天津职业大学 文本 Dynamic Accommodation Responses to Stationary Colored Targets_眼屈光学

【T中0m。+学号方 ORIGINAL ARTICLE Dynamic Accommodation Responses to Stationary Colored Targets DAVID A.ATCHISON,PhD,FAAO,NIALL C.STRANG,PhD, and LAWRENCE R.STARK,PhD,FAAO Schoi fOptamery.Qaremiand Uuiweriy f Trealngy Keiin Grv,Qed Asurraiu (DAA.LRS)and Depurtmenr ofVa Sciemces Glngow Caledouian Univenity,GLegn,Scoland (NCS) ABSTRACT:Purpose.When the targets or the background in a display are different colors,longitudinal chromatic aberration ensures that there is no single correct accommodative response.The purpose of the present study was to determine whether the response becomes more variable when viewing certain multicolor displays.Methods.Accom- modative responses of five young participants were measured with a dynamic infrared optometer while they viewed steady targets at a nominal stimulus level of 3 D.Target-on-background color combinations were black on white,black on blue,black on red,blue on red,red on blue,dark blue on red,and dark red on blue.Resufts.When compared with the standard hlack-on-white target,responses to targets with reduced spectral bandwidth were not significantly more variahle.In most participants,responses to near-isoluminant targets (e-g.red on blue and bluc on red)were not more variable than to the standard target.However,calculated confidence intervals cannot rule out moderate to large changes in variability near isoluminance.Responses to these multicolor targets tended to favor the blue focus. Concfusions.In most individuals,viewing multichromatic targets does not increase significantly the variation in accommodative response as compared with hroadhand black-and-white targets.(Optom Vis Sci 2004:81:699-711) Key Words:chromatic aberration,dynamic accommodatin,color,color visual display unit (VDU).lighting any studies have mesured accommodation when an curyand stability of the acoommodarion reponse to coloe appropriace accommodation respone can provide an srimuls. unambiguously focusedeinaimgRelatively ewer Redundancy of Accommodative System Cues.Accom studies have used oupditions in which there is an ambiguous ac modacion can be driven by a mumber of stimuli.including defocus, commodation stinlus.Such ambiguities can result from the physical lyout of objects in the environmento from optical Perccived distance,nryr,and vergThe response chrfhyessuchsgo to defocus is o maltiply redundant.driven by blur,1CA. and an unnown"achnomai"ceFxamples are availble in the and longitudinal chrumatic aberration (LCA). literature to show that a process operaes to compensate effectively causc of the cyc's四ral LCA,”ffenent colored objocts and for poor or absent stimuli.Thus,displys that upset the or bacgrounds in the display are imaged at dfferent plnes in the eye mal culor stimuli to accummodation may not noocorily disurh with respec to thereina thus.there may be no single position of optimal focus.Some situations in which this ocur in daily lifear the reiponse if ocher reluable cues and stimuli can substinute when viewing (1)multiolorod diplnys aich as comnputer visal Spectra!Bandwidth Under beuadhand illumination.LCA disply units (VDUg)or cnlnt televiion scrern&(2)enlor printed prowides an imporanstimuls to accommodation for moving and matter under broadband illuminants,or (3)achromatic material stationry uargets.On a color VDU.broadband spcral under trichmic illuminantssuh asprime oolor"lamps. litirs are obtainnd using all thne (or xnmnctimes twn)ofthe red, green.and blue primaries.Ie sbould be noted that for achromtic targets,the various wavelengths that malce up the whise hack- Factors in the Response to Color Stimuli ground ar imagod at dit形rn时planes in the ye.nouting in blt spread fianctions with characterisic colored fringes.Thus, Recene work accommodatian.colar,and ICA provides though the wavelengths are food at different planes.this cand busis for understanding a mumber of factors that are involved an the tion is pecesay for refles accommodation to func omy. Opowwiry and Vila Vol 81.No S.Spimbe 2001

ORIGINAL ARTICLE Dynamic Accommodation Responses to Stationary Colored Targets DAVID A. ATCHISON, PhD, FAAO, NIALL C. STRANG, PhD, and LAWRENCE R. STARK, PhD, FAAO School of Optometry, Queensland University of Technology, Kelvin Grove, Queensland, Australia (DAA, LRS), and Department of Vision Sciences, Glasgow Caledonian University, Glasgow, Scotland (NCS) ABSTRACT: Purpose. When the targets or the background in a display are different colors, longitudinal chromatic aberration ensures that there is no single correct accommodative response. The purpose of the present study was to determine whether the response becomes more variable when viewing certain multicolor displays. Methods. Accom￾modative responses of five young participants were measured with a dynamic infrared optometer while they viewed steady targets at a nominal stimulus level of 3 D. Target-on-background color combinations were black on white, black on blue, black on red, blue on red, red on blue, dark blue on red, and dark red on blue. Results. When compared with the standard black-on-white target, responses to targets with reduced spectral bandwidth were not significantly more variable. In most participants, responses to near-isoluminant targets (e.g., red on blue and blue on red) were not more variable than to the standard target. However, calculated confidence intervals cannot rule out moderate to large changes in variability near isoluminance. Responses to these multicolor targets tended to favor the blue focus. Conclusions. In most individuals, viewing multichromatic targets does not increase significantly the variation in accommodative response as compared with broadband black-and-white targets. (Optom Vis Sci 2004;81:699–711) Key Words: chromatic aberration, dynamic accommodation, color, color visual display unit (VDU), lighting Many studies have measured accommodation when an appropriate accommodation response can provide an unambiguously focused retinal image.1 Relatively fewer studies have used conditions in which there is an ambiguous ac￾commodation stimulus. Such ambiguities can result from the physical layout of objects in the environment2–4 or from optical characteristics of the eyes, such as astigmatism,5 anisometropia,6, 7 and longitudinal chromatic aberration (LCA).8 Because of the eye’s natural LCA,9 different colored objects and backgrounds in the display are imaged at different planes in the eye with respect to the retina; thus, there may be no single position of optimal focus. Some situations in which this occurs in daily life are when viewing: (1) multicolored displays such as computer visual display units (VDUs) or color television screens, (2) color printed matter under broadband illuminants, or (3) achromatic material under trichroic illuminants such as “prime color” lamps.10 Factors in the Response to Color Stimuli Recent work on accommodation, color, and LCA provides a basis for understanding a number of factors that are involved in the accuracy and stability of the accommodation response to color stimuli. Redundancy of Accommodative System Cues. Accom￾modation can be driven by a number of stimuli, including defocus, perceived distance, voluntary effort, and vergence.1 The response to defocus is also multiply redundant, driven by blur,11 LCA,12, 13 and an unknown “achromatic” cue.14 Examples are available in the literature to show that a process operates to compensate effectively for poor or absent stimuli.15, 16 Thus, displays that upset the nor￾mal color stimuli to accommodation may not necessarily disturb the response if other reliable cues and stimuli can substitute. Spectral Bandwidth. Under broadband illumination, LCA provides an important stimulus to accommodation for moving and stationary targets.12, 13, 17–20 On a color VDU, broadband spectral qualities are obtained using all three (or sometimes two) of the red, green, and blue primaries.21 It should be noted that for achromatic targets, the various wavelengths that make up the white back￾ground are imaged at different planes in the eye, resulting in blur spread functions with characteristic colored fringes.12 Thus, al￾though the wavelengths are focused at different planes, this condi￾tion is necessary for reflex accommodation to function normally.12 1040-5488/04/8109-0699/0 VOL. 81, NO. 9, PP. 699–711 OPTOMETRY AND VISION SCIENCE Copyright © 2004 American Academy of Optometry Optometry and Vision Science, Vol. 81, No. 9, September 2004

