glaucoma silent substitution stimulation of s-cone pathway ... · silent substitution stimulation...

Silent Substitution Stimulation of S-cone Pathway andL- and M-cone Pathway in Glaucoma

Patrick Bessler,1 Sascha Klee,1 Ulrich Kellner,2 and Jens Haueisen1,3

PURPOSE. The study aimed for objective detection of primaryopen-angle glaucoma with selective color channel stimulationbased on the silent substitution technique. In addition, anobjective was analysis of the characteristics of individual colorchannels based on visual evoked potentials.

METHODS. Visual evoked potentials were recorded in 141 sub-jects (39 patients with glaucoma, 71 healthy subjects, and 31age-matched healthy subjects) with two scalp electrodes aftercone-specific flash stimulation. Silent substitution stimulationwas presented with a 30-in. liquid crystal display. Separateresponses were obtained for short-, medium-, and long-wave-length–sensitive cones. Age-matched subgroups were used tocompare patients with glaucoma and healthy subjects.

RESULTS. The S-cone responses of age-matched healthy subjectshad significantly different slopes for the first positive wave,compared with the responses of patients with moderate glau-coma. This difference was not observed in the L- and M-coneresponses. Distinct changes in the S-cone response profileswere observed with increasing severity of glaucoma. Patientswith severe glaucomatous damage were recognizable by thealtered profiles of their visual evoked potentials. Healthy sub-jects showed significant differences between color channels.

CONCLUSIONS. Glaucoma and its severity were objectively de-tected by using the silent substitution technique. The stimula-tion technique and signal analysis enabled assessment of thevisual evoked potentials of individual color channels. (InvestOphthalmol Vis Sci. 2010;51:319–326) DOI:10.1167/iovs.09-3467

Primary open-angle glaucoma is a common disease through-out the world and is one of the most common causes of

blindness.1 Early detection is essential in fighting its progres-sion.2

Routine screening tests detect only a portion of glaucomacases. Furthermore, these tests have limited sensitivities in theearly stages of the disease.3,4 In recent years, significant ad-vances have been made in imaging techniques. Their suitabilityfor the early diagnosis of glaucoma and the effect on thetechniques of influencing factors (e.g., cornea and anatomicvariability) are currently being assessed. Imaging techniques

such as retinal tomography and cornea pachymetry remaininadequate as single diagnostic tools.5,6 Standard diagnosticmethods such as subjective perimetry are dependent on thepatient’s cooperation. Aside from their subjectivity, they havethe disadvantage that significant retinal damage must bepresent for deficits in visual function to be diagnosed.4,7 Inten-sive research has been performed on developing objectivevision diagnosis techniques.8–11 For example, the pattern elec-troretinogram (ERG) is a sensitive test for diagnosing glau-coma.12 Another objective examination method is the visualevoked potential (VEP), whereby vision deficiencies are objec-tively evaluated by analyzing the electrodiagnostic responses ofthe visual cortex.11,13–15

Several studies have demonstrated that glaucoma is associ-ated with disturbances in blue color perception.7,8,13,16 Thephotoreceptors (L-, M-, and S-cones) differ in some certainproperties. In addition, differences are present in the photo-transduction cascade.17–20 With glaucoma, the damage to in-traocular neurons causes changes, even cell death, in higherlevels of the visual pathway. Thus, changes in VEP may be ofclinical use in diagnosing glaucoma.21,22 The resulting specificvulnerability to various clinical pictures could be useful indiagnosis if appropriate stimulation technology is ap-plied.15,23–25

For the specific stimulation of individual photoreceptorcells, tests have been performed to determine the characteris-tics of the color channels, functioning of the retina, and colorperception.26–31

In the present study, we sought to objectively analyze thecortical response signals after selective color channel stimula-tion, based on the silent substitution technique (SST).29,32 Bythis method, the characteristics of the color channels can bedetermined and parameters can be established. We applied ouranalysis to the VEP of healthy subjects and patients with glau-coma at different stages of the disease.

