Action spectrum for photophobia

1852 J. Opt. Soc. Am. A/Vol. 20, No. 10 /October 2003 Stringham et al.

Action spectrum for photophobia

James M. Stringham, Kenneth Fuld, and Adam J. Wenzel

Department of Psychology, University of New Hampshire, Durham, New Hampshire 03824

Received January 31, 2003; revised manuscript received May 27, 2003; accepted May 28, 2003

Thresholds for photophobia (light-induced discomfort) were determined at wavelengths from 440 to 640 nm forthree subjects. Photophobia was assessed by means of electromyography, which was used to measure subjects’level of squinting. After correction for absorption by macular pigment and the ocular media, subjects’ func-tions displayed a trend of increasing sensitivity with decreasing wavelength. We propose that the correctedfunction is indicative of increased sensitivity to potential retinal damage by short-wavelength light. It istherefore suggested that photophobia serves a function of biological protection. Results also suggest that pho-tophobia is significantly mitigated by macular pigment in the short wavelengths. © 2003 Optical Society ofAmerica

OCIS codes: 330.1720, 330.6180.

1. INTRODUCTIONPhotophobia results from exposure to light that is of suf-ficient intensity to produce discomfort. Typical behav-ioral responses to this discomfort include squinting, winc-ing, directing the eyes down toward the ground (in thecase of intense sunlight1), or simply looking away from anintense light stimulus. The perceived level of discomfortis typically commensurate with the level of aversive re-sponse. In the clinical literature, photophobia has beendefined as resulting in the induction or the exacerbationof pain upon exposure to light.2 Little is known aboutthe neurophysiological processes that mediate photopho-bia. Because pain-signaling fibers of the trigeminalnerve terminate in the dilator and constrictor muscles ofthe iris, it has been suggested that the vasodilation ac-companying the trigemino-pupillary reflex is the sensitiz-ing factor that makes the pupillary light reactionpainful.2,3 Indeed, it has been demonstrated that an in-tact trigeminal nerve is necessary to experiencephotophobia.2 Iris constriction, as might be expected,has been shown to increase with increasing illumination;at the point when the light intensity becomes truly dis-comforting, however, the iris appears to constrict and di-late irregularly every few seconds. This phenomenon isknown as hippus and is thought to occur because of simul-taneous, antagonistic activation of sympathetic and para-sympathetic pupillary responses.4 More recently, it hasbeen shown that hippus is not consistently associatedwith subjective reports of visual discomfort.5 King6 sug-gested that the rapidity of iris constriction, and intensestretching under lighting conditions that produce maxi-mal constriction, appear to be closely related to visual dis-comfort. Unfortunately, these factors have proved diffi-cult to measure and thus cannot be reliably used asobjective measures of visual discomfort.

In the clinical literature, photophobia is recognized as asymptom of various pathological conditions such as iritis,achromatopsia, conjunctivitis,3 trigeminal neuralgia,4 tu-mors compressing the anterior visual pathways,2 and,most prominently, migraine headache.7–13 Photophobia

1084-7529/2003/101852-07$15.00 ©

is also commonly noted as a symptom in albinism14 andhas also been found to coincide episodically with panicdisorder and depression.15–17 In the migraine headacheliterature, it has been suggested that photophobia experi-enced by migraineurs between attacks may be due to ab-normalities in spatial and temporal visual processing.10

Given that there are considerable backprojections fromthe striate cortex to the lateral geniculate nucleus, it hasbeen suggested that the apparent subcortical processingabnormalities (i.e., spatial and temporal) in migraineursproduce a geniculostriate feedback loop, which results incortical hypersensitivity and thus abnormal sensitivity tolight.11

A more thoroughly studied phenomenon related to pho-tophobia is discomfort glare. Discomfort glare is gener-ally defined as a subjective impression of discomfort uponexposure to light. Researchers studying discomfort glarehave adopted subjective rating scales, termed visual ana-log scales. The most popular rating scale used is the deBoer visual analog scale,18 with subjective ratings rang-ing from ‘‘just noticeable’’ to ‘‘unbearable.’’ There havebeen a few studies (e.g., Refs. 19 and 20), however, report-ing the use of electromyography as an objective measureof discomfort. Investigators from these studies measuredthe muscular activity associated with squinting andblinking around the ocular orbit upon exposure to a glarestimulus and found good agreement with subjective rat-ings of discomfort. General findings from discomfortglare studies (see Ref. 21) include the following: (a) In-creases in luminance give rise to increases in visual dis-comfort; (b) increases in pretest adaptation luminance re-sult in a decrease in visual discomfort; (c) increases insize of the test stimulus, while maintaining a specified lu-minance level, result in increased visual discomfort; (d)visual discomfort is greater for stimuli viewed centrallyrather than peripherally; and (e) discomfort thresholdsmeasured psychophysically are especially variable amongsubjects.