700 Accummodation for Staliorury Colured Tangels-Alchison el al. In optical sytems designed to rduie proximal cuimesigated temporal stability of the response,on was limitodtu mudation,the accmmodation respons to dynamic tats rends ambyopicandno to became ponrer in manochromatie or reduerd spectral hand- Collins ct al.used natural viewing conditioes and widch lighs.22 For sedy tgethe accommodation re- found no signifcant difference in response vabiliry berween var- sponse becomes less stable,shoing low frtquency drifs and in. ious achromatic and unequally luminant chromaric (e red on creased fluctuations that suggest subjective difficulry.The blue,blue on red,and green on blackl color conditions presented amplitde of accommodation may also fatigue slighty during oa VDU.Three studies hive imestigted the fluctuations af work perfoemed under monochromatic umination.2 Neverthe- accommodarion.Denicul and Corao-Mtinused combinations les,the time averged response to steady targets can be ccuae of black or whice text ona colored field or colored text on ablack regardess of spectra bandwidtho the aount of LCA o white field.Combinations of colored text on a colored back present.A minority of individuals also appear to focus normally ground were not investigated.For black text or black backgrounds, in monochromatic lhe.12.19..12 they poted an increase in the ratio of high temporal fiequency to /sokmrninance.Scudies using targets presented in Badal opti low temporal frequency xctiviy (HF/LF)s the color portion be cal systems have found that multicolored targets near nominal came ess saturated.For nominally isoluminant tarpers (e.g.whice ilaminance provide a poor stimulus to reflexive accummoda- on culor and oolor on white),there were no cunsistent changes in tion.Comerady,udics using natural viewing cunditions xccommodanive Actuations wich chromariciry.Gray ct al. and sbjecively brightess matched targets have found that their found no difference in reponse variabiliry (RMS value)or the participants'steady accommodation was reasocably accurate with flucuations of accommodation between black-and-white targets these nominally isolminant targets.One posible explna- and a saturaned red target on a bbk background.Simmercal tion fur thesc contradictury noults is that in natural viewing on- imegigated the accommodative flucruations with a muamber of ditions.pricpunes use prosiml and voluntary accommodation different tintd eindividually preserihed to rdeasthenopic to compenste for peoe"reflexive'accommodation. complaints However,the spectral bandwidths and color carac Conficting Stim.Multicnlored displays have the poten- teristics of the indidly proibed s vaid widdy within tial to pmvide cunflicring xrimuli ro the accommndation system. each eaperimental oomdirian.making it difficnle to apply thoss such that no sinle accommodation response provides an oprimal findangs to the questions under investigatinn in the present study level of clariry.This will occur when the rwo colars ae sufficendy WWhen viewing multicolored displays,the respoeue coul be- satuatedand the dference in refractive eror berween their come more varable because (1)the stimulus to accommodation daminint wavdengthe is larger than the cye's depth of focus.If impoverishad in see wary,perhaps beeause of reduced spectral there are patches of test of different colots on an achromatic bandwidth or reduced luminance contrast;(2)the background,then accommodation may be forced to flucruate conficting dioperi stimuli provided by the calors in the target up and down with eich new fixarion.For example,combina destabilize the neural control mechanisms of accommodation;oc tions of red and black have generally been found to lead to (3)there is a conscious strategy toer cmmodation dynami higher accommodation responses than combinations of blue clly berween the focus levels that prowide distinct vision for the and black,.21.25.31.although depending on the tar- colrs peesent in the display.We hypotheed that the accomm gets involved,this is not always the case,.10 and some dation response becomes more variable under some coloc combi practice may be required to change focus between the colors. ons of arget and backgroand and cested this hypothesis using For a olord target on a different clored bockground,there an infrarcd dymamic objoctive optumter. may he no singe accummndative npons that can mke the targst boundary distinct.The group averige sedystate reponse when METHODS viewing sch targets in printed fomoron color VDUs does nat vary grtly bctwoen colar comhintions.Huwever,Charman Fiv individuals3dtn22nd3引er(man,266生 nd similar targrts hut conmtad an the rsponse patterns of 3.6 yers)participared in the sudy.Thry had reglighle aigma- individual purticipants When viewing red-an-ble ce hle-on-ted tism (D_25 D)and narmal suhjective amplitudes of accommo- targets.participants were found who consistently focused for (1) dation (6 D)in their tested right eyes.All the participants apart the nd wavdengths,(2)the bluc wadength,(3)the bckgromnd from rwo participating authons were naie as to the purpos of tha calor,(4)the target cnlar,ar (5)who were equally hippy to foeus experiment.This study adhered to the Dheelararion of Helsinki. fur either the red or blue warvelngth,ar (61 who focused at much The QUT Researh Ethics Committee approved the study.with the same level for all the targets.This study demonstraes that infoemed coment prowided by the participants therc an iclinsyrcratic individual cliferrnces in arcommocatinn to Acommocrion of parti:ipinrs'cyo wa measured using a multioolored displays that cannoe be eaphined by a simple appli- modified Ratsch and Lamh Ophthalmetron infrared opeometer ction of the various factors tht have previously been discsd (Rochester.NY).Targets were presented with a Maxwellian- view Badal optomeer tched to the Ophthametron.The Response Stability viewing sysem contributed 0.I D of LCA across the visible spectrum.To preveat spurious redings by the Ophthametron,I Although there have been studies of the time-averaged response drop of 25%phenylephrine was used to dilate pupils.A 5mm to sedy colored targets and of the response to dynamic coloced aperture conjugate with the eye's entrance pupil formed the aper- targets.less i known of the sabaliry and variability of the steady ture stop of the optical system. stae repoose to stationary oolored tarpets.Of those studies that Absolute calibrations of the Ophthalmesron were obtained for Opromrty aty Viniow Sciener,Vol.51.No.9.Scgarbet 2004

In optical systems designed to reduce proximal cues to accom￾modation, the accommodation response to dynamic targets tends to become poorer in monochromatic or reduced spectral band￾width light.12, 18, 19, 22 For steady targets, the accommodation re￾sponse becomes less stable, showing low frequency drifts and in￾creased fluctuations that suggest subjective difficulty.20 The amplitude of accommodation may also fatigue slightly during work performed under monochromatic illumination.23 Neverthe￾less, the time-averaged response to steady targets can be accurate, regardless of spectral bandwidth20, 22, 24–27 or the amount of LCA present.28 A minority of individuals also appear to focus normally in monochromatic light.12, 19, 20, 22 Isoluminance. Studies using targets presented in Badal opti￾cal systems have found that multicolored targets near nominal isoluminance provide a poor stimulus to reflexive accommoda￾tion.29, 30 Conversely, studies using natural viewing conditions and subjectively brightness-matched targets have found that their participants’ steady accommodation was reasonably accurate with these nominally isoluminant targets.8, 31, 32 One possible explana￾tion for these contradictory results is that in natural viewing con￾ditions, participants use proximal and voluntary accommodation to compensate for poor “reflexive” accommodation.33 Conflicting Stimuli. Multicolored displays have the poten￾tial to provide conflicting stimuli to the accommodation system, such that no single accommodation response provides an optimal level of clarity. This will occur when the two colors are sufficiently saturated,34, 35 and the difference in refractive error between their dominant wavelengths9 is larger than the eye’s depth of focus. If there are patches of text of different colors on an achromatic background, then accommodation may be forced to fluctuate up and down with each new fixation. For example, combina￾tions of red and black have generally been found to lead to higher accommodation responses than combinations of blue and black,8, 21, 25, 31, 32, 34–38 although depending on the tar￾gets involved, this is not always the case,35, 39, 40 and some practice may be required to change focus between the colors.25 For a colored target on a different colored background, there may be no single accommodative response that can make the target boundary distinct. The group average steady-state response when viewing such targets in printed form40 or on color VDUs31, 41 does not vary greatly between color combinations. However, Charman8 used similar targets but concentrated on the response patterns of individual participants. When viewing red-on-blue or blue-on-red targets, participants were found who consistently focused for (1) the red wavelengths, (2) the blue wavelengths, (3) the background color, (4) the target color, or (5) who were equally happy to focus for either the red or blue wavelength, or (6) who focused at much the same level for all the targets.8 This study demonstrates that there are idiosyncratic individual differences in accommodation to multicolored displays that cannot be explained by a simple appli￾cation of the various factors that have previously been discussed. Response Stability Although there have been studies of the time-averaged response to steady colored targets and of the response to dynamic colored targets, less is known of the stability and variability of the steady￾state response to stationary colored targets. Of those studies that investigated temporal stability of the response, one was limited to amblyopic eyes,42 and three others applied no statistical analy￾ses.20, 37, 39 Collins et al.41 used natural viewing conditions and found no significant difference in response variability between var￾ious achromatic and unequally luminant chromatic (e.g., red on blue, blue on red, and green on black) color conditions presented on a VDU. Three studies have investigated the fluctuations of accommodation. Denieul and Corno-Martin37 used combinations of black or white text on a colored field or colored text on a black or white field. Combinations of colored text on a colored back￾ground were not investigated. For black text or black backgrounds, they noted an increase in the ratio of high temporal frequency to low temporal frequency activity (HF/LF) as the color portion be￾came less saturated. For nominally isoluminant targets (e.g., white on color and color on white), there were no consistent changes in accommodative fluctuations with chromaticity. Gray et al.27 found no difference in response variability (RMS value) or the fluctuations of accommodation between black-and-white targets and a saturated red target on a black background. Simmers et al.43 investigated the accommodative fluctuations with a number of different tinted lenses individually prescribed to reduce asthenopic complaints. However, the spectral bandwidths and color charac￾teristics of the individually prescribed lenses varied widely within each experimental condition, making it difficult to apply those findings to the questions under investigation in the present study. When viewing multicolored displays, the response could be￾come more variable because (1) the stimulus to accommodation is impoverished in some way, perhaps because of reduced spectral bandwidth12, 18–20 or reduced luminance contrast;29, 30 (2) the conflicting dioptric stimuli provided by the colors in the target destabilize the neural control mechanisms of accommodation; or (3) there is a conscious strategy to alter accommodation dynami￾cally between the focus levels that provide distinct vision for the colors present in the display. We hypothesized that the accommo￾dation response becomes more variable under some color combi￾nations of target and background and tested this hypothesis using an infrared dynamic objective optometer. METHODS Five individuals aged between 22 and 31 years (mean, 26.6
3.6 years) participated in the study. They had negligible astigma￾tism (0.25 D) and normal subjective amplitudes of accommo￾dation (
6 D) in their tested right eyes. All the participants apart from two participating authors were naive as to the purpose of the experiment. This study adhered to the Declaration of Helsinki. The QUT Research Ethics Committee approved the study, with informed consent provided by all the participants. Accommodation of participants’ eyes was measured using a modified Bausch and Lomb Ophthalmetron infrared optometer (Rochester, NY).44 Targets were presented with a Maxwellian￾view Badal optometer attached to the Ophthalmetron.44 The viewing system contributed 0.1 D of LCA across the visible spectrum. To prevent spurious readings by the Ophthalmetron, 1 drop of 2.5% phenylephrine was used to dilate pupils. A 5-mm aperture conjugate with the eye’s entrance pupil formed the aper￾ture stop of the optical system. Absolute calibrations of the Ophthalmetron were obtained for 700 Accommodation for Stationary Colored Targets—Atchison et al. Optometry and Vision Science, Vol. 81, No. 9, September 2004

