from the prestimulus re- noted following either drug was the routinely ob- cording. served transient burning sensation with thymoxa- Nine subjects also were tested after conjunctival mine.1819 The pupillary light reflex was studied dur- administration, in the consensual eye, of one to two ing a partial blockade (ie, 10 to 15 min after drug drops of either 0.5% tropicamide (Alcon Laborato- administration) rather than at the time of maximal ries, Forth Worth, TX) or 0.5% thymoxamine effect, after 30 min. This allowed us to obtain re- (Warner-Lambert, Morris Plains, NJ). These agents sponses during a blockade but at levels of baseline produce a parasympathetic or sympathetic (alpha- pupil area comparable to some normal subjects. This adrenergic) block, respectively.1718 All subjects were simplified the data analysis since several parameters

Downloaded from iovs.arvojournals.org on 10/02/2021 No. 1 AUTONOMIC COMPONENTS OF THE PUPILLARY LIGHT REFLEX / Heller er ol 159

A 11.2 +/- 0.8 n 8.1 +/- 2.2 * 30 - Xsm 10.2 +/- 0.8 Xlg 15.7 +/- 1.5

£ 20 X x x*

yx x A X X x *x x # X ° A ^ X A 10 - a *A* xX - X X a o 0 . I.I.1 . 1 , . 1 . 1 . 10 20 30 40 50 60 70

Fig. 3 (continued) (C, D). BPA See legend under Figure 3 (A, B). 30

A 6.3 +/- 1.3 Q 7.2 +/- 1.0

Xsm 8.8 +/- 1.2 Xlg 6.1 +/- 0.7 20

X |

10 X X| a X X X AX xj x, a X X X a 1 10 20 30 40 50 60 70 D BPA

used in the analysis are dependent on pupil area in a line pupil area (BPA), latency to constriction (tc), size 20 nonlinear fashion, confounding an interpretation of of constriction [C(%)]} maximum rate of constriction control and drug data from the same individual. In and its latency (mrc, tmrc), maximum rate of pri- addition, since at the limits of pupil size mechanical mary redilation and its latency (mrdi, tmrd^, and factors can influence pupillary response,21 it was im- percentage recovery [rec(%)].3 Under our experimen- portant to compare groups with comparable baseline tal conditions, employing a short, Maxwellian-view pupil area. stimulus, unlike the stimulus used in the study of For each pupillary response a smoothed curve22 of Lowenstein and Loewenfeld,3 the secondary redila- pupil area versus time and its time differential curve tion was often partially overlapped by the primary were plotted. These curves were analyzed using pa- redilation. This has been observed by others employ- rameters either identical to or similar to those found ing a short stimulus.23 Therefore, we used the percent by Lowenstein and Loewenfeld to reflect the auto- redilation occurring during the time interval 2.2-2.6 nomic components of the pupillary light reflex: base- sec after the stimulus (ie, 2.0 to 2.4 sec after the end of

Downloaded from iovs.arvojournals.org on 10/02/2021 160 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Jonuory 1990 Vol. 01

100 Q A X a 90 X A XX ixx x x x 80 X X xi ^ 70 S 60 o Fig. 3 (continued) (E). See Q> 50 legend under Figure 3 (A, k_ A 86.7 +/- 2.9 Q 95.0 +/- 2.8 B). Xlg 81.8 +/- 3.7 40 Xsm 82.8 +/- 3.0 30 20 10 0 10 20 30 40 50 60 70 BPA