Whereas discomfort glare is certainly related to photo-phobia, it is difficult to make a direct comparison betweenthe two. In discomfort glare studies, subjects are rarely,

2003 Optical Society of America

Stringham et al. Vol. 20, No. 10 /October 2003 /J. Opt. Soc. Am. A 1853

if ever, exposed to lights that are subjectively judged as‘‘intolerable.’’ By contrast, photophobia involves an acuteintolerance to light, typically marked by some sort of be-havioral aversion (e.g., squinting or closing the eyes).

Because photophobia is usually thought of as a symp-tom of pathology, its assessment in normal subjects hasbeen limited.8,22,23 As a result, little is known about howbasic stimulus variables affect photophobia. For ex-ample, it is not known how the wavelength of light affectsphotophobia thresholds, and that is the purpose of thepresent study: to determine the action spectrum for pho-tophobia. Discomfort, not photophobia, thresholds tolights of various spectral compositions have been docu-mented in a few studies. Main et al.24 obtained subjec-tive discomfort thresholds in both migraineurs andhealthy subjects for three broadband spectral stimuli,which were judged to correspond to ‘‘blue,’’ ‘‘green,’’ and‘‘red’’ lights. Their subjects directly viewed light from aslide projector at a distance of 0.6 m. The intensity of thelights was increased by the experimenter until the subjectnoted that the stimulus was uncomfortable to view.Whether there was an aversion of some sort by the sub-jects was not stated. The method and the rate of increas-ing the intensity were not specified by the authors. Ad-ditionally, the angular size of the stimuli was notdisclosed (although we calculate the angle, assuming astandard projection lens diameter, to be approximately4.8 deg). Their healthy control subjects were found tohave lower discomfort thresholds, measured in photomet-ric units, for the short- and long-wavelength weightedstimuli, relative to the medium-wavelength weightedstimulus. Our effort to transform their photometric datainto radiometric data, in order to provide a rough energy-based action spectrum, proved to be heavily dependent onassumptions (e.g., spectral emission of the light source).It is therefore difficult to speculate on actual sensitivity.Moreover, because broadband (not monochromatic)stimuli were used, it was not possible to generate a trueaction spectrum from these data.

Flannagan et al.25 measured discomfort glare to six1-deg, monochromatic lights (ranging from 480 to 650nm), presented 7 deg peripheral to fixation. The re-searchers used a slide projector and interference filters topresent the stimuli. The subjects viewed light from theslide projector directly from a distance of 2.69 m. Sub-jects were instructed to rate, using a modified de Boer vi-sual analog scale, the various stimuli according to relativesubjective discomfort. Like Main et al.,24 they found thatrelative discomfort (based on photometric measures) wasgreater in the short- and long-wavelength regions as com-pared with that in the middle-wavelength region. To ex-amine their results as an action spectrum, we used radio-metric data provided by Flannagan et al.25 and convertedratings of ‘‘disturbing’’ into sensitivity measures. The re-sulting function is negatively sloped, indicating increas-ing sensitivity with decreasing wavelength.

The aforementioned studies have attempted to charac-terize the effect of intense lights of various spectral com-positions on visual discomfort. Although it appears thatthe action spectrum for visual discomfort may be differentfrom V(l) or any other common spectral sensitivity func-tion, the specific characterization of the wavelength de-

pendence of photophobia has yet to be sufficiently deter-mined. An action spectrum for photophobia could revealthe receptor mechanisms involved in mediating photopho-bia. Furthermore, the shape of the function could yieldinformation about postreceptoral contribution. For thepresent study, we consider photophobia to be an experi-ence of intolerable light intensity that induces a behav-ioral aversive response. We hypothesize that the squint-ing response would provide such a measure of aversion.The squinting response would seem to be an ecologicallyvalid measure of discomfort, as there would be no otherexplanation for a person to squint in the presence of anintense light other than to alleviate the discomfort asso-ciated with that light. From this reasoning, we derivedour operational definition of photophobia: Photophobiais experienced at the point when a light becomes suffi-ciently intense so as to elicit a criterion squinting re-sponse. In a manner similar to that used by Bermanet al.19 and Murray et al.,20 we used an objective method,electromyography (EMG), to record squinting responsesto briefly presented, intense lights.