Accommodation for Staticnary Colored Targets-Axchison al.701 each participant Various levels of accommodation were simulted with trial lenses in fromt of the eyclopleged cye.Tanear regressinn equations were fimed to the optometer reading (volt (V)]as a furction of simulatod refractive crror,taking care to accuunt for the distance ocular refraction,referened to the entrance pupil,and 0.8 the verex distance of trial knses Slopes [VD (dioper)l of regres- sions were simlar,but the intererpts varied betwren pirticipints To test the signal-to-noisc ratio ofthe instrument,10 ncordings 0.6 were taken from the mudel cyc supplicd with the Ophthalmeron, and 5 and 9 reoording were taken from two cydoplegrd human eyes(1 drop of 1%cyclopentolate:subjective amplitude of accom modation.<04 D).The rexponse was sampled at 35.97 Hx for 0.4 illuminant A+whKe 14.2 s.The root-mean-square levd in the model cye was 0.024 D. red and the root-mcan-quire kevds in the tau cydopkged cyes were dark red 0.06 and 0.10 D,indicating little influence of noise on optometer 0.2 dark blue readings. Stimuli were 35-mm slide transparencics of a Maltese cross blue whoee dtails are shown in Fig.1.Kodak Ortholith and Kodak 常 Ektachrome (64 ASA)transparency films (Ruchester.NY)were usd.Seven target-on-badiground color combinations were used 02 D.4 D.6 D.8 (1)black on white:(2)black oa blue:(3)black on red:(4)blue on red:(5]dark blue on red:(G)red on blue:and (7)dark red on blue. X The luminanes and color characteristics of the side targets were FIGURE 2. measured at the Maxwdlian view imge plane with a Topcon Chmmalicity toudinles (x.y)fur fe coer langts pinted in the cnlr BM-7 luminance colorimeter (Tokyo.lapan).Spectral transmit space of the CE 1931 observer hhite:0.523,0.373;red:0.705,0284: tances of the slide targets wereo measured separatey (Beckman bluw a150069:dack redl:0604.0337:;dark bl0.200.158: DU-650 spectrophotometer.Fullerton.CA). lluminant A0.4476,0.4075l. Chromaticiry ooordinates for the targets are shown in Fig 2. Dominant wavelengths (reative to CIE illuminant A)were 648 nm (re),466 nm (blc),632 nm (dark rod).and 467 nm (dark bluc).Estimatod spoctral charactcristics of the target colors weighted by the phocopic luminous efficienty function'V)an shown in Fig.3.The spectral charcteristics of the tungsten source were approximated by that of a black body based on the measured DtRo对 0 6.49- 410 70 Wavelength (nm) 10 FIGURE 3. FIGLRE 1. Relative spectral energy distributions of the targes colors.weighoed by the Maltese rro tnpet ued n perment.The hkmdd of the 10 fuld is pronvidod by a field stop at +83 D distal to the far point of the eye.Various ther respective peaks.Plts ane given for the whie,red.blue,dark red. rrons and harkgmind cnor combintions wer uned (e Mfehnckl and clark hhe rrlnrs

each participant. Various levels of accommodation were simulated with trial lenses in front of the cyclopleged eye. Linear regression equations were fitted to the optometer reading [volt (V)] as a function of simulated refractive error, taking care to account for the distance ocular refraction, referenced to the entrance pupil, and the vertex distance of trial lenses. Slopes [V/D (diopter)] of regres￾sions were similar, but the intercepts varied between participants. To test the signal-to-noise ratio of the instrument, 10 recordings were taken from the model eye supplied with the Ophthalmetron, and 5 and 9 recordings were taken from two cyclopleged human eyes (1 drop of 1% cyclopentolate; subjective amplitude of accom￾modation, 0.4 D). The response was sampled at 35.97 Hz for 14.2 s. The root-mean-square level in the model eye was 0.024 D, and the root-mean-square levels in the two cyclopleged eyes were 0.06 and 0.10 D, indicating little influence of noise on optometer readings. Stimuli were 35-mm slide transparencies of a Maltese cross, whose details are shown in Fig. 1. Kodak Ortholith and Kodak Ektachrome (64 ASA) transparency films (Rochester, NY) were used. Seven target-on-background color combinations were used: (1) black on white; (2) black on blue; (3) black on red; (4) blue on red; (5) dark blue on red; (6) red on blue; and (7) dark red on blue. The luminances and color characteristics of the slide targets were measured at the Maxwellian view image plane with a Topcon BM-7 luminance colorimeter (Tokyo, Japan). Spectral transmit￾tances of the slide targets were also measured separately (Beckman DU-650 spectrophotometer, Fullerton, CA). Chromaticity coordinates for the targets are shown in Fig. 2. Dominant wavelengths (relative to CIE illuminant A) were 648 nm (red), 466 nm (blue), 632 nm (dark red), and 467 nm (dark blue). Estimated spectral characteristics of the target colors weighted by the photopic luminous efficiency function45 (V) are shown in Fig. 3. The spectral characteristics of the tungsten source were approximated by that of a black body based on the measured FIGURE 1. Maltese cross target used in experiment. The blurred edge of the 10° field is provided by a field stop at 8.3 D distal to the far point of the eye. Various cross and background color combinations were used (see Methods). FIGURE 2. Chromaticity coordinates (x, y) for the color targets plotted in the color space of the CIE 1931 observer (white: 0.523, 0.373; red: 0.705, 0.284; blue: 0.153, 0.068; dark red: 0.604, 0.337; dark blue: 0.230, 0.158; illuminant A: 0.4476, 0.4075). FIGURE 3. Relative spectral energy distributions of the target colors, weighted by the photopic relative luminous efficiency function (V), and normalized to their respective peaks. Plots are given for the white, red, blue, dark red, and dark blue colors. Accommodation for Stationary Colored Targets—Atchison et al. 701 Optometry and Vision Science, Vol. 81, No. 9, September 2004

702 Accommodation for Stationary Colorod TargetsAtchison ct al. color temperamre ofthe source (2469 K).5 The esciration puritics blockes of7 rrials cach.Within a block,the proentation order for of the targets were 0.28 for whie,0.93 for nod and blue,0.60 for cach of the seven target-on-background combinations was random dark rod,and 0.69 for dark bluc.All the colored targets show without replacement.Participants were instructed to"lookt che secondary mudes in their spectra distributions(Fig.3).In the case tanget naturally,the sme as you would when nommlly reading a of the red and blue targers,these modes are small,and the excita book or sign at the same distance." tion purities remain high (0.93 in boch cases).However,the modes Dark focus measurements were made with a Canon Aucoref RI are important for the dark nod and dark blue tangets (Fig.3) autorcfracor (Tokyo,Japan).Each participant was left in the prowiding a pocential conflicting stimulus to ccommodation and dark for 3 minutes before measurement to allow any accommoda reducing the relative brightness of the primary peak.However,it tive adaptatinn effee to subside.7The dark foeus was bele the should be poted that the dark red color was only ever paired with accommodation rospons for all the particpants at most times.For blue (i.e.,in the dark red on blue condition)and that the second example,the mean difference berween the dark focus and the re- ary mude of the dark rod color is cls to that of the spectral peak sponse to the black-on-white target varied from0D to 2.1 D for for the blue (Fig.3).A similar argument may be made for the our participants.Thus,it is unlikely that the resting ofaccom- pairing of dark blue with red (Fig.3).Thus,the effecrs of secondary modation played a major role in measurements. modes in dark red and dark blue targets are likely to have been Previous studies hve reported wide inerindividua differencesn small the rspon to olor targstherfore,it smed unlikdy that the Using the Chromatic eye model of Thibos ex al.,"the LCA homogeneiry of variances asumpion of common prmtric statisti- between the blue and red culors was 1.05 D,and the LCA berween cal trsts suchas analsisafvariance(ANOVA)wnuld he mrt.Arhses dark blue and dark red colors was 0.99 D.Estimated retinal illu ofgroup mean effs in this sitsrion alscabe misleading"making minances for the white,red,and blue backgrounds were 3.45.3.53 the singe-ce foe singl-subjecr)design a necesiry.Common pr and 3.19 log tolands respecrively,and 2.79 and 3.00 log molands for the dark red and dark blue target components.respectively. metric tesrs (eg.ANOVA)camnnot he used in single-cauve designs Weber luminance contrasts (AL/L)between croes and background becase seril cometion between redings on the same participant viclaes the indeperdence ssampcions of those tests Randomi were 0.97 for the black on white,black on red,and black on blue tangess:0.80 for the dark red on-blue targee:0.71 for the darkc tion tots for sing-cse designs are vaid altematives in this cas blue-on-red rarget:-0.079 for the red-oo-blue targer and 0.074 Appendisandrused in the pecsent sudy. for the blue-on-ted target.Thus.the blue-on-red and red-on-blue The geometrical tes (GEO)is a randomiration test (seAppen- targets were close to pominal isoluminance.A negative oontrast dix)and was usad as a nonprametrie alternative to ANOVA.Ies vaue for the red on blue tage indicates that the cross rarget was test sraristic is the averape difference in the dependent variable berween all the postble pairs of creatment conditioes.Omnibus brighter than the background in this condition. The black-on-white tange was the standard target.A highly tests were performed with 5.04 X 10 random enumerations of che accurate and stable respoese would be expected with this target data Nonparametric confidence intervals [C.I's)on pairwise com because of its wide spectra bandwidthand high lumine parisons were then determined with the GEO test using an itera- contrast!The black on blue and black on-red targets provided tive procdures with cither 5000 or 15,000 random enumera- high luminance contrast but had reduced spectral bandwidh.In tions.CI's could not be calculated in some cases because ofmissing these cooditions.the response may be less accurate and more vari- valuex The nonparametricClis gencrally not symmetrie shaut the ahle in some individuls becase of the reduced spectral band- mean but may be described by the CI lower limit (C)and CI width.a.2 Hawever,there wrre no conflicing stimuhs levels upper limit (C.When it is necesary co describe CI's for a suhset caused by LCA,and it wis expecred that the respoese woald be of study participants,these will be given in the text and in Table greater foe the black-on-red than for the black-on-blue targee (see as the range of individual CI lower limits,followed by the range of Factors in the Respoese to Coloe Stimuli,above). individual Cl upper limits.For esampl,if a subsct of three par- The blue-on-red and red-on-blue targets have small luminance ticipants had Cl's on a particular variahle of 0.1 to 1.8,0.2 to 1.9, contrasts,and it was expected that these wold show deeneased and 3to 2.0.repectively,then this would be repeeted as 0.1 accuracy.The response also may become more variabl,either C≤03nl18≤C11≤20. because the near isoluminant conditions provide poor chromaric To exanaine the efects of target color on short-cerm within-trial cu0or becase the participant is making active changes in response variabiliry,lo(Var)was used as a dependentvariable in the accommodation herween the focus levels for the red and bhe target GEO test,where Var is the measured reponse variance in each components.The dark bluc-on-tod target was expeeted to provide 12.8-s trial.In this way,tests of diferences in In(Var)become exact responses inrermediate berween those of the black-on-red and tests of ratios of within trial variances (RWTVI.CI's foe RWTV's blue-on-red targers,whereas the dark red-on-hlue rarget was ex- were caluated by the itrativ procdure described previously perted to pronide respondes intermrdute hetwern those of the Tost in the er prosntation ofraioof varin data.the following blck-on-blue and rod-on-hlue targets caltices were performed.Foeaofaofvinc veseped Inall the cxperimental runs,accommedtion was ncurdod at 40 moticnts…星d,hctQ,Q}was obindl Ha for 12.8 s.The target was set at a position,depending on cch where nTheg nodirectiono of vinces wis de participane's refractive eroe,corresponding to an accommodative finod as T-op代，wherE RN-mx.Smh中，the smallest srimlus of ahour3D far the hlack-an-white combination (rang nondirectonal ratio ofvarianos wis definedsr-exp(R whereR 285 to 36 D).This stimulus was chooen as rpresentative af min Q For cxample,in the sct of three vahos 1.1.0.2.and 3.5,the arange of near-visiun tasks.Data were collectod in 10 cxpcrimental minimum marioo varrianos (RV)0.2,and the mnimum RVis 3 5,bur Optaueny ad Vuien Seiner.Yal 81,Na.$Septerhet 200