MATERIALS AND METHODS

Subjects

We studied 141 subjects (39 patients with glaucoma and 102 healthysubjects) at the AugenZentrum Siegburg and the Institute of Biomedi-cal Engineering and Informatics, Ilmenau University of Technology(Fig. 1). All the patients were selected from AugenZentrum Siegburg.All subjects gave informed consent after the details and purpose of thestudy had been explained. The study was performed according to thetenets of the Declaration of Helsinki.

The group of healthy subjects consisted of 102 medically diagnosedvolunteers without relevant diseases or defects in color perception.Optical deficits were corrected with existent visual aids. The subjectshad visual acuities between 0.8 and 1.0, intraocular pressures less than21 mm Hg, and normal visual fields. Sequential examinations wereperformed on the right and left eyes. Both eyes were subjected toS-cone stimulation and combined L-and M-cone stimulation. Thus, fourrecordings were collected for each subject. Because of the experimen-

From the 1Institute of Biomedical Engineering and Informatics,Ilmenau University of Technology, Ilmenau, Germany; 2AugenZentrumSiegburg, Siegburg, Germany; and the 3Department of Neurology,Friedrich Schiller University of Jena, Jena, Germany.

Supported by Grants 13N8521 and 03IP605 from the GermanFederal Ministry of Education and Research.

Submitted for publication January 28, 2009; revised June 5 andAugust 10, 2009; accepted August 13, 2009.

Disclosure: P. Bessler, None; S. Klee, None; U. Kellner, None;J. Haueisen, None

Corresponding author: Patrick Bessler, Ilmenau University ofTechnology, Institute of Biomedical Engineering and Informatics, POB100 565, 98684 Ilmenau, Germany; [email protected].

Glaucoma

Investigative Ophthalmology & Visual Science, January 2010, Vol. 51, No. 1Copyright © Association for Research in Vision and Ophthalmology 319

tal setup, technical and biological artifacts occurred. Consequently,some of the recordings were excluded from further analysis.

The measurements were repeated in eight subjects, to assess re-producibility. Only one of the measurements was used in the subse-quent analysis (besides the reproducibility test).

Furthermore, 39 patients with glaucoma were examined in ananalogous manner. Four patients were excluded due to artifacts (Fig.1). The patients were medically prediagnosed and divided into twogroups: moderate and severe glaucomatous damage. All the patientswith glaucoma had a history of intraocular pressure above 21 mm Hg.The intraocular pressures of all the patients were normalized by med-ical therapy. The group with moderate glaucoma consisted of 29patients with visual acuities of between 0.6 and 1.0. Their visual fieldsshowed areas of mild to moderate sensitivity loss but no absolutescotomata. The severe glaucoma group consisted of six patients whohad absolute scotomata (Bjerrum scotoma [BS] and concentric con-striction [CC]) in their visual fields (Table 1).

An age-matched healthy subject group of approximately equal size(31 subjects from the 102 originally enrolled) was established (Fig. 1)to compare the control group of healthy subjects with the group of 29patients with moderate glaucoma.

Stimulation

A 30-in. liquid crystal display (Myrica V30–1; Fujitsu Siemens, Munich,Germany) was used with native display resolution (1280 � 768 pixels)and a frame rate of 60 Hz (Fig. 2D). The display was connected to astimulation unit (THERA PRAX; neuroConn GmbH, Ilmenau, Ger-

many). The frame rate was verified with a PIN-diode (BPX 65; Siemens,Munich, Germany) and a digital memory oscilloscope (TDS 3045;Tektronix, Beaverton, OR) by applying a 30 Hz black–white flickersequence to the display, which was generated by the stimulationsystem. The display was adjusted to the chromatic coordinates x �0.33 and y � 0.33 with a compact-array spectrometer (CAS 140B;Instrument Systems, Munich, Germany). The spectral distribution andtemporal constancy of the emission spectra of the red, green, and bluecolor channels were measured with the spectrometer. Gamma correc-tion was performed by determining the characteristic gamma curve forthe stimulator and the graphic card (8-bit VGA). The color channelintensities were 117 (red), 282 (green), and 47 (blue) cd/m2.