2. METHODSA. SubjectsTwo of the authors (males, aged 27 and 30 yr) served assubjects. Additional data were obtained from a thirdsubject (female, aged 29 yr), who was naive to the specificaims of the experiment. All subjects were normaltrichromats with no history of visual pathology. To en-sure normal trichromacy, subjects’ color vision was as-sessed by using the Ishihara Pseudoisochromatic TestPlates and the Farnsworth Dichotomous Panel D-15 Test.The study was approved by the University of New Hamp-shire’s Institutional Review Board for the Protection ofHuman Subjects, and all subjects willingly consented toparticipate in the experiments.

B. ApparatusA three-channel standard Maxwellian-view system with a1000-W xenon arc lamp was used. In one channel, twolow-level red fixation points, oriented diagonally aboutthe center of the dark-adaptation stimulus, were used asa guide for subjects’ fixation. The second channel wasused for the presentation of a shuttered low-intensitywhite light, which served to assess subjects’ pretest dark-adaptation level. The third channel provided the teststimulus. Lights in the test stimulus channel were ren-dered monochromatic by Ditric Optics interference filters(mean half-bandwidth of 8.5 nm, ranging from 7 to 10.5nm). A neutral-density wedge was used to adjust the in-tensity of the test stimulus. Photophobia-inducing en-ergy levels were measured after each session with a UDTradiometer (Optometer 61).

A Grass Instruments model 7P3B EMG preamplifier/amplifier was coupled to one channel of a Gould Brush220 chart recorder for the analysis of squint muscle po-tentials. For precise determination of the portion of thechart recording that corresponded to stimulus presenta-tion, a photocell was positioned orthogonally to a beamsplitter through which light from the test channel passed.


Output from the photocell provided input for the secondchannel of the chart recorder.

C. ProcedureSubjects maintained a stable alignment by biting down ona premade dental impression and leaning their headagainst a forehead rest assembly. A pupillary alignmentprocedure was performed to ensure that the light fromthe optical system was in focus in the plane of the sub-ject’s pupil and passing through the center of the subject’spupil. A constant pretest dark-adaptation level was en-sured by subjects’ detection of a 5.0-log cd/m2, 1-s flash ofa white-light stimulus, presented every 6 s to the central5.75 deg of the retinas of their right eyes. Subjects wereinstructed to fixate the center of an imaginary line drawnbetween the two low-level, diagonally oriented, red fixa-tion points (separated by 7 deg) and attempt to detect thelow-level, white-light stimulus upon shuttered presenta-tion. Upon detection of three consecutive flashes, thesubject signaled the experimenter. The scotopic-levelwhite light was then presented continuously for 1 min,and the subject was instructed to maintain fixation. Thiswas done to ensure equal adaptation of the retinal area tobe stimulated by the test field. After the 1-min adapta-tion period, the subject was presented with the teststimulus, a light of predetermined wavelength and inten-sity, subtending 5.75 deg of visual angle, for 5 s.

EMG was used to record muscle potentials associatedwith subjects’ squinting responses. Surface electrodeswere attached to the right temple (reference), to the uppercheek, below the right eye’s lateral canthus (test), and onthe back of the neck (ground). EMG was used solely forthe determination of a threshold photophobia response;the criterion threshold photophobia response was basedon the amplitude and the duration of squinting. A con-tinuous squinting response that lasted at least half theduration of the test stimulus presentation and thatreached, at any point during the continuous squinting, a4:1 signal-to-noise ratio (squinting activity/baseline) wasconsidered a threshold photophobia response. Occasion-ally, a subject blinked or exhibited a ‘‘startle squint’’ uponpresentation of the test stimulus, followed by normal fixa-tion or some squinting (depending on the discomfortcaused by the test stimulus) for the remainder of thestimulus presentation. We attributed the startle reac-tion to the photic startle response.26 On traces where thephotic startle response was evident, we did not factor thestartle amplitude or duration into the photophobiathreshold criteria. Figure 1 shows an example of a typi-cal criterion photophobia response.

The method of ascending limits was employed. A rela-tively low-intensity test stimulus was initially presented,and, in steps separated by dark adaptation to a constantlevel, the intensity of the stimulus was increased in orderto reach the photophobia threshold.