the stimulus) to measure secondary redilation drug-treated subjects, that is, tropicamide-treated [d2(%)]. This time interval was selected because: (1) subjects were compared with controls with large base- in some curves there was a distinct secondary redila- line area and thymoxamine-treated subjects were tion noted at this time interval; (2) evidence for a compared with controls with small baseline area (Fig. secondary redilation at this time interval was consis- 3). Using this comparison, tropicamide significantly tently seen in the derivative curve; and (3) this time affected five of the eight parameters, decreasing the interval corresponds to the time during redilation of size of constriction [C(%)]; the maximum rate of the distinct secondary phase observed by Lowenstein constriction (mrc) and the maximum rate of primary and Loewenfeld.3 redilation (mrd^, and increasing the latency to con- Statistical analysis was done using the student t- striction (tc) and the percentage recovery [rec(%)]. test. The measure of secondary redilation [d2(%)]5 the time to maximum rate of constriction (tmrc), and the Results time to maximum rate of primary redilation (tmrdi) were unaffected by tropicamide. Mean baseline pupil area for the normal subjects Thymoxamine treatment, despite significantly was 31.9 ± 2.0 mm2 (mean ± SE). For tropicamide- lowering baseline pupil area, failed to produce signifi- treated subjects the mean baseline area was 53.9 ± 2.6 cant differences in any of the parameters. mm2, and for thymoxamine-treated subjects the 2 mean was 19.2 ± 2.6 mm . Thus, both drugs pro- Discussion duced the expected autonomic block (P < 0.05). Several sequential measurements of the pupillary In this study we used a partial, local cholinergic or light reflex in a single subject are shown in Figure 1. alpha-adrenergic blockade to investigate the auto- The latency, maximum rates of change, time to peak, nomic components of the pupillary light reflex in and times of maximum rate of change showed little humans. Baseline pupil area, measured under meso- variation from trial to trial. There was minor nonsys- pic light conditions, was similar to that observed by tematic fluctuation in baseline pupil area. Figure 2 others.24 The cholinergic antagonist tropicamide sig- demonstrates a single response with its time differen- nificantly increased baseline pupil area. This effect tial and illustrates the parameters used to analyze the was probably due to antagonism of the known cho- response. linergic stimulation of the sphincter muscle, but an The values for the parameters for normal and effect on the recently described cholinergic inhibition drug-treated subjects are listed in Table 1 and plotted of the dilator muscle may have contributed. The as a function of baseline pupil area in Figure 3 (A-E). alpha-adrenergic antagonist thymoxamine signifi- For statistical analyses drug-treated groups of subjects cantly decreased baseline pupil area. This effect was were compared only to the subset of normal subjects probably due to antagonism of the known alpha- who had a baseline area comparable to that of the adrenergic stimulation of the dilator, but antagonism

Downloaded from iovs.arvojournals.org on 10/02/2021 No. 1 AUTONOMIC COMPONENTS OF THE PUPILLARY LIGHT REFLEX / Heller er ol 161