The RGB color values were transformed into the standard colorvalues of virtual XYZ space.33 Figure 2A shows the XYZ coordinates ofthe stimuli for the 10° CIE observer. Hunt fundamentals for 10° andlarger viewing conditions were used for the transformation to LMS-space.34 The transformation matrices for the applied stimulator systemare as follows:

MRGB3XYZ � �0.54 0.30 0.160.28 0.60 0.120.02 0.12 0.87

�MXYZ3LMS � � 0.39 0.69 �0.08

�0.23 1.18 0.050.00 0.00 1.00

�

FIGURE 1. Summary of patient characteristics, including the number of examined subjects and subgroups for analysis, number of subjects in eachsubgroup, age in years, and sex. The group of 102 healthy subjects was used to examine several color channels. The group of 31 healthy subjectswere age-matched and compared with the 29 patients with moderate glaucoma and 6 with severe glaucoma.

TABLE 1. Diagnosis of Patients with Severe Glaucoma

Patient 110 Patient 129 Patient 140 Patient 144 Patient 175 Patient 176

OS OD OS OD OS OD OS OD OS OD OS OD

Diagnosis BS Artifact BS BS BS Artifact CC CC BS BS BS BS

320 Bessler et al. IOVS, January 2010, Vol. 51, No. 1

A color combination for the S-cone stimulation (L1 � L2, M1 � M2,S1 � S2) and the combined L-and M-cone stimulation (L1 � L2, M1 �M2, S1 � S2) was calculated for the display. The activation of severalcone types and stimulation contrasts was determined (Fig. 2E), andthey could be selectively changed by adjusting the color combinations.We verified the S-cone isolation with the help of an adaption andbleaching experiment, which is based on the isolation technique fromStiles.35 The cone contrast (C) was determined by the following for-mula (according to the Michelson contrast)36

C ��EON � EOFF�

�EON � EOFF�

where, EON and EOFF represent the cone activation levels during theON and OFF phase, respectively. E is the amount of light absorbed perunit of retinal area for the three cone types.34

The PIN-diode was also used to analyze the electro-optical transientbehavior of the display caused by different color combinations of thestimulation sequences. The time between the starting and endingpoints of the presented color combination was measured by applyingthe S-cone and combined L- and M-cone sequence. The transient timeswere 6.2 ms for the S-cone stimulation and 5.4 ms for the L- and M-conestimulation. To correct the frame delay in the electroencephalogram(EEG), we analyzed the photodiode signal and the synchronizationtrigger from the stimulation unit. The trigger coincided with the onsetof stimulus generation by the unit, and the photodiode signal coin-cided with stimulus imaging by the display. The frame delay was 7.0 msfor all stimulations.

During the examination, the subject’s head was positioned on achin–forehead rest. The left and right eyes were tested sequentially;the eye not being tested was covered with an eye patch. The eye beingtested was positioned central to the stimulator at a distance of 0.5 m.This resulted in a maximum visual angle of 64° � 42° (Fig. 2B).

Circular flash stimulation with a fixation point was selected.37 Abalanced sequence was used for the stimulation order and for the orderin which the eyes were tested. The stimulus size was �11°. Allvolunteers were light adapted. Furthermore, the contribution of therod system was suppressed with an alternating stimulus and by theambient room luminance of 100 cd/m2. There was an adaption area

(�375 cd/m2, cold cathode fluorescent lamp spectrum with dominantwavelengths of �1 � 488 nm, �2 � 544 nm, �3 � 611 nm) surroundingthe circular stimulus area to eliminate rod responses during stimulationof the S-cones. The luminance of the adaption area resulted in a retinalilluminance of 3.3 log td.

The ON stimulus time was 17 ms and the OFF time was 767 ms. Anadditional random interstimulus interval (ISI) was chosen at 17 to 600ms to prevent the influence of periodic disturbance signals and habit-uation effects (Fig. 2C). A total of 160 repetitive stimulations wereperformed per examination.

Data Recording

The EEG signal was recorded simultaneously to the stimulation via theexamination system (THERA PRAX; neuroConn GmbH). EEG signalswere recorded with an electrode cap and Ag/AgCl ring electrodes(Easycap, Herrsching, Germany) placed over the visual cortex at Ozand the reference position Fz. The sample rate was 512 Hz. Thehardware trigger signal from the stimulation unit was used and ad-justed regarding the delay (see stimulation section), to precisely cor-relate the EEG signal with the excitation.