The intervening dark-adaptation trials extended theduration of the experimental session considerably. As away of expediting each photophobia threshold measure-ment, the following procedure was employed. During theexperiment, after each trial, subjects were instructed torate the discomfort level of the test stimulus on a scale

from 1 to 10, 10 being their subjective experience of pho-tophobia. The rating was then inserted into a simpleequation,

I 5 ~10-R ! 3 0.05, (1)

where I 5 neutral density to be removed from the testchannel and R 5 subjective rating. For example, if thesubjective rating of the test stimulus were 7 on a giventrial then, for the subsequent trial, 0.15 log unit of neu-tral density would be subtracted from the test beam bymeans of the neutral-density wedge. This procedure wasrepeated until the subject reached the criterion photopho-bia response, as described above. A rating of 10 was notconsidered a measure of threshold photophobia—theEMG criteria described above were exclusively used to de-termine photophobia thresholds. In our many trials,however, we had only one coincidence of a 10 rating andnonphotophobia EMG criteria. In that case, we sub-tracted 0.10 log unit of neutral density from the testbeam; the ensuing trial produced a threshold photophobiaresponse, as determined by EMG.

Wavelengths from 440 to 640 nm in steps of 20 nm wereassessed. Because of the extreme length of each session(typically 2–3-h duration), three sessions were needed tocomplete an action spectrum; curves from each sessionwere least-squares fitted on the basis of photophobiathresholds at three normalizing wavelengths, obtainedduring each session: 520, 600, and 640 nm. In this way,any response variability arising from differences in elec-

Fig. 1. Illustration of a typical EMG recording from the presentstudy. Top panel: photocell output (amplitude spikes corre-spond to beginning and termination of test stimulus presenta-tion); bottom panel: typical threshold photophobia squinting re-sponse. Note startle response with ensuing brief refractoryperiod (a), followed by continuous squinting (b).


trode placement, subjects’ day-to-day absolute sensitivity,or output of the optical system could be offset by a scalarshift along the axis of ordinates. Two action spectra werecompleted for each subject.

3. RESULTSThe two photophobia action spectra @P(l)# for observersJS and AW are presented in Fig. 2. Peak sensitivities (onan energy basis) are normalized to zero. The similarityof the curves within and between subjects suggests thatthe methods used facilitated reliable measures. Bothsubjects’ data exhibit a large notch, with a minimum-sensitivity point corresponding to 460 nm. Both subjects’curves are relatively broad when compared with eitherV(l) or V8(l). In fact, an envelope of the combination ofV(l) and V8(l) makes for an approximate fit to our ob-server’s P(l) functions when either V(l) or V8(l) is al-lowed to shift along the Y axis for a least-squares fit (Fig.3). The prominent notch, however, remains as a majordiscrepancy between P(l) and the combined V(l) andV8(l) functions. The primary difference between the twosubjects’ curves is the wavelength of peak sensitivity.JS’s function peaks at 500 nm, whereas AW’s peaks at520 nm.

Fig. 2. Two complete photophobia action spectra @P(l)# for sub-jects JS and AW. Peak sensitivities are normalized to zero.

Fig. 3. Subjects’ averaged P(l) compared with the envelope ofV(l) and V8(l), adjusted along the Y axis for best fit.

Of particular interest to us was the large notch cen-tered at 460 nm. Common spectral sensitivity functionssuch as V(l) and V8(l) do not show this pronouncednotch. To explore further the nature of the notch and toexamine more closely the possible trend of increasing pho-tophobia sensitivity with decreasing wavelength below460 nm, we had both observers participate in a session inwhich photophobia thresholds were assessed at wave-lengths from 440 to 500 nm in steps of 10 nm. Addition-ally, a third observer was enlisted to confirm the overallshape of P(l), including three wavelengths within the re-gion of the notch. The results of these supplemental ses-sions are presented in the top panel of Fig. 4, where it canbe seen that 460 nm maintained its position as theminimum-sensitivity point relative to the other shortwavelengths tested. The curve of the third observer, NK,closely approximated that of the other subjects, includingthe appearance of the pronounced notch at 460 nm.