color temperature of the source (2469 K).45The excitation purities of the targets were 0.28 for white, 0.93 for red and blue, 0.60 for dark red, and 0.69 for dark blue. All the colored targets show secondary modes in their spectral distributions (Fig. 3). In the case of the red and blue targets, these modes are small, and the excita￾tion purities remain high (0.93 in both cases). However, the modes are important for the dark red and dark blue targets (Fig. 3), providing a potential conflicting stimulus to accommodation and reducing the relative brightness of the primary peak. However, it should be noted that the dark red color was only ever paired with blue (i.e., in the dark red-on-blue condition) and that the second￾ary mode of the dark red color is close to that of the spectral peak for the blue (Fig. 3). A similar argument may be made for the pairing of dark blue with red (Fig. 3). Thus, the effects of secondary modes in dark red and dark blue targets are likely to have been small. Using the Chromatic eye model of Thibos et al.,9 the LCA between the blue and red colors was 1.05 D, and the LCA between dark blue and dark red colors was 0.99 D. Estimated retinal illu￾minances for the white, red, and blue backgrounds were 3.45, 3.53 and 3.49 log trolands, respectively, and 2.79 and 3.00 log trolands for the dark red and dark blue target components, respectively. Weber luminance contrasts (L/L) between cross and background were 0.97 for the black-on-white, black-on-red, and black-on-blue targets; 0.80 for the dark red-on-blue target; 0.71 for the dark blue-on-red target; 0.079 for the red-on-blue target; and 0.074 for the blue-on-red target. Thus, the blue-on-red and red-on-blue targets were close to nominal isoluminance. A negative contrast value for the red-on-blue target indicates that the cross target was brighter than the background in this condition. The black-on-white target was the standard target. A highly accurate and stable response would be expected with this target because of its wide spectral bandwidth18, 22 and high luminance contrast.1 The black-on-blue and black-on-red targets provided high luminance contrast but had reduced spectral bandwidth. In these conditions, the response may be less accurate and more vari￾able in some individuals because of the reduced spectral band￾width.18, 22 However, there were no conflicting stimulus levels caused by LCA, and it was expected that the response would be greater for the black-on-red than for the black-on-blue target (see Factors in the Response to Color Stimuli, above). The blue-on-red and red-on-blue targets have small luminance contrasts, and it was expected that these would show decreased accuracy. The response also may become more variable, either because the near isoluminant conditions provide poor chromatic cues12, 29, 30 or because the participant is making active changes in accommodation between the focus levels for the red and blue target components. The dark blue-on-red target was expected to provide responses intermediate between those of the black-on-red and blue-on-red targets, whereas the dark red-on-blue target was ex￾pected to provide responses intermediate between those of the black-on-blue and red-on-blue targets. In all the experimental runs, accommodation was recorded at 40 Hz for 12.8 s. The target was set at a position, depending on each participant’s refractive error, corresponding to an accommodative stimulus of about 3 D for the black-on-white combination (range, 2.85 to 3.6 D). This stimulus was chosen as representative of arange of near-vision tasks. Data were collected in 10 experimental blocks of 7 trials each. Within a block, the presentation order for each of the seven target-on-background combinations was random without replacement. Participants were instructed to “look at the target naturally, the same as you would when normally reading a book or sign at the same distance.”15 Dark focus measurements were made with a Canon Autoref R-1 autorefractor (Tokyo, Japan).46 Each participant was left in the dark for 3 minutes before measurement to allow any accommoda￾tive adaptation effects to subside.47 The dark focus was below the accommodation response for all the participants at most times. For example, the mean difference between the dark focus and the re￾sponse to the black-on-white target varied from 0.6 D to 2.1 D for our participants. Thus, it is unlikely that the resting state of accom￾modation played a major role in measurements. Previous studies have reported wide interindividual differences in the responses to color targets;8 therefore, it seemed unlikely that the homogeneity of variances assumption of common parametric statisti￾cal tests such as analysis of variance (ANOVA) would be met. Analyses of group mean effects in this situation also can be misleading,8making the single-case (or single-subject) design a necessity. Common para￾metric tests (e.g., ANOVA) cannot be used in single-case designs because serial correlation between readings on the same participant violates the independence assumptions of those tests.48 Randomiza￾tion tests for single-case designs are valid alternatives in this case (see Appendix)48, 49 and were used in the present study. The geometrical test (GEO) is a randomization test (see Appen￾dix) and was used as a nonparametric alternative to ANOVA.50 Its test statistic is the average difference in the dependent variable between all the possible pairs of treatment conditions. Omnibus tests were performed with 5.04  105 random enumerations of the data. Nonparametric confidence intervals (CI’s) on pairwise com￾parisons were then determined with the GEO test using an itera￾tive procedure51 with either 5000 or 15,000 random enumera￾tions. CI’s could not be calculated in some cases because of missing values. The nonparametric CI is generally not symmetric about the mean but may be described by the CI lower limit (CL) and CI upper limit (CU). When it is necessary to describe CI’s for a subset of study participants, these will be given in the text and in Table 1 as the range of individual CI lower limits, followed by the range of individual CI upper limits. For example, if a subset of three par￾ticipants had CI’s on a particular variable of 0.1 to 1.8, 0.2 to 1.9, and 0.3 to 2.0, respectively, then this would be reported as 0.1 CL 0.3 and 1.8 CU 2.0. To examine the effects of target color on short-term within-trial response variability, ln(Var) was used as a dependent variable in the GEO test, where Var is the measured response variance in each 12.8-s trial. In this way, tests of differences in ln(Var) become exact tests of ratios of within-trial variances (RWTV). CI’s for RWTV’s were calculated by the iterative procedure described previously.51 Toassistinthelaterpresentationofratioofvariancedata,thefollowing calculations were performed. For a set of ratio of variance values expressed as quotients {q1, q2, ... qN}, the set {Q1, Q2, ... QN} was obtained, where Qi 
lnqi
. The largest nondirectional ratio of variances was de￾fined as rN  exp (RN), where RN  max Qi . Similarly, the smallest nondirectionalratioofvarianceswasdefinedasr1exp(R1),whereR1 min Qi . For example, in the set of three values 1.1, 0.2, and 3.5, the minimumratioofvariances(RV)is0.2,andthemaximumRVis3.5,but 702 Accommodation for Stationary Colored Targets—Atchison et al. Optometry and Vision Science, Vol. 81, No. 9, September 2004