Preprocessing

Signal processing and analysis of EEG and VEP signals were performed(MatLab software; The MathWorks, Natick, MA). A digital softwarefilter was used to filter the electrode drift (0.8 Hz high-pass). Inaddition, a 30 Hz digital low-pass filter was used. The signal was filteredwith elliptic infinite impulse response filters in the forward and back-ward directions to prevent phase shifts. In addition, trials that hadphysiological signal distortions (e.g., muscle activity, increased alpha-activity, and eye movement) were detected by using artifact detectorsand excluded from the response signal analysis.38 A strong technicaldistortion in the area of the physiological response signal (17 Hz) alsohad to be eliminated by adapting and applying the matching pursuitmethod.39 A separate classification of the 160 recorded trials wasperformed. The correlation coefficient between the average signal ofall valid trials and each single valid trial was determined. The 100 trialswith the highest correlation coefficient were averaged. The constant

FIGURE 2. Schematic representationof the timing, configuration, and geom-etry of the stimuli, as well as the majorcolor values. (A) The CIE coordinates(X, Y, Z) of the S-cone stimulation (S)and combined L- and M-cone stimula-tion (LM) were based on the 1964 stan-dard colorimetric observer. (B) TheON and OFF terms corresponded tothe two colors of each stimulation se-quence and the stimulus size. (C) Thetimings and durations of the stimuli.A random ISI was chosen to reducehabituation effects, which can causealpha spindles. (D) Additional informa-tion about the stimulator (L, lumi-nance; R, red; G, green; B, blue).(E) The computed LMS values and re-sulting cone contrasts. S-cone stimula-tion (S) means the substitution (re-spectively no change in activation) ofthe L- and M-cones. Hence, the L and Mvalues for the ON and OFF terms mustbe equal. For the combined L- and M-cone stimulation (LM), the S-cones aresubstituted, and the S values (activa-tion of S-cones) are equal. Both stimu-lation conditions are optimized for ap-proximately equal and maximum conecontrast (C).

IOVS, January 2010, Vol. 51, No. 1 SST Stimulation in Different Glaucoma Stages 321

number of averaged trials guaranteed comparability of the responsesignals.

Analysis

As test parameters, we used the peak latency of the first negative wave(N1) and the first positive wave (P1), the peak-to-peak amplitude, theslope, and the area of the S-cone and combined L- and M-cone responsesignals (Fig. 3). These values were calculated by using automatedanalysis algorithms. The peak latency was determined by finding locallatency minima and maxima. These latency periods and correspondingamplitudes were used to calculate the peak-to-peak amplitude andslope. The area was determined by integrating all the amplitudes in therange of the main response signal (between the first and secondnegative wave). A manual correction was performed in the case of falseparameter detection.

The data were analyzed by multivariate analysis of variance(MANOVA) (SPSS software; SPSS Inc., Chicago, IL).

RESULTS

Healthy Subjects

Figure 3 shows two response signals for the stimulated colorchannels of a healthy subject; it also shows the parametersused to analyze the signals. The two typical response signalsdiffer in the curve shape.

A grand average response for the two stimulated colorchannels and the corresponding standard deviation (SD) wascalculated from the recordings of healthy subjects (Fig. 4).

We found variations in the curve shapes of the stimulatedcolor channels. The peak-to-peak amplitude of the S-cone re-sponse was 9 �V, which is markedly less than that of the L- andM-cone response (13 �V). The N1 latency of the S-cone re-sponse (111 ms) was greater than that of the L- and M-coneresponse (90 ms). The P1 latency was 182 ms for the S-coneresponse and 140 ms for the L- and M-cone response. Themaximum variance of the response signals occurred during themain response (50–400 ms), with an average SD of 2.6 �V forthe S-cone response and 3.8 �V for the L- and M-cone response.

Parameter analysis was performed to investigate the differ-ences between the color channels (Table 2). Statistical analysiswith MANOVA revealed significant differences between theselective stimulated color channels for all parameters.

Repeated measurements showed no differences in the re-sults.