From the results presented in the top panel of Fig. 4, itwas apparent that, at wavelengths below 460 nm, oursubjects’ sensitivity increased with decreases of wave-length. In an attempt to determine if subjects’ sensitiv-ity would continue to increase at wavelengths shorterthan 440 nm, we tested at 420 nm. We did not have

Fig. 4. Top panel: Supplemental wavelengths (440–500 nm,10-nm steps) added to existing P(l) functions (subjects JS andAW). Subject NK, tested on six wavelengths, is plotted for com-parison. Functions are normalized to 500 nm. Bottom panel:P(l) functions, based on the scaling procedure described in thetext, for JS and AW. Functions are normalized to 500 nm.


enough energy at 420 nm to elicit a criterion squinting re-sponse in either subject, but both subjects viewed thiswavelength at maximum transmittance and each ratedthe experience a 9. Because we had obtained scalingdata during the experiment, we were then able to takerelative energy measurements for all wavelengths attheir 9 rating. For the few wavelengths that were neverrated a 9 but were rated 8 and 10, a linear interpolationwas used to determine our criterion response of 9. Al-though this procedure is admittedly crude, the resultingaction spectra bear similarity to those derived by EMG, asshown in the bottom panel of Fig. 4. In fact, a strong sig-nificant correlation was found between subjects’ ratings ofthreshold photophobia and EMG-derived thresholds (JS:r 5 0.976, p , 0.001; AW: r 5 0.982, p , 0.001).More importantly, the results suggest that sensitivity tolights below 440 nm increases.

4. DISCUSSIONA possible explanation for the pronounced notch found inour subjects’ P(l) functions could be the absorption oflight by macular pigment (MP). The wavelength of peakabsorption for MP is 460 nm,27–29 which is preciselywhere the point of minimum sensitivity is found in ourobservers’ P(l) functions. MP is symmetrically distrib-uted about the foveola; the optical density of MP de-creases exponentially from its peak in the foveola out-ward to an asymptote at approximately 5.5 deg of angularsubtense.30–32 Although the exact function of MP is notcertain, a protective role against damaging short-wavelength light has been hypothesized.33–36

MP optical density (MPOD) is conventionally measuredpsychophysically by the method of heterochromatic flickerphotometry (HFP). HFP thresholds are largely deter-mined by receptors stimulated by the edges of the testfields used.36 Thus the derived optical density values arefor MP at the retinal loci corresponding to where theedges of the field are. During the course of our study, wefound preliminary evidence for spatial summation of thephotophobia response over relatively large field sizes(from 1 to 6 deg). Preliminary data collected for an up-coming study also indicate that the photophobia responseexhibits robust spatial summation throughout a range ofstimulus sizes (from 5 to 30 deg). We therefore hypoth-esize that photophobia thresholds are determined by reti-nal mechanisms summed over the total area stimulatedby the test field, not just those located at the edges of thefield. Thus the attenuation factor by MP would be con-siderably greater than simply the HFP-derived MPODvalue, which reflects optical density at only the edges ofthe test field.

To correct our observers’ P(l) functions for the attenu-ation of MP, we mathematically integrated MPOD overthe area stimulated by the test field. This was done byusing spatial profiles of MPOD (previously obtained forthe temporal retina), measured at 460 nm, for subjects JSand AW. The integration was then simply doubled to ac-count for the absorption of light by MP over the nasal halfof the retina stimulated by the test stimulus. In obtain-ing the total absorption for MP in this way, one assumesthat MP is distributed symmetrically about the foveola.

This has generally been found to be the case.32 A furtherassumption is that the putative spatial summation effectfor photophobia is linear. Our preliminary data regard-ing spatial summation of the photophobia response sug-gest an approximately linear relationship between stimu-lus energy and stimulus area for stimuli subtending from5 to 30 deg.

Because the relative absorption of MP as a function ofwavelength has been well documented,27–29,34 we wereable to correct for MP attenuation at the other wave-lengths used in the study, relative to our 460-nm correc-tion value. We used the MP absorption template of Boneet al.29 to accomplish this. Ocular media absorption wascorrected for by using the values recommended byWyszecki and Stiles.27 The MP- and ocular-media-corrected P(l) functions are presented in Fig. 5. Withthe correction for MP attenuation and ocular media ab-sorption, the shape of the P(l) functions is essentiallymonotonic—a general trend of increasing sensitivity withdecreasing wavelength is apparent. As a general test ofour correction, we determined thresholds for photophobiato 460- and 540-nm lights at a retinal eccentricity outsidethat which contains MP. To elicit photophobia, we in-creased the stimulus to 22 deg. The subject was in-structed to fixate a small point of light located 31 deg tem-poral to the center of the test stimulus. This stimulusconfiguration ensured that the entire test stimulus wasplaced beyond the spatial extent of MP. In support of ourcorrection, the relationship between 460 and 540 nm wasfound to be reversed with respect to whether viewing wasfoveal or peripheral (Fig. 6). Owing to the large differ-ence in stimulus size and potential differential wave-length effects of spatial summation, however, direct com-parisons between central and peripheral viewingconditions regarding MP should be made tentatively.