Accommodation for Staticnary Colored Targess-Axchison c al.704 TABLE 1. Summary of statistical tests and comparisons Kesult Description Test Patticipants Test Resubts Numbers Omnibus Statistical lests Moan ARh:target color GED 1a AB D,E pu.01 1b P=.021 Short-torm AR vanability:tarpet color GEO 2a B P=01.DO8 2b A C.D,E 0015≤p≤0.71 Long-erm AR variability:target color MBF 3 AB 0.0005￥p￥0.002 3b C,D.E 0.46≤p≤075 Reduced spectral bandwidth:bluc wersus red Mean AR:black on-red less black-on- CEO 4a AB D,E mc1n+0.0D,+0.35≤C≤+0.42D,+0.91 hluc Cu +1.32 D 4b mc1n-0.12D:CL-0.64to+0.34D Shurl-erm AR varianke:black-on-red CIGEO 5a dn035:Cl.0.18lD095:r1-1.l:rx-5.6 ÷black-on-blue 5b A B D,E medn12,0.15sC1s069.1.2sCus58: h1-12:rw-68 Long-erm AR variance for black-on- MBF 6 AB 0.012￥p0.79 red versus black-on-blue Reduced spectral bandwidth conditions wersus black onwhie Mean AR:black on-red less black on CGEO 33 AB mcan+0.15D.+0.00≤C￥+004D,+0.23 white ￥Cu≤+028D 716 C D dn+008D,-0.23三Cg-007D+0.10 C 10.56 D 7c E meany +0.06 D:Cl undelined Mean AR:black-on-blue less black- CIGEO 8a ABD m-0.66D:-0.955C15-074D-0.49 un-while Cu-044 D 8b CE man-0.25D:-0.93sCg-031D.+0.02 ￥Cug+0.47D Short-term AR variance:black-on- CIGEO 9为 ABCD mn1.2:0.18gC1≤071,1.5gC:g5.4 white +black-on-red r1-14:w-55 h F mean 1.1;C undefined Short-term AR variance:black-on- CIGFO 10 A meary 097:0.14CL=0.73,1.2 s Cys 40: white÷lak-A-lhe 11-12:w-70 Long-er AR variarke for black-on- MBF 11 AB 0295pG042 white versus black-on-red Long-emm AR variane for black-un- MBF 12 AB 0115p5016 white versus black-on-blue Multi-color displays versus black-on white Mean AR:med-an-hlue less black-on- 10 13a ABE man-0.55D:-1.D55Cs-D63D-0.21 white 5Cu5-007D 13b CD man-0.06D;-0.53三C1≤-D0D,+0.02 Cv +0.57D Mean AR:blue-on-red less black-n- CIGFO 14a B man-0.77D:Ql-1,28m-0.43D whire 146 AC.D am-0.21D:-0.84GCG-038D,+0.17 C 10.60 D 14c E d汽-0.57D:Cl undelined Shurl-erm AR variare:black-un- CIGEO 15a B 4n042:C013w097:1-1.0w-80 white red-on-blue 15b AC.D,E mo1n1.20.14三C4s0.71,1.Bs Cus5.9g 11=14w=7.0 Short-tem AR variance:black-on-white CKO 161 AB C D mean,D.76;01.06 s CL 0.37,1.66 Cu s hlue-an-red 3.07:1=1.7:f=16 16h mean 1.9:C undefined Lonp-emm AR variance for black-on- MBF 17 AB 00125p50,048 white versus red-on-hluae long-erm AR variance for black-on- MBF AB 0022s050.13 white versus hlue-on-ted (paameng and Vaion Srimm Yal 81.No S.Septemhet 2004

TABLE 1. Summary of statistical tests and comparisons Description Test Resulta Numbers Participants Test Results Omnibus Statistical Tests Mean ARb : target color GEO 1a A, B, D, E p 0.001 1b C p  0.021 Short-term AR variability: target color GEO 2a B p  0.008 2b A, C, D, E 0.015 p 0.71 Long-term AR variability: target color MBF 3a A, B 0.0005 p 0.002 3b C, D, E 0.46 p 0.75 Reduced spectral bandwidth: blue versus red Mean AR: black-on-red less black-on￾blue CIGEO 4a A, B, D, E mean, 0.80 D, 0.35 CL 0.62 D, 0.91 CU 1.32 D 4b C mean, 0.12 D; CI, 0.64 to 0.34 D Short-term AR variance: black-on-red CIGEO 5a C mean, 0.35; CI, 0.18 to 0.95; r1  1.1; rN  5.6 black-on-blue 5b A, B, D, E mean, 1.2, 0.15 CL 0.69, 1.2 CU 5.8; r1  1.2; rN  6.8 Long-term AR variance for black-on￾red versus black-on-blue MBF 6 A, B 0.012 p 0.79 Reduced spectral bandwidth conditions versus black-on-white Mean AR: black-on-red less black-on￾white CIGEO 7a A, B mean, 0.15 D, 0.03 CL 0.04 D, 0.23 CU 0.28 D 7b C, D mean, 0.08 D, 0.23 CL 0.07 D, 0.10 CU 0.56 D 7c E mean, 0.06 D; CI, undefined Mean AR: black-on-blue less black￾on-white CIGEO 8a A, B, D mean, 0.66 D; 0.95 CL 0.74 D, 0.49 CU 0.44 D 8b C, E mean, 0.25 D; 0.93 CL 0.31 D, 0.02 CU 0.47 D Short-term AR variance: black-on￾white black-on-red CIGEO 9a A, B, C, D mean, 1.2; 0.18 CL 0.71, 1.5 CU 5.4; r1  1.4; rN  5.5 9b E mean, 1.1; CI, undefined Short-term AR variance: black-on￾white black-on-blue CIGEO 10 All mean, 0.97; 0.14 CL 0.73, 1.2 CU 4.0; r1  1.2; rN  7.0 Long-term AR variance for black-on￾white versus black-on-red MBF 11 A, B 0.29 p 0.42 Long-term AR variance for black-on￾white versus black-on-blue MBF 12 A, B 0.11 p 0.16 Multi-color displays versus black-on￾white Mean AR: red-on-blue less black-on￾white CIGEO 13a A, B, E mean, 0.55 D; 1.05 CL 0.63 D, 0.21 CU 0.07 D 13b C, D mean, 0.06 D; 0.53 CL 0.30 D, 0.02 CU 0.57 D Mean AR: blue-on-red less black-on- CIGEO 14a B mean, 0.77 D; CI, 1.28 to 0.43 D white 14b A, C, D mean, 0.21 D; 0.84 CL 0.38 D, 0.17 CU 0.60 D 14c E mean, 0.57 D; CI, undefined Short-term AR variance: black-on- CIGEO 15a B mean, 0.42; CI 0.13 to 0.97; r1  1.0; rN  8.0 white red-on-blue 15b A, C, D, E mean, 1.2; 0.14 CL 0.71, 1.8 CU 5.9; r1  1.4; rN  7.0 Short-term AR variance: black-on-white blue-on-red CIGEO 16a A, B, C, D mean, 0.76; 0.06 CL 0.37, 1.66 CU 3.07; r1  1.7; rN  16 16b E mean, 1.9; CI, undefined Long-term AR variance for black-on￾white versus red-on-blue MBF 17 A, B 0.012 p 0.048 Long-term AR variance for black-on￾white versus blue-on-red MBF 18 A, B 0.022 p 0.13 Accommodation for Stationary Colored Targets—Atchison et al. 703 Optometry and Vision Science, Vol. 81, No. 9, September 2004

704 Accommodation for Stationary Colored Targets-Alchison et al. TABLE 1. Summary oi statistical tests and comparisons (continurd) Description Test Result Icst Kesuhts Numbers Participants Multicolor deplays vensus blue and red foci Mean AR:red-on-blue less red focus CGEO 19为 A B,D.E mean-0.63D:-1.19￥C1￥-071D.-0.40 ￥C≤-015D 1为 C anm+0.16D:Cl-0,27m+074D Mean AR:red-n-blue less blue foous OIGFO 204 D man+0.44D:Cl+0.08to+0.80D 20 A B C 4m+0.10D:032sC1g008D,10.34 C 1055D 20 meany-0.03:Cl.undefined Mean AR:blue-on-red less red focus CIGEO 21a ABD medn-0.65D:-1.95sC-060D,-0.52 C,9-008D 21b mn+0.08D:C-0.42o+0.70D 21c mean,-0.62 D:Cl,undefined Mean AR:hlue-on-red less blue focus CFO 22a mam+0,44D:l+0.D5tm+0.0D 72b B,C,D meany +0.03 D:-0.44 CL -D.13 D,+0.2D SCS +067 D 22c meany +0.13 D:Cl undelined Multioolor displays:effect of luminance conlrasl Mean AR:ced backgroundl MCTT 23 A 0.135p078 Mean AR:blue background MCTT 24a C pc0.001 24b AB,上 D039s p 0.48 Short-torm AR varance:red MCII 25 DH1sp三0.98 hackground Short-term AR varance:bluo MCTT 26 D.115psD79 hackgyound Statistically significant effects are denoted by bold rosult numbors. AK,accommodation response;Cl.99%confidence interval for an individual paricipant:CIG,confidence inerval by the geometrical test C.lower bound of the 99%confidence interval:Cu upper bound of the 99%confidence interval:MBF,modified Brown-Forsythe s:MCTT,modified correlation trend test:GEO,geumetrical test smallst nondirectional ratio of variances; largest nondirectional ratio of variances.These statistical tests and parameters are described fully in the Mechods section the lgest nondirectionl RV (i.e.,the one furthest fiom 10)is 5.0 (ie. medium-comas target is midway berween the repective high- IKL2,and the smalleg nondsectional RVfLe.,the ane dosest to 1.0 is contrast and low-contras targets.Probabiliry values were deter- 11. mined using 6X 10 random enumerations. A test wis sought for changes in accommodation response vari Finally.to compensate for family wise emor raes cased by single abiliry over a longer period comeponding to the time berween case analyses on the five separare participants,the Bonferroniinequal trials.The previous test (see abowe)had only considered accomno- ity was imoked to mduce the significanoe levd from 5%to 1 for all dative vartabeliry for the brief 12.8-&duration of an individual trial. the tests.Accordingly,9%CI's were usd theoughout. Unforunatey,there are no exact tests amenable to our data therefore.a modified form of the Brown Forsythe (BF)test was used.The test sistic was calculated by randomization proce- dure (GEO test)to free it from the independence assumption. RESULTS which is usually untenhle in single-case designs Omnibas tests Frequency distributions of each of the accommodative re. were performed with 5.04 X 10'random enumerations and pair sponse runs for cach target-on-background cambination are wise comparisons wich eitber 5000 or 15.000 random enumera shown for participants A.B.C.D.and E in Figs.4.5,and 6. tions This test is not amenable to the calcultion of CI's. respectivcly.Group average accommodation responses for all To esamine the efTects of luminance contrast.the Edgington the target-an-hackground comhinations are plotted in Fig.7. correlatinn trend test (CTT)was mndified for repeated measures To abeain this figure,each participant's response for the black- and to allow a nondirectional mll hypothesis Far a given hack- on-whire target was subtracted from the mean response for the ground color (red or blue)the three matching tangetc(iother conditions.Consequcntly,the group mean and SD for blck,dark blue,and hlue foe the ted background:black.dark red. the black-on-white combination are both xero.Intratrial ac- and red for the blue hackground)were given dummy codings to commodation response variahiliry in the various target condi- rest for deparurs from the mull hypochesis that the meponse to the tions is ploeted in Fig.8