Given the large differences in the ages of the subject groups(healthy subjects and patients with moderate glaucoma), therelationship between the analysis parameters and age wasevaluated in the healthy subject group. Examination of latencyshifts revealed an average increase in the S-cone response in N1latency of 4 ms per decade and in P1 latency of 3 ms perdecade. The average L- and M-cone response values werecalculated to be 5 ms per decade (N1 latency) and 4 ms perdecade (P1 latency).

Figure 5 shows the age structure of the test subjects. Basedon this distribution, two groups were formed (age ranges,10–44 and 47–85 years). As an example, Figure 6 shows therelationship between the calculated values and age for the N1latency parameter.

The statistical significance of the differences in analysis param-eters for the two age groups was studied by using MANOVA. Theresulting probabilities are PN1 � 1E-13, PP1 � 3E-4, Pslope � 0.13,Parea � 0.61, and Ppeak-to-peak amplitude � 0.02. The peak-to-peakamplitude, N1 and P1 showed a significant difference betweenthe two age groups (P 0.05), whereas the slope and areashowed no significant difference (P 0.05). However, the statis-tical test results do not rule out an age dependency of the param-eters. Therefore, the following observations were made whenworking with age-matched groups.

Patients with Glaucoma

Table 3 lists the averages and SDs of the analysis parameters forage-matched healthy subjects and patients with moderate glau-coma, the differences between both groups (�), and the sig-nificance of the differences (P-value). Significant differenceswere observed in some parameters between patients withmoderate glaucoma and healthy subjects. These differenceswere not observed in the L- and M-cone response.

Distinct changes in the S-cone response profiles were ob-served in patients with severe glaucoma damage comparedwith that of a healthy subject after selective S-cone stimulation(Fig. 7).

The response signal of the healthy subject shows a typical,distinct VEP (Fig. 7, top). Six of the patients with glaucoma (10S-cone response signals) were diagnosed with severe glaucomadamage. Six of the 10 response signals of these patientsshowed no VEP (Fig. 7, patient 176, left eye). Negativity ofresponse (Fig. 7, patient 110, left eye) instead of the typicalVEP was observed for three evaluations. One response showeda weak VEP despite diagnosis of concentric constriction. Thesecond eye of this patient showed no VEP. In summary, it canbe concluded that severe glaucoma damage is manifested inthe S-cone response for 9 of the 10 eyes.

FIGURE 4. Grand averages (solid curves) and SD (dashed curves) ofthe combined L- and M-cone response (top) and the S-cone response(bottom) of the healthy subjects.

FIGURE 3. Examples of S-cone responses (dashed curve) and com-bined L-and M-cone responses (solid curve) to SST stimulation and theanalysis parameters (N1, P1, slope, area, and peak-to-peak amplitude).


A change in the S-cone response with disease severity canalso be shown in the analysis of the grand averages of therespective group (Fig. 8).

A strong VEP (Fig. 8, top) was found in the S-cone responsegrand average of the age-matched healthy subject group. Aweaker signal and a change in latency were observed in theS-cone response grand average of patients with moderate glau-coma (Fig. 8, middle). The signal strength was distinctlyweaker for patients with severe glaucoma (Fig. 8, bottom). Thegrand average shows that there is no response signal for thisgroup.

DISCUSSION

To our knowledge, this study is the first in which optimalselective color channel stimulation has been performed basedon the SST method, together with an objective validation withVEP, in a large number of healthy subjects and in patients withglaucoma. We observed damage to the S-cone pathways inpatients with glaucoma and changes in the S-cone responsewith the severity of glaucoma.

Cone isolation was demonstrated by an adaption andbleaching experiment. The absence of the S-cone responseafter the bleaching procedure in contrast to the unaffected L-and M-cone response was the evidence of the successful coneisolation. As shown in Figure 2E, both stimulation conditionswere optimized for nearly equal and maximum cone contrast(�90%). However, optimization for equal cone contrast wasperformed, since it is known that latency variations are much

larger for different cone contrasts than for different luminancecontrasts.40 Circular flash stimulation was selected because thestudy was designed to have as few restrictions as possible interms of the health, age, and cooperation of the subjects.37

Onset–offset stimuli produce larger and more distinctive re-sponse signals to chromatic stimuli than do transient reversalsor pattern-reversal stimuli.41,42 The light adaption of the vol-unteers and the adaption area (retinal illuminance of 3.3 log td)should prevent the rod contribution to the response signals.43