The negative slope of the corrected functions is similarto the discomfort sensitivity data from Flannagan et al.,25

plotted for comparison in Fig. 5. It is important to notethat, in their study, the glare source was presented 7 degin the periphery (a retinal region that contains no MP).For this reason, we can make a direct comparison be-

Fig. 5. P(l) functions (subjects JS and AW), corrected for MPand ocular media absorption, compared with threshold-energyretinal damage function (Ham et al.37) and glare sensitivity data(Flannagan et al.25). The latter two functions are least-squaresfits to P(l).


tween our corrected functions (which reflect zero absorp-tion by MP) and their function. We plotted data for onlythe three wavelengths (out of six) from the study of Flan-nagan et al. for which radiometric data were providedthat corresponded to ratings of ‘‘disturbing’’ (i.e., equiva-lent to reaching photophobia threshold). Because Flan-nagan et al.25 presented their stimuli in free view, pupil-lary modulation of light entering the eyes may be thesource of the difference in slope between the two.

Our corrected functions are fitted to a first approxima-tion by the rhesus macaque threshold-energy retinaldamage function reported by Ham et al.37 (Fig. 5). Theyused laser illumination of various wavelengths at aparafoveal location to produce a threshold change in fun-dus appearance, which came about primarily as a resultof retinal pigment epithelium damage. Ham et al. notethat their action spectrum approximates the absorptionspectrum of melanin; melanosomes in the retinal pigmentepithelium appeared to be involved in the lesions. Theirtesting at a parafoveal region of the retina effectively pre-cluded the absorptive effects of MP. Although all four ofthe exposure-time-based sensitivity functions of Hamet al. fit our data well, their 1000-s data were used forcomparison, as those light levels were the closest in inten-sity to the light levels used in our study, albeit substan-tially more intense. In particular, the general similaritybetween the functions of increased sensitivity with de-creasing wavelength is striking. A similar relationshiphas been established by Gorgels and van Norren38 for therat retina.

We speculated that forming the basis of P(l) could bean uncommon photopigment, such as the recently charac-terized melanopsin,39 or perhaps a short-wavelength-absorbing photoproduct of rhodopsin, like metarhodopsinII.40 The absorption spectra of these pigments can beseen plotted together with the average corrected P(l) inFig. 7. Because Ham et al.37 suggested that melaninmay have played a role in their damage function, retinalmelanin (absorption) is also plotted for comparison. Fig-ure 7 shows that no single comparison pigment can ad-equately describe P(l). With the pigment absorptionspectra individually adjusted along the Y axis, a compos-ite fit to P(l) accounts to some extent for the shape of the

Fig. 6. Logarithm of the 460-nm/540-nm sensitivity ratio forsubjects JS and AW: central viewing versus 7-deg peripheralviewing.

corrected P(l). We suspect, however, that this very ap-proximate fit is the result of curve fitting and not the rev-elation of a physiological basis of photophobia.

Given our results, one could hypothesize that the func-tion of photophobia is biological protection. An aversiveresponse (squinting) that is biased in sensitivity towardpotentially damaging short-wavelength light is sugges-tive of this. Moreover, the close approximation of oursubjects’ corrected P(l) functions to the threshold-energyretinal damage function in rhesus monkeys found by Hamet al.37 further supports this conclusion.

Our data suggest that MP plays a major role in the at-tenuation of photophobia, substantially greater thanwhat would be expected from spatial averaging of MPOD.For photophobia, it appears that MP acts as a spatiallyintegrated filter. For this reason, even a small amount ofspatially integrated MP could prove to significantly re-duce photophobia mediated by central viewing.

With respect to MP’s role in attenuating photophobia, ifwe assume the P(l) notch to be primarily the result ofMP, then, hypothetically, a subject with very little or noMP might produce a P(l) function that indeed decreasesas a function of wavelength in a fashion similar to thatexhibited by our corrected P(l) functions. The authorsintend to pursue this point in a future study.

We maintain that the squinting response is a validmeasure of photophobia. However, the high correlationbetween subjects’ ratings of threshold photophobia andobjectively determined thresholds for photophobia sug-gests that, in future studies, photophobia thresholds mayconfidently be determined psychophysically by using care-fully controlled experimental procedures and scalingmethods such as those described in the present study.