the largest nondirectional RV (i.e., the one furthest from 1.0) is 5.0 (i.e., 1/0.2), and the smallest nondirectional RV (i.e., the one closest to 1.0) is 1.1. A test was sought for changes in accommodation response vari￾ability over a longer period corresponding to the time between trials. The previous test (see above) had only considered accommo￾dative variability for the brief 12.8-s duration of an individual trial. Unfortunately, there are no exact tests amenable to our data;52 therefore, a modified form of the Brown-Forsythe (BF) test53 was used. The test statistic was calculated by a randomization proce￾dure (GEO test) to free it from the independence assumption, which is usually untenable in single-case designs. Omnibus tests were performed with 5.04  105 random enumerations and pair￾wise comparisons with either 5000 or 15,000 random enumera￾tions. This test is not amenable to the calculation of CI’s. To examine the effects of luminance contrast, the Edgington correlation trend test (CTT)49 was modified for repeated measures and to allow a nondirectional null hypothesis. For a given back￾ground color (red or blue) the three matching target colors (i.e., black, dark blue, and blue for the red background; black, dark red, and red for the blue background) were given dummy codings to test for departures from the null hypothesis that the response to the medium-contrast target is midway between the respective high￾contrast and low-contrast targets. Probability values were deter￾mined using 6  104 random enumerations. Finally, to compensate for family-wise error rates caused by single￾case analyses on the five separate participants, the Bonferroni inequal￾ity was invoked to reduce the significance level from 5% to 1% for all the tests. Accordingly, 99% CI’s were used throughout. RESULTS Frequency distributions of each of the accommodative re￾sponse runs for each target-on-background combination are shown for participants A, B, C, D, and E in Figs. 4, 5, and 6, respectively. Group average accommodation responses for all the target-on-background combinations are plotted in Fig. 7. To obtain this figure, each participant’s response for the black￾on-white target was subtracted from the mean response for the other conditions. Consequently, the group mean and SD for the black-on-white combination are both zero. Intratrial ac￾commodation response variability in the various target condi￾tions is plotted in Fig. 8. TABLE 1. Summary of statistical tests and comparisons (continued) Description Test Resulta Numbers Participants Test Results Multicolor displays versus blue and red foci Mean AR: red-on-blue less red focus CIGEO 19a A, B, D, E mean, 0.63 D; 1.19 CL 0.71 D, 0.40 CU 0.15 D 19b C mean, 0.16 D; CI, 0.27 to 0.74 D Mean AR: red-on-blue less blue focus CIGEO 20a D mean, 0.44 D; CI, 0.08 to 0.80 D 20b A, B, C mean, 0.10 D; 0.32 CL 0.08 D, 0.34 CU 0.55 D 20c E mean, 0.03; CI, undefined Mean AR: blue-on-red less red focus CIGEO 21a A, B, D mean, 0.65 D; 1.95 CL 0.60 D, 0.52 CU 0.08 D 21b C mean, 0.08 D; CI, 0.42 to 0.70 D 21c E mean, 0.62 D; CI, undefined Mean AR: blue-on-red less blue focus CIGEO 22a A mean, 0.44 D; CI, 0.05 to 0.80 D 22b B, C, D mean, 0.03 D; 0.44 CL 0.13 D, 0.20 CU 0.67 D 22c E mean, 0.13 D; CI, undefined Multicolor displays: effect of luminance contrast Mean AR: red background MCTT 23 All 0.13 p 0.78 Mean AR: blue background MCTT 24a C p 0.001 24b A, B, D, E 0.039 p 0.48 Short-term AR variance: red background MCTT 25 All 0.081 p 0.98 Short-term AR variance: blue background MCTT 26 All 0.11 p 0.79 a Statistically significant effects are denoted by bold result numbers. b AR, accommodation response; CI, 99% confidence interval for an individual participant; CIGEO, confidence interval by the geometrical test; CL, lower bound of the 99% confidence interval; CU, upper bound of the 99% confidence interval; MBF, modified Brown-Forsythe test; MCTT, modified correlation trend test; GEO, geometrical test; r1, smallest nondirectional ratio of variances; rN, largest nondirectional ratio of variances. These statistical tests and parameters are described fully in the Methods section. 704 Accommodation for Stationary Colored Targets—Atchison et al. Optometry and Vision Science, Vol. 81, No. 9, September 2004

Accommodation for Staticnary Colared TargetsAxchison c al.705 七c政aag 04 D.8 03 金鞋 04 04 婴AM Urh同 08 Dak Rrd e想作s =博投s Dark Bie 0,8 EaH国 等 4 4 d 08 08 ad 04 en 0. 4 B Accommodation Response (D) FIGURE 4. IA)Hiskgtants of dynamic utccentodtiont tpurras fur participt A ke the suvun larut-on-ckgpuurd conditions:black on whoe,black on blac. dark red on hlue.cark hlue on red,hlack on red,red on hlue,and blue on red.The ordinate gives the proporton of time spent ocusing at the particular dn nspute levd The hiob widl is 0.I D.Eath cuntinaus line repnsets dala from ant invidaal tril and cuvens accommodative response ranpe for thot tal ie..the hnes ane not shown falling to zero on each sidek.Datted lines indicane the appropriate responses fur the cfffrenl cakr.The datk fucus for tho partiLipart is 1.7 D.(B)Hisogrants uf cymamk accuermnlalicea npurnes fut partitipant B for the seven targeton-bockground condtore.The dark iocus for hs pamcpansD.ther detals are as ior Fig 4A capton. Omnibus Statistical Tests presented)were only pertormed for the two individuals whose There was a highly significant overall effect of coloe condition omnibus tesrs were significant (ie parieipants A and B). on the mean accommodatinn rexpanse for four of five participants (Table 1,reslt la)hut mot for parricipant C(Fig SA and Table 1. roult 1b).By contrast,target olor had a significant cffoct on Reduced Spectral Bandwidth:Blue vs.Red short-term accommodation respons variability in only one of five As prodicted from the LCA of the eye,the response to the participanes (Fig4B,participant B and Table 1.resules 2a and 2b). black-on-red tange was greater than o the blckc-on-blue tanger in Target calar had a significant effeet on long-term respoese vari- moot participants The mean dfference in rsponse for those four ahility in enly rwo of five participnts (Table 1,resnlts 3a and 3b). participants (A.B,D,and E)who showed a significant effect was Accordingly,pairwise cumparicons on long-tomm variability (to be 0.80D(Table 1,result 4a),a litlelos than the cudifference in Optaueny ad Vuien Seiner.Yal 81,Na.$Septerhet 200

Omnibus Statistical Tests There was a highly significant overall effect of color condition on the mean accommodation response for four of five participants (Table 1, result 1a) but not for participant C (Fig. 5A and Table 1, result 1b). By contrast, target color had a significant effect on short-term accommodation response variability in only one of five participants (Fig. 4B, participant B and Table 1, results 2a and 2b). Target color had a significant effect on long-term response vari￾ability in only two of five participants (Table 1, results 3a and 3b). Accordingly, pairwise comparisons on long-term variability (to be presented) were only performed for the two individuals whose omnibus tests were significant (i.e., participants A and B). Reduced Spectral Bandwidth: Blue vs. Red As predicted from the LCA of the eye, the response to the black-on-red target was greater than to the black-on-blue target in most participants. The mean difference in response for those four participants (A, B, D, and E) who showed a significant effect was 0.80 D (Table 1, result 4a), a little less than the actual difference in FIGURE 4. (A) Histograms of dynamic accommodation responses for participant A for the seven target-on-background conditions: black on white, black on blue, dark red on blue, dark blue on red, black on red, red on blue, and blue on red. The ordinate gives the proportion of time spent focusing at the particular accommodation response level. The histogram bin width is 0.1 D. Each continuous line represents data from an individual trial and covers the accommodative response range for that trial (i.e., the lines are not shown falling to zero on each side). Dotted lines indicate the appropriate responses for the different colors. The dark focus for this participant is 1.7 D. (B) Histograms of dynamic accommodation responses for participant B for the seven target-on-background conditions. The dark focus for this participant is 0.6 D. Other details are as for Fig. 4A caption. Accommodation for Stationary Colored Targets—Atchison et al. 705 Optometry and Vision Science, Vol. 81, No. 9, September 2004