It has been shown that glaucoma can be diagnosed basedon deficits in the blue color channel, making a selective, objec-tive examination of this color channel of great inter-est.7,8,10,13,16,21,24,44–46 Many of these previous studies wereperformed by using perimetry,7,45,46 which has the consider-able disadvantage of being a subjective method. In contrast,ERG (especially the pattern ERG) is a useful objective testingmethod.24,44,47–49 However, disease-related changes in the lay-ers of the visual pathways and the influence of cortical pro-cessing in the case of glaucoma could not be determined.21,22

VEP allows testing of the entire visual pathways. Some re-searchers, have performed VEP testing on patients with glau-coma.8,10,13,50 However, rather than the SST method, blue-on-yellow stimulus,8 black-white pattern reversal stimulus,10,50

and heterochromatic flicker photometry (HFP)13 were used tostimulate color channels.

In the present study, we performed an objective examina-tion of the visual pathways from the retina to the visual cortexby analyzing VEP. The results show that the S-cone response ofpatients with moderate glaucoma is different from that ofage-matched healthy subjects (Table 3). Latency shifts of 3 (N1)and 11 (P1) ms were determined. We found a significantdifference in the parameter slope (0.07 �V/ms) between the

FIGURE 5. Histogram of the age distribution in the healthy subjectgroup.

FIGURE 6. Example of age dependency: parameter N1 in the healthysubject group.

TABLE 2. Parameters for the S-cone Response and the Combined L-and M-cone Response of theHealthy Subjects

ParameterS-cone Response

(n � 178)L- and M-cone Response

(n � 183) � P

Nl (ms) � 110.0 95.0 15.0 0.001SD 16.2 18.6

P1 (ms) � 183.0 150.0 33.0 0.001SD 29.3 28.7

Slope (�V/ms) � 0.2 0.4 0.2 0.001SD 0.1 0.2

Area (�Vms) � 552.0 806.0 254.0 0.001SD 278.3 402.2

Peak-to-peak amplitude (�V) � 14.0 19.0 5.0 0.001SD 5.3 7.4


two groups. Therefore, the S-cone response revealed distinctdifferences between the two groups compared with the differ-ences of the combined L-and M-cone response (�N1 � 2 ms,�

P1� 3 ms, �slope � 0.03 �V/ms). A latency shift of the S-cone

response has also been observed by other researchers. Forexample, with blue-on-yellow stimulation, Horn et al.8 found apeak time difference of 7 to 19 ms between a glaucoma patientgroup and a control group. Rodarte et al.10 found differences ofapproximately 3 to 6 ms. However, we found only a minordifference in the latency shift of the L- and M-cone response.

This finding supports the assumption that glaucoma damage isinitially manifested in the blue color channel. Note that theanalyzed data originate from patients with moderate glaucomaand healthy subjects.

In addition to the latency shift, distinct changes occurred inthe configuration of the S-cone response toward the absence ofresponse (Fig. 7), especially in patients with severe glaucomadamage. It was not possible to determine parameters when theresponse signals were highly deformed or absent. However,these abnormal S-cone response profiles were visually deter-mined for 9 of 10 eyes in the group with severe glaucomadamage. Only one recording of an eye with severe glaucoma-tous damage had a detectable VEP; no VEP was recordablefrom the second eye of this patient. The remaining response inone eye could be explained by the different progression of thedisease in each eye.

Figure 8 shows a reduction in the response signal and a shiftin the latency of the grand average response signal of thepatients with moderate glaucoma (Fig. 8, middle) comparedwith the grand average response signal of the age-matchedhealthy subjects (Fig. 8, top). Furthermore, there is no defined

TABLE 3. Parameters of Age-Matched Groups of Healthy Subjects and Patients with Moderate Glaucoma by Selective Cone Stimulation

Parameter

S-cone Response L- and M-cone Response

Patient(n � 45)

Volunteer(n � 62) � (P)

Patient(n � 49)

Volunteer(n � 63) � (P)

Nl (ms) � 121.0 118.0 3.0 108.0 106.0 2SD 18.1 13.1 (P � 0.32) 28.6 14.9 (P � 0.58)