of alpha-adrenergic inhibition of the sphincter, which pupil area. If secondary redilation were due to periph- has been reported,10"12 may have contributed. eral sympathetic activity, we would predict a decrease Since the size of the pupillary constriction to light in the secondary redilation measure, d2(%), in the is known to be a function of pupil area,21 we ana- presence of thymoxamine. Since we did not see any lyzed, as well, the dependence on pupil area, of all the change in d2(%), we also looked at redilation in thy- individual parameters in our analysis. Absolute mag- moxamine-treated subjects during the time period of nitude of constriction, as mentioned above, is highly 2.6-5.0 sec post-stimulus to see if there was a later dependent on pupil area (r = -0.8), while the per- effect of thymoxamine, but none was seen. The percentage constriction [C(%)] was only moderately cor- centage recovery, rec(%), was also unaffected in thy- related with area (r = -0.4). The maximum rate of moxamine-treated subjects, further suggesting a min- constriction, mrc, and the maximum rate of primary imal or absent contribution by peripheral alpha- redilation, mrdi, both absolute measures, were adrenergic activity (either dilator stimulation or greater at larger pupil area (r = 0.8 and 0.7, respec- sphincter inhibition) to the redilatory phase. The tively), presumably secondary to the increased con- thymoxamine-treated pupils were compared to nor- striction which occurred in a reflex of essentially the mal pupils of similar size since this was the only ap- same duration. The two percentage measures, sec- propriate control available. It is possible that normal ondary redilation [d2(%)] and percentage recovery small pupils, which have increased central parasym- [rec(%)], were independent of pupil area. Of the tem- pathetic tone, have a generally inhibited sympathetic poral parameters, only latency (tc) was a function of system and thereby an inhibited dilator activity. A pupil area, inversely so, as has been previously ob- small pupil also might have diminished dilator activ- served.2325 The values we observed for the parameters ity because of the length-tension curve properties for [except d2(%), a parameter we devised] were similar the stretched dilator muscle. While it is not possible to those previously observed.32627 to rule out these effects, if they were operative in the Because of the dependence of some of the parame- normal small pupil controls, we should have ob- ters on pupil area, we compared the data from drug- served an increased value for d2(%) (our measure to treated subjects with that from control subjects who detect secondary active sympathetic redilation) in had a baseline pupil area comparable to that of the normal pupils with a larger BPA, but this was not drug-treated subjects. seen (ie, 62% was independent of BPA). Tropicamide affected five of the eight parameters: While it is true that adrenergic antagonists such as C (%), mrc, mrdi, tc and rec (%). The decreases in C thymoxamine are not pure in terms of receptor ac- (%) and mrc and the increase in tc are readily ex- tion (alpha vs. beta or between alpha subclasses) or plained by the well established antagonism by tropic- might have other actions, the biochemical evidence amide of the cholinergic parasympathetic activation suggests strongly that thymoxamine as employed by of the sphincter pupillae muscle. The maximum rate us would have been expected to affect sympathetic of primary redilation, mrdi, which is thought to rep- activity of the dilator. It is possible that the difference resent parasympathetic relaxation, was also de- between our finding and that of Lowenstein and creased. This decrease in mrdi probably was due to Loewenfeld is due to the different stimulus used. In the significantly attenuated constriction after tropic- their study3 they employed a prolonged (1 sec) non- amide in two of the four subjects. Unexpectedly, the Maxwellian-view stimulus. Another possibility is that percentage recovery, rec (%), was increased in the in the ablation techniques they employed, on which presence of tropicamide. This observation would not their interpretation was based, other components be- be expected if redilation was due primarily to para- sides alpha-adrenergic efferents were affected (eg, sympathetic relaxation and peripheral sympathetic cholinergic inhibitory outflows28). Our data do bring activity. This finding can be explained, however, by into question the role of peripheral sympathetic activ- assuming that cholinergic inhibition of the dilator ity in the mechanism for the finding of a secondary pupillae10"12 plays a role in the light reflex. If phasic redilation of the pupil following a light stimulus. It cholinergic inhibition of the dilator normally is pres- should be emphasized that there is ample evidence ent during redilation then a cholinergic blocker such that active central and peripheral sympathetic activ- as tropicamide would be expected to decrease this ity is necessary for the phenomenon of reflex pupil- inhibition, thereby enhancing redilation. Since rec(%) lary dilatation in response to sensory or psychosen- was measured at 5 sec after the stimulus, we conclude, sory stimuli.29 that this cholinergic inhibition of the dilator still It is possible that redilation represents predomi- would be present or have its effect still present at 5 sec nantly a continual return to baseline tone (ie, para- post-stimulus. sympathetic relaxation) modulated by changes in Thymoxamine did not significantly affect any of cholinergic inhibition of the dilator as described the parameters, although it did decrease baseline above. This model, however, does not easily explain

Downloaded from iovs.arvojournals.org on 10/02/2021 162 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / January 1990 Vol. 31