70h Accommodation for Stationary Colorod largets-Atchison et al. 08 lllack on Whhe Bkg件e 04 0.4 0.8 Hack on Bse 08 0.8 DOR Dark Red 期 oa Blwe 04 0 0.8 rl买 0.8 IrL Kh 0.4 0.4 0 08 0.8 Bbckon民 04 08 R时 08 04 04 0.8 Bee en Rod 08 Hee on Red 0,4 04 0 2 B Accommodation Response (D) FIGURE 5. (A)Histograms af dynamic accommodation esporses for participant C for the seven tanetonbackpound conditions.The dark focus for this parScipunt is 0.7 D.Other dutaib are n ke Fg.4A capfcn.(B)Hisogyan of dymarit commodatiun nopumes for prliipan!D fur the weven tarzet-on-bockground conditiore.The dark iocus for this particlpant is 1.4 D.Other details are as for Hig.4A caption. ocular refraction (caused by LCA)of1.05 D.Participant Cdid not pants'responses were significantly more varable when comparing shom a significant differential mexponse to ICA(Fig SA and Tahle the hlack-on-rod and hlack-on-hlue coelirions (Table 1,result 6) 1,result 4b). In most cases,the accommodation respunse was not signifi- cantly more variable in the short term when comparing the two Reduced Spectral Bandwidth Conditions vs.Black- on-White reduced spectral bandwidth conditions with cach ocher (partic ipants A.B,D,and E:Fig 8 and Table 1,result 5h).Participant The reponse to the bhck-on-ted target was slightly higher than C had a less variable respans when vicwing the black-on-rod v to the gandard black-on-white target for twooffive paricipants(A black-on-blue target (Fig.5A:RWTV,0.35;Table 1,roult 5a). and Bs Figs.4and 7;mean diffierence,+0.15 D:Table 1,result 7a) When considering kng-term variability,none of the partici- but noc for the other three participants (C.D.and E:'Table 1. Oaameng and Vaien Sinmn Val 81.No S.Sepeemhe 2004

ocular refraction (caused by LCA) of 1.05 D. Participant C did not show a significant differential response to LCA (Fig. 5A and Table 1, result 4b). In most cases, the accommodation response was not signifi￾cantly more variable in the short term when comparing the two reduced spectral bandwidth conditions with each other (partic￾ipants A, B, D, and E; Fig. 8 and Table 1, result 5b). Participant C had a less variable response when viewing the black-on-red vs. black-on-blue target (Fig. 5A; RWTV, 0.35; Table 1, result 5a). When considering long-term variability, none of the partici￾pants’ responses were significantly more variable when comparing the black-on-red and black-on-blue conditions (Table 1, result 6). Reduced Spectral Bandwidth Conditions vs. Black￾on-White The response to the black-on-red target was slightly higher than to the standard black-on-white target for two of five participants (A and B; Figs. 4 and 7; mean difference, 0.15 D; Table 1, result 7a) but not for the other three participants (C, D, and E; Table 1, FIGURE 5. (A) Histograms of dynamic accommodation responses for participant C for the seven target-on-background conditions. The dark focus for this participant is 0.7 D. Other details are as for Fig. 4A caption. (B) Histograms of dynamic accommodation responses for participant D for the seven target-on-background conditions. The dark focus for this participant is 1.4 D. Other details are as for Fig. 4A caption. 706 Accommodation for Stationary Colored Targets—Atchison et al. Optometry and Vision Science, Vol. 81, No. 9, September 2004

Accommodation for Staticnary Colared Targets-Achisan al.707 (1 Blad cn Whitc same was te for long-term variability (participants A and B. Table 1,resules 11 and 12). 04 Multicolor Displays vs.Black-on-White 0 Respoeses to the red-on-blue and blue-on-red targets were first 08 利kgn后树 compared with the standard black-on-white target.The respoese tothe rd-on-blue targst was significantlyower than to the black- 04 on-white tanget for threc participants (A.B.and E:Fig.4 and6 mean,-0.55 D:Table 1,resalt Isa)hut not for the ocher rwo participunts (C and D:Fig,5 and Table 1.result 13h).In contrast, the rrsponse to the bhe-on-md target w nat sigificantly differ- 08 Durk Hed en Blae ent from that for the black-on-white tanget in four of five partici- pants (A.C.D,and E;Figs.4A.5,and 6 and Tabk 1,roults 146 04 and 14c).In the other participant (B:Fig.4B).the response to the 0 8 Dark Blue 8月 0.4 0.8 04 R Red an 125 0.7541542 11 04 A Din.in response from Bk/w target (D) FIGURE 7. Group accommodasve response drferences tor target.on-background cumbiratium.The ore lun incicale 1 5D of the meart tiBenerces.Each participant's mean res军onse is relative to his or her mean res军enea中e b以.kk-ct-whea, 0.4 Accommodation Response (D) FIGURE 6. Histodram ol dynamic accommodation responses lor paricipant E kor the seven tarpet-nn-hackzround conditioes.The dark focus for this participont is 0.5 D.Other detaib are as lor Fia.4A caption. results 7b and 7c).Ilowever,the response to the bladk-oo-blue targe was much lower than to the bladc-oo-white target for three of five participants(代，B.andD:g4,5B,nd7 ean differ- 0.I n ence,-0.66 D:Table 1,resule 8a)but not for two puarticipunts [C and E:Fig 5A and 6 and Table 1,reoult 8h). Cimup mean standard deviarinn (D) In the shoet term.none of the participants had significanty FIGURE 8. more variable responses in the reducedec handwidth condi- Gmup mean intratrial S's of the arenmmmodative nmponse rins icr exch tioes [Fig.8;black-on-white v&hlack-on-red,Table I.results 9a target-on-backgrcund combination.The emor bors indicate 1 SD of the and Sh;hlack-on-whire vs back-on-hlue,Tahle 1.mauk 10).The mrin SO's

results 7b and 7c). However, the response to the black-on-blue target was much lower than to the black-on-white target for three of five participants (A, B, and D; Figs. 4, 5B, and 7; mean differ￾ence, 0.66 D; Table 1, result 8a) but not for two participants (C and E; Figs. 5A and 6 and Table 1, result 8b). In the short term, none of the participants had significantly more variable responses in the reduced spectral bandwidth condi￾tions (Fig. 8; black-on-white vs. black-on-red, Table 1, results 9a and 9b; black-on-white vs. black-on-blue, Table 1, result 10). The same was true for long-term variability (participants A and B; Table 1, results 11 and 12). Multicolor Displays vs. Black-on-White Responses to the red-on-blue and blue-on-red targets were first compared with the standard black-on-white target. The response to the red-on-blue target was significantly lower than to the black￾on-white target for three participants (A, B, and E; Figs. 4 and 6; mean, 0.55 D; Table 1, result 13a) but not for the other two participants (C and D; Fig. 5 and Table 1, result 13b). In contrast, the response to the blue-on-red target was not significantly differ￾ent from that for the black-on-white target in four of five partici￾pants (A, C, D, and E; Figs. 4A, 5, and 6 and Table 1, results 14b and 14c). In the other participant (B; Fig. 4B), the response to the FIGURE 6. Histograms of dynamic accommodation responses for participant E for the seven target-on-background conditions. The dark focus for this participant is 0.5 D. Other details are as for Fig. 4A caption. FIGURE 7. Group accommodative response differences for target-on-background combinations. The error bars indicate 1 SD of the mean differences. Each participant’s mean response is relative to his or her mean response for the black-on-white combination. FIGURE 8. Group mean intratrial SD’s of the accommodative response runs for each target-on-background combination. The error bars indicate 1 SD of the mean SD’s. Accommodation for Stationary Colored Targets—Atchison et al. 707 Optometry and Vision Science, Vol. 81, No. 9, September 2004