ACKNOWLEDGMENTSThe authors thank Rob Drugan and Bill Stine and tworeferees for their comments and helpful suggestions re-garding the manuscript.

Fig. 7. Average of corrected P(l) functions (subjects JS and AW)compared with absorption spectra of retinal melanin, melanop-sin, and metarhodopsin II. Pigments have been adjusted on theY axis to allow for a composite best fit to P(l).


Address correspondence to Kenneth Fuld, Departmentof Psychology, Conant Hall, University of New Hamp-shire, Durham, New Hampshire 03824; e-mail, [email protected].

REFERENCES1. D. H. Sliney, ‘‘Ocular exposure to environmental light and

ultraviolet—the impact of lid opening and sky conditions,’’Dev. Ophthalmol. 27, 63–75 (1997).

2. J. E. Lebensohn, ‘‘The nature of photophobia,’’ Arch. Oph-thalmol. (Chicago) 12, 380–390 (1934).

3. J. E. Lebensohn, ‘‘Photophobia: mechanism and implica-tions,’’ Am. J. Ophthalmol. 34, 1294–1300 (1951).

4. R. Hopkinson, ‘‘Glare discomfort and pupil diameter,’’ J.Opt. Soc. Am. 46, 649–656 (1956).

5. P. Howarth, G. Heron, D. Greenhouse, I. Bailey, and S. Ber-man, ‘‘Discomfort from glare: the role of pupillary hippus,’’Light. Res. Technol. 25, 37–42 (1993).

6. V. King, ‘‘Discomfort glare from flashing sources,’’ J. Am.Optom. Assoc. 43, 53–56 (1972).

7. H. G. Wolff, Headache and Other Head Pain, 2nd ed. (Ox-ford U. Press, New York, 1963).

8. B. Eckhardt, J. M. McClean, and H. Goodell, ‘‘Experimentalstudies on headache: the genesis of pain from the eye,’’Res. Publ. Assoc. Res. Nerv. Ment. Dis. 23, 209–227 (1943).

9. P. Drummond, ‘‘A quantitative assessment of photophobiain migraine and tension headache,’’ Headache 26, 465–469(1986).

10. D. M. Coleston, E. Chronicle, K. H. Ruddock, and C. Ken-nard, ‘‘Precortical dysfunction of spatial and temporal vi-sual processing in migraine,’’ J. Neurol. Neurosurg. Psy-chiatry 57, 1208–1211 (1994).

11. E. P. Chronicle and W. M. Mulleners, ‘‘Visual system dys-function in migraine: a review of clinical and psychophysi-cal findings,’’ Cephalalgia 16, 525–535 (1996).

12. P. Drummond and A. Woodhouse, ‘‘Painful stimulation ofthe forehead increases photophobia in migraine sufferers,’’Cephalalgia 13, 321–324 (1993).

13. P. Drummond, ‘‘Photophobia and autonomic responses to fa-cial pain in migraine,’’ Brain 120, 1857–1864 (1997).

14. J. Perez-Carpinell, P. Capilla, and J. Morales, ‘‘Vision de-fects in albinism,’’ Optom. Vision Sci. 8, 623–628 (1992).

15. M. Kellner, K. Wiederman, and J. Zihl, ‘‘Illumination per-ception in photophobia patients suffering from panic disor-der with agoraphobia,’’ Acta Psychiatr. Scand. 96, 72–74(1997).

16. H. Gerbaldo, ‘‘Depression and photophobia behavior (let-ter),’’ Am. J. Psychiatry 145, 1479–1480 (1988).

17. S. F. Signer and Y. D. Lapierre, ‘‘Photophobia in depression(letter),’’ Am. J. Psychiatry 10, 1355–1356 (1989).

18. J. B. de Boer, ‘‘Visual perception in road traffic and the fieldof vision of the motorist,’’ in Public Lighting, J. B. de Boer,ed. (Phillips Technical Library, Eindhoven, The Nether-lands, 1968), pp. 11–96.

19. S. Berman, M. Bullimore, R. Jacobs, I. Bailey, and N. Gan-dhi, ‘‘An objective measure of discomfort glare,’’ J. Illum.Eng. Soc. 23 (No. 2), 40–48 (1994).

20. I. Murray, S. Plainis, and D. Carden, ‘‘The ocular stressmonitor: a new device for measuring discomfort glare,’’Light. Res. Technol. 34, 231–242 (2002).

21. C. Waters, R. Mistrick, and C. Bernecker, ‘‘Discomfort glarefrom sources of nonuniform luminance,’’ J. Illum. Eng. Soc.24 (No. 2), 73–84 (1995).