Pl (ms) � 199.0 188.0 11.0 165.0 162.0 3SD 33.1 28.9 (P � 0.07) 39.9 30.6 (P � 0.66)

Slope (�V/ms) � 0.16 0.23 0.07 0.31 0.34 0.03SD 0.1 0.1 (P � 0.01) 0.2 0.2 (P � 0.37)

Area (�Vms) � 486.0 573.0 87.0 707.0 715.0 8SD 408.6 307.9 (P � 0.23) 386.1 366.2 (P � 0.91)

Peak-to-peak amplitude (�V) � 12.0 14.0 2.0 14.0 16.0 2SD 10.2 6.3 (P � 0.21) 7.9 6.0 (P � 0.22)

The difference between the two groups and the significance are expressed by � and the P-value. respectively. The patient group did notinclude the patients with severe glaucoma.

FIGURE 7. S-cone response of a healthy subject and of patients withsevere glaucomatous damage; clear changes are apparent in the S-coneresponse profiles of the patients.

FIGURE 8. Change in the S-cone response (grand averages: solidcurves; SD: dashed curves) is dependent on glaucoma status. TheS-cone response of an age-matched healthy subject (top), patients withmoderate glaucoma (middle), and patients with severe glaucoma (bot-tom) are shown. Severity of glaucoma results in a latency shift andreduction of the response signal (middle) up to loss of VEP (bottom).


VEP in the case of severe glaucoma damage (Fig 8, bottom).Therefore, the severity of the disease can be assessed by ob-jective electrophysiological testing. Aldebasi et al.,13 using HFPstimulation, reported distinct changes in the configuration ofthe S-cone response in a comparison of a control group with agroup of patients with primary open-angle. Using blue-on-yellow stimulation, Horn et al.8 also found an increase in thelatency of the S-cone response in the progression of glaucoma.These findings support the hypothesis of a change in S-coneresponse with glaucoma progression.

In comparing the stimulated channels within the healthygroup (Fig. 4), the grand average response signals exhibitlarger amplitudes, a greater area, and an increased slope of theL- and M-cone response. The latency of N1 and P1 against theS-cone response was distinctly lower. It can be concluded thatprocessing is faster for the red-green channel. Porciatti andSartucci 40 and Rabin et al.42 reached the same finding by usingblue-yellow stimulation in conjunction with sinusoidal grat-ings. The SDs of the calculated grand averages (Fig. 4) mirrorthe large variances of the interindividual response signals. Thisfinding can be attributed to the fundamentally high variabilityof the EEG, in addition to the individually varying responsesignals of the stimulations.

The extracted VEP parameter also showed significant differ-ences (P 0.001) between the two stimulated color channels(Table 2), as seen in the grand average curve shape (Fig. 4).The higher peak-to-peak amplitude, slope, and area of the L-and M-cone response represent a distinctly stronger responsesignal in the stimulation of L- and M-cones. The latency shiftbetween the S-cone response and the combined L- and M-coneresponse was 15 to 33 ms. Other researchers have also deter-mined differences between color channels. Latency shifts be-tween color channels have been observed even in infants.41

Rabin et al.42 determined a difference of 25 to 30 ms in adultsin blue-yellow stimulation with sinusoidal gratings, and Robsonand Kulikowski51 found a latency shift of 13 ms between colorchannels with blue-yellow stimulation. Porciatti and Sartucci40

found comparable waveforms with greater amplitudes andshorter latency for L- and M-cone response after onset VEPexaminations for S-cone response and L- and M-cone response.Differences in reaction times during color channel examina-tions have been attributed to the visual processing system.52

Processing and further transmission of color-selective stimula-tion up to the visual cortex occurs in different systems (par-vocellular layers, red-green information; koniocellular layers,blue-yellow information)21,53,54 and probably causes differ-ences in further signal processing.55,56

The latency shifts of the selectively stimulated color chan-nels observed in the present study correspond to the physio-logical characteristics of color-opponent neurons and process-ing in the parvocellular and koniocellular visual pathways.40

Possible reasons for latency shifts can be determined even inthe very early processing stages of the visual system. ERGstudies reveal that the response signals for different colorchannels have different latencies and amplitudes from the firstprocessing steps on the retina.57,58 The latency is also depen-dent on the luminance of stimuli and the respective contrast ofactivation.