70 Accommodation for Stationary Colorod TargetsAtchison ct al blue-on-nd target was significantly lower (mcan,-0.77:Table 1,DISCUSSION roolt 14a), With ooc exceprion,the responses to the near-isoluminant tar- Overall,there was noconistent support for or hypathesis that gets were not significantly more variable than to the standard multicnlor targets lead to inercasod accommodtion reponse vari- black on white target in the short term (Fg,8:black on-white vs abiliry.When compared with the standard blck-on-white target. red on blue.participants A.C.D,and E,Table 1.result 15b. rexponses to targets with reduced spectral bandwidth (black on red blck-on-white v.bluc-on-rod.participants A.B,C.D,and E and blck en blue)were not mare vartble in either short-ar Tabk 1,results 16a and 16b).The exoeption was that participant long term viewing (mean RWTV,0.97 to 1.2:Table 1,results %a B had a less variable response to the black-on-white target than to to 10).I lowever.CI's for ratios of variances cannot rule our mod. the red on-blue targe:(Fig.4B:RWTV,0.42:Table 1.result 15a). erate increaes or decreases in response variabiliry in some individ. In the long,term,none of the participants showed moce variable uals (r7.0.Table 1.results 9a to 10).These results apparendy roponsis to the near-isoluminant targets (participants A and B: contradict earier sudies that found poorer accommodative sil Tabk 1,results 17 and 18). iry in reduced spectral bandwidth canditions.How. ewer,accummodative gain falls off gradually with spoctral band- width:therefore,the bandwidths in the present study may Multicolor Displays vs.Blue and Red Foci have been adequate for operation of the normal chromatic stimu Respons to the ncar-isoluminant multicolor targets (rod on ls to accommodation.Another possihiliry is that participants blue and blue on red)were gererally doeer to the blue response were roorting blrand ocher csdrive the respos in (black on-blue target)than to the red response (black on-red the proence of a degraded chromatic stimulus.In any case,the cargec). present study supports other findings that accommodation s rea- For the red-on-blue target,the response was significanthy lower obsttowdtand that rypically than the red focus for four of five participanes (A.B.D.and E:Figs enouuntered situratod oolors are not tou dabilitaring to the control 4.5B,and 6:mean,-0.63 D:'Table 1.result 19a)hut not for mechanisms of accommodation.s participant C(Fig 5A and Table 1.result 19).Comversely.the With one exception (participant B.see Fig.4B),responses to red on-blue respoose was noc significantly different from the blue the near-isoluminant tangets(red on blue and blue an red)were focus for four participants (Fig 4.5A.and G and Table 1.results not more variable than to the standard black on white target 20h and 20c).The exception was participant D.whose red-on- (mean RWTV,0.76 to 1.9:Table 1,results 15b to 16b).Ilow. blue response was greater than the blue focus (mean,+0.4 D:Fig. ever,in these participants our Cl's cannot rule out moderate to 5B and Table 1,result 20a). large increases or decreases in variability near isoluminance for Similarly.for the blue-on-red target.the response was signifi some individuals (r 16:Table I.results 15b to 16b)at least cantly lower than the red focus foe three of five participants (A.B. for short-term (12.Ss)viewing.In the present study,the targets and D:Figs.4 and SB:mean.-0).65 D:Table 1.resule 21a)but were not cntirely isoluminant.Colors were not subjectively pot for two participants (Figs.5A and 6 and Table 1.results 21b brightness matched (although they were matched for retinal and 21c).Conversely.the blus on-red response was not signifi illuminance),and no attempt was made to correct for the par cantly different from the hlue focus for four participants (B.C,and tikcipants'transvene chromatic aherration or lngitudinal chro- D.Figs.4B and 5 and Table 1,resule 2Zh:E.Fig.G and Table 1. matic aberration.Thus,there wire likely luminance and chro- resut 22c).The exception was participant A (Fig.4A).whose matic artifacts in the retinal images that could be used to drive ble-on-red response was greater than the blue fooss (mean. accommodatian.Despite evidence that isoluminnt targets +0.i D:Table 1.result 22a). prowvide a poor stimulus to accommedation,the present sudy sgssts that in conditions outside the laboratory (whene Multicolor Displays:Effect of Contrast arrempis are not made to maintain ioluminance),these targets would lead to no mare than moderate increases in response It was hypothesized that the ropons to the medism-contrast variabality in many individuals Whether it is advisahle to view cages for particular background color would be intermediate such degraded stimuli routinely is an issue not addressed in the berween the respective high-contrast and near-isoluminant target present srudy.Other srudies using natural viewing conditions reponses.When considering the mean response leve,and with have also found teasonably accurate responses to subjectively one excepeion,this mll hypethess could nat he rejecod in any of brightness-matched multicolor displaysand to displays the participants(Fig 7;rod hackgromnd,all particpints;Table 1. in which there was no brightness matching. result 25:blue backgound.paricipants A,B.D.and E:Table 1. In four of five participans.there was a differential response result 24b).The escepcion was parricipant C with a blue back- to black-on-red and black-on-blue targets that cosdy matched ground (Fig 5A).In this ese,the respoese a the daek red-on-blue the LCA of the eye (response of 0.80 D vs.LCA of 1.05 D). targr (2.30 D)was significantly higher (Tahle 1,result 24a)than However,when presented with near isoluminant red on-blue to the other rao cunditions (black on bluc,1.70 D;rod on bluc. and blue-on-rod tangets,four of frve participants did not adopt I.751D). more variable roponses and so were not focusing bttween the When considering the short-term vartnce of the accommoda- blue and red focus levels as we originally hypothesized.We tion resporse,the null hypocheis (of no effeet of luminance com- cannne rule out that participant B was adopting such a respanse trast)could not be md in any cae (Fig,8 and Tahle 1,results pattern.Alternatively,the near-isnluminant conditions may 253nd261. have been detrimental to this individual's accummodative Optaueny ad Vuien Seiner.Yal 81,Na.$Septerhet 200

blue-on-red target was significantly lower (mean, 0.77; Table 1, result 14a). With one exception, the responses to the near-isoluminant tar￾gets were not significantly more variable than to the standard black-on-white target in the short term (Fig. 8; black-on-white vs. red-on-blue, participants A, C, D, and E, Table 1, result 15b; black-on-white vs. blue-on-red, participants A, B, C, D, and E, Table 1, results 16a and 16b). The exception was that participant B had a less variable response to the black-on-white target than to the red-on-blue target (Fig. 4B; RWTV, 0.42; Table 1, result 15a). In the long term, none of the participants showed more variable responses to the near-isoluminant targets (participants A and B; Table 1, results 17 and 18). Multicolor Displays vs. Blue and Red Foci Responses to the near-isoluminant multicolor targets (red on blue and blue on red) were generally closer to the blue response (black-on-blue target) than to the red response (black-on-red target). For the red-on-blue target, the response was significantly lower than the red focus for four of five participants (A, B, D, and E; Figs. 4, 5B, and 6; mean, 0.63 D; Table 1, result 19a) but not for participant C (Fig. 5A and Table 1, result 19b). Conversely, the red-on-blue response was not significantly different from the blue focus for four participants (Figs. 4, 5A, and 6 and Table 1, results 20b and 20c). The exception was participant D, whose red-on￾blue response was greater than the blue focus (mean, 0.44 D; Fig. 5B and Table 1, result 20a). Similarly, for the blue-on-red target, the response was signifi￾cantly lower than the red focus for three of five participants (A, B, and D; Figs. 4 and 5B; mean, 0.65 D; Table 1, result 21a) but not for two participants (Figs. 5A and 6 and Table 1, results 21b and 21c). Conversely, the blue-on-red response was not signifi￾cantly different from the blue focus for four participants (B, C, and D, Figs. 4B and 5 and Table 1, result 22b; E, Fig. 6 and Table 1, result 22c). The exception was participant A (Fig. 4A), whose blue-on-red response was greater than the blue focus (mean, 0.44 D; Table 1, result 22a). Multicolor Displays: Effect of Contrast It was hypothesized that the response to the medium-contrast targets for a particular background color would be intermediate between the respective high-contrast and near-isoluminant target responses. When considering the mean response level, and with one exception, this null hypothesis could not be rejected in any of the participants (Fig. 7; red background, all participants; Table 1, result 23; blue background, participants A, B, D, and E; Table 1, result 24b). The exception was participant C with a blue back￾ground (Fig. 5A). In this case, the response to the dark red-on-blue target (2.30 D) was significantly higher (Table 1, result 24a) than to the other two conditions (black on blue, 1.70 D; red on blue, 1.75 D). When considering the short-term variance of the accommoda￾tion response, the null hypothesis (of no effect of luminance con￾trast) could not be rejected in any case (Fig. 8 and Table 1, results 25 and 26). DISCUSSION Overall, there was no consistent support for our hypothesis that multicolor targets lead to increased accommodation response vari￾ability. When compared with the standard black-on-white target, responses to targets with reduced spectral bandwidth (black on red and black on blue) were not more variable in either short- or long-term viewing (mean RWTV, 0.97 to 1.2; Table 1, results 9a to 10). However, CI’s for ratios of variances cannot rule out mod￾erate increases or decreases in response variability in some individ￾uals (r 7.0; Table 1, results 9a to 10). These results apparently contradict earlier studies that found poorer accommodative stabil￾ity in reduced spectral bandwidth conditions.12, 18–20, 22 How￾ever, accommodative gain falls off gradually with spectral band￾width;18, 22 therefore, the bandwidths in the present study may have been adequate for operation of the normal chromatic stimu￾lus to accommodation.12 Another possibility is that participants were resorting to blur11 and other cues14 to drive the response in the presence of a degraded chromatic stimulus. In any case, the present study supports other findings that accommodation is rea￾sonably robust to spectral bandwidth18, 22, 27, 41 and that typically encountered saturated colors are not too debilitating to the control mechanisms of accommodation.18 With one exception (participant B, see Fig. 4B), responses to the near-isoluminant targets (red on blue and blue on red) were not more variable than to the standard black-on-white target (mean RWTV, 0.76 to 1.9; Table 1, results 15b to 16b). How￾ever, in these participants our CI’s cannot rule out moderate to large increases or decreases in variability near isoluminance for some individuals (r 16; Table 1, results 15b to 16b) at least for short-term (12.8 s) viewing. In the present study, the targets were not entirely isoluminant. Colors were not subjectively brightness matched (although they were matched for retinal illuminance), and no attempt was made to correct for the par￾ticipants’ transverse chromatic aberration or longitudinal chro￾matic aberration. Thus, there were likely luminance and chro￾matic artifacts in the retinal images that could be used to drive accommodation. Despite evidence that isoluminant targets provide a poor stimulus to accommodation,29, 30 the present study suggests that in conditions outside the laboratory (where attempts are not made to maintain isoluminance), these targets would lead to no more than moderate increases in response variability in many individuals. Whether it is advisable to view such degraded stimuli routinely is an issue not addressed in the present study. Other studies using natural viewing conditions have also found reasonably accurate responses to subjectively brightness-matched multicolor displays8, 31, 32 and to displays in which there was no brightness matching.41 In four of five participants, there was a differential response to black-on-red and black-on-blue targets that closely matched the LCA of the eye (response of 0.80 D vs. LCA of 1.05 D). However, when presented with near-isoluminant red-on-blue and blue-on-red targets, four of five participants did not adopt more variable responses and so were not focusing between the blue and red focus levels as we originally hypothesized. We cannot rule out that participant B was adopting such a response pattern. Alternatively, the near-isoluminant conditions may have been detrimental to this individual’s accommodative 708 Accommodation for Stationary Colored Targets—Atchison et al. Optometry and Vision Science, Vol. 81, No. 9, September 2004