the often observed abrupt attenuation in the rate of the autonomic inncrvation of the human iris. Br J Pharmacol redilation observed in mid-redilation (Fig. 1). This 38:462P, 1970. attenuation, however, could be the result of a sudden 9. Lowenstein O and Loewenfeld IE: Effect of physostigmine and pilocarpine on iris sphincter of normal man. Arch Ophthalmol decrease in the level of the increased central sympa- 50:311, 1953. thetic inhibition of the E-W nucleus, which is 10. Kern R: The adrencrgic receptors of the intraocular muscles in thought to commence with and be responsible for the man: An in vitro study. Graefcs Arch Klin Exp Ophthalmol secondary phase of constriction.3 A decrease in this 180:231, 1970. ongoing central sympathetic inhibition during the pe- 11. van Alphen GWH: The adrencrgic receptors of the intraocular riod of increased parasympathetic tone compared to muscles of the human eye. Invest Ophthalmol 15:502, 1976. 12. Yoshitomi T, Ito Y, and Inomata H: Adrcnergic excitatory and baseline would result in a decreased rate of redilation. cholinergic inhibitory innervalions in the human iris dilator. Indeed, Lowenstein and Loewenfeld proposed the Exp Eye Res 40:453, 1985. mechanism of rapidly alternating levels of central 13. Johnson SH, Brubaker RF, and Trautman JC: Absence of an sympathetic inhibition to explain the cyclical changes effect of timolol on the pupil. Invest Ophthalmol Vis Sci in pupil area that they observed during pupillary light 17:924, 1978. 14. Namba K, Utsumi T, and Nakajima M: Effect of timolol on reflexes in excited monkeys. the pupillary dynamics under open-loop photic stimulus. Folia In summary, we have demonstrated the feasibility Ophthalmol Jpn 31:118, 1980. of employing pupillometry combined with pupillary 15. Stark L: Neurological Control Systems—Studies in Biocngi- pharmacological blockade to dissect the individual neering. New York, Plenum Press, 1968. 16. Wcstheimcr G: The Maxwelliam view. Vision Res 6:669, parasympathetic and sympathetic contributions to 1966. the pupillary light reflex in humans. We found the 17. Pollack SL, Hunt JS, and Poise KA: Dose-response effects of expected cholinergic contribution to the constrictive tropicamide HCL. Am J Optom Physiol Optics 58:361, 1981. phase but no evidence for peripheral sympathetic ac- 18. Wand M and Grant WM: Thymoxaminc hydrochloride: An tivity during secondary redilation. We propose that alpha-adrenergic blocker. Surv Ophthalmol 25:75, 1980. 19. Pfeiffer MA, Cook D, Brodsky J, Tice D, Parrish E, Rcenan A, redilation may be a complex event involving para- Halter JB, and Porte D Jr: Quantitative evaluation of sympa- sympathetic relaxation, cholinergic inhibition of the thetic and parasympathetic control of iris function. Diabetes dilator muscle and central sympathetic inhibition of Care 5:518, 1982. the Edinger-Westphal nucleus. 20. Sun F and Stark L: Pupillary excape intensified by large pupillary size. Vision Res 23:611, 1983. Key words: pupil, human, autonomic, light reflex, cholin- 21. Loewenfeld IE and Newsome DA: Iris mechanics: I. Influence ergic, adrenergic of pupil size on dynamics of pupillary movements. Am J Ophthalmol 71:347, 1971. References 22. Vellcman PF and Hoaglin DC: Applications, Basics, and Computing of Exploratory Data Analysis. Boston, Duxbury 1. Davson H: Physiology of the Eye, 4th ed. New York, Aca- Press, 1981. demic Press. 1980. 23. Meyer ME, Ogle KN, Hollenhorst RW, and Mover NJ: Deriv- 2. Lowenstein O and Loewenfeld IE: The pupil. In The Eye, ative curve in evaluation of pupillary reflex response to light. Davson H, editor. New York, Academic Press, 1969, pp. Exp Eye Res 8:355, 1969. 255-337. 24. Ellis CJK: The pupillary light reflex in normal subjects. Br J 3. Lowenstein O and Loewenfeld IE: Mutual role of sympathetic Ophthalmol 65:754, 1981. and parasympathetic in shaping of the pupillary reflex to light. 25. Alpern M, McCready DW Jr, and Barr L: The dependence of Arch Neurol Psychiatr 64:341. 1950. the photopupil response on flashduratio n and intensity. J Gen 4. Lowenstein O: Clinical diagnosis of disturbances of the central Physiol 47:265, 1963. sympathetic system by means of pupillography. Arch Neurol 26. Lowenstein O and Loewenfeld IE: Influence of retinal adapta- Psychiatr 55:682, 1946. tion upon the pupillary reflex to light in normal man. Am J 5. Lowenstein O: Clinical pupillary symptoms in lesions of the Ophthalmol 51:644, 1961. optic nerve, optic chiasm, and optic tract. Arch Ophthalmol 27. Morgan SS, Hollenhorst RW, and Ogle KN: Speed of pupillary 52:385, 1954. light response following topical pilocarpine or tropicamide. 6. Lowenstein O: Pupillary reflex shapes and topical clinical Am J Ophthalmol 66:835, 1968. diagnosis. Neurology 5:631, 1955. 28. Yamuchi A, Lever JD, and Kemp KW: Catccholaminc load- 7. Lowenstein O and Friedman ED: Pupillographic studies: I. ing and depletion in the rat superior cervical ganglion. J Anat Present state of pupillography: Its method and diagnostic sig- 114:271, 1973. nificance. Arch Ophthalmol 27:969. 1942. 29. Loewenfeld I: Mechanisms of reflex dilatation of the pupil. 8. Lind NA and Shincbourne E: Studies on the development of Doc Ophthalmol 12:185, 1958.

Downloaded from iovs.arvojournals.org on 10/02/2021