22. K. Siegwart, ‘‘Zur Frage nach dem Vorkommen und demWesen des Blendungsschmerzes,’’ Schweiz Med. Wochen-schr. 50, 1165–1172 (1920).

23. J. D. Wirtschafter and C. M. Bourassa, ‘‘Binocular facilita-tion of discomfort with high luminances,’’ Arch. Ophthal-mol. (Chicago) 75, 683–688 (1966).

24. A. Main, I. Vlachonikolis, and A. Dowson, ‘‘The wavelengthof light causing photophobia in migraine and tension-typeheadache between attacks,’’ Headache 40, 194–199 (2000).

25. M. Flannagan, M. Sivak, M. Ensing, and C. Simmons, ‘‘Ef-fect of wavelength on discomfort glare from monochromaticsources,’’ Tech. Rep. UMTRI-89-30 (University of MichiganTransportation Research Institute, Ann Arbor, Mich., 1989),pp. 1–19.

26. S. K. Yates and W. F. Brown, ‘‘Light-stimulus-evoked blinkreflex: methods, normal values, relation to other blink re-flexes, and observations in multiple sclerosis,’’ Neurology 3,272–281 (1981).

27. G. Wyszecki and W. S. Stiles, Color Science, 2nd ed. (Wiley,New York, 1982).

28. J. J. Vos, ‘‘Literature review of human macular absorptionin the visible and its consequences for the cone receptor pri-maries,’’ Rep. IZF 1972-17 (Institute for Perception RVO-TNO, Soesterberg, The Netherlands, 1972).

29. R. A. Bone, J. T. Landrum, and A. Cains, ‘‘Optical densityspectra of the macular pigment in vivo and in vitro,’’ VisionRes. 32, 105–110 (1992).

30. D. M. Snodderly, J. D. Auran, and F. C. Delori, ‘‘The macu-lar pigment. II. Spatial distribution in primate retinas,’’Invest. Ophthalmol. Visual Sci. 25, 674–685 (1984).

31. R. A. Bone, J. T. Landrum, L. Fernandez, and S. L. Tarsis,‘‘Analysis of the macular pigment by HPLC: retinal distri-bution and age study,’’ Invest. Ophthalmol. Visual Sci. 29,843–849 (1988).

32. B. R. Hammond, B. R. Wooten, and D. M. Snodderly, ‘‘Indi-vidual variations in the spatial profile of macular pigment,’’J. Opt. Soc. Am. A 14, 1187–1196 (1997).

33. M. Schultze, Ueber den gelben Fleck der Retina, seinen Ein-fluss auf normales Sehen und auf Farbenblindheit (Max Co-hen & Sohn, Berlin, 1866).

34. D. M. Snodderly, P. K. Brown, F. C. Delori, and J. D. Auran,‘‘The macular pigment. I. Absorbance spectra, localiza-tion and discrimination from other yellow pigments in pri-mate retinas,’’ Invest. Ophthalmol. Visual Sci. 25, 660–672(1984).

35. B. R. Hammond, B. R. Wooten, and D. M. Snodderly, ‘‘Pres-ervation of visual sensitivity of older subjects: associationwith macular pigment density,’’ Invest. Ophthalmol. VisualSci. 36, 1–10 (1998).

36. J. S. Werner, S. K. Donnelly, and R. Kliegl, ‘‘Aging and hu-man macular pigment density,’’ Vision Res. 2, 257–268(1987).

37. W. T. Ham, H. A. Mueller, and D. H. Sliney, ‘‘Retinal sensi-tivity to damage from short wavelength light,’’ Nature (Lon-don) 260, 153–155 (1976).

38. T. Gorgels and D. van Norren, ‘‘Ultraviolet and green lightcause different types of damage in rat retina,’’ Invest. Oph-thalmol. Visual Sci. 5, 851–863 (1995).

39. R. J. Lucas, S. Hattar, M. Takao, D. M. Berson, R. G. Foster,and K. W. Yau, ‘‘Diminished pupillary light reflex at high ir-radiances in melanopsin-knockout mice,’’ Science 299, 245–247 (2003).

40. C. Grimm, C. E. Reme, P. O. Rol, and T. P. Williams, ‘‘Bluelight’s effects on rhodopsin: photoreversal of bleaching inliving rat eyes,’’ Invest. Ophthalmol. Visual Sci. 41, 3984–3990 (2000).

Documents

Action spectrum for photophobia