Our analysis of the influence of age on VEP parametersshowed significant differences between the two age groups inN1, P1, and peak-to-peak amplitude. The increase in latencywas approximately 4 ms per decade. Similar age-related influ-ences on VEP have been observed by other researchers.9,59,60

Latency shifts of approximately 8 ms per decade have beenobserved in onset and reversal studies, amplitude valuesdropped slightly, although a fundamental change in form couldnot be determined.9,60 In a multifocal black–white patternreversal stimulation, Rodarte et al.10 determined a slight age-

related dependency of 1.3 ms per decade. In the present studythe parameters slope and area did not show any significantdifferences between the two age groups; however, an age-related dependency could not be ruled out. Therefore, age-related dependencies should be taken into consideration whentesting. Possible reasons for age-related dependencies of VEPare the effects of aging on the photoreceptors, a reduction incontrast sensitivity, or an increase in the opacity of the eye.60

We presented the results of tests that involved optimalselective color channel stimulation based on SST methodologyand an objective evaluation with VEP, as performed on a largenumber of healthy subjects and patients with glaucoma. Theapplied stimulation methods offered advantages in objectivityand flexibility of selective color channel stimulation. This ap-proach could enable new diagnostic possibilities and improvedearly diagnosis. A receiver operating characteristic analysis ofthe classification, based on the developed parameters, shouldbe performed to quantify the impact of the method for diag-nosis. To further validate the significance of the methodology,progressive follow-up studies or studies with a large number ofpatients who are at various stages of the disease as well as thedetailed observation of accompanying diseases are necessary.

Acknowledgments

The authors thank Silke Weinitz and Sylvi Herzog for performing theexaminations.

References

1. Resnikoff S, Pascolini D, Etya’ale D, et al. Global data on visualimpairment in the year 2002. Bull World Health Organ. 2004;82:844–851.

2. Rait JL. Management of ocular hypertension and primary openangle glaucoma. Clin Exp Optom. 2000;83:136–144.

3. Vistamehr S, Shelsta HN, Palmisano PC, et al. Glaucoma screeningin a high-risk population. J Glaucoma. 2006;15:534–540.

4. Cockburn DM. Diagnosis and management of open angleglaucoma: suggested guidelines for optometrists. Clin Exp Optom.2000;83:119–127.

5. Hong SM, Ahn H, Ha SJ, Yeom HY, Seong GJ, Hong YJ. Earlyglaucoma detection using the Humphrey Matrix perimeter, GDxVCC, stratus OCT, and retinal nerve fiber layer photography. Oph-thalmology. 2007;114:210–215.

6. Medeiros FA, Ng D, Zangwill LM, Sample PA, Bowd C, WeinrebRN. The effects of study design and spectrum bias on the evalua-tion of diagnostic accuracy of confocal scanning laser ophthalmos-copy in glaucoma. Invest Ophthalmol Vis Sci. 2007;48:214–222.

7. Johnson CA, Adams AJ, Casson EJ, Brandt JD. Progression of earlyglaucomatous visual-field loss as detected by blue-on-yellow andstandard white-on-white automated perimetry. Arch Ophthalmol.1993;111:651–656.

8. Horn FK, Jonas JB, Budde WM, Junemann AM, Mardin CY, KorthM. Monitoring glaucoma progression with visual evoked potentialsof the blue-sensitive pathway. Invest Ophthalmol Vis Sci. 2002;43:1828–1834.

9. Crognale MA. Development, maturation, and aging of chromaticvisual pathways: VEP results. J Vis. 2002;2:438–450.

10. Rodarte C, Hood DC, Yang EB, et al. The effects of glaucoma onthe latency of the multifocal visual evoked potential. Br J Ophthal-mol. 2006;90:1132–1136.

11. Tobimatsu S, Celesia GG. Studies of human visual pathophysiologywith visual evoked potentials. Clin Neurophysiol. 2006;117:1414–1433.

12. Bach M. Electrophysiological approaches for early detection ofglaucoma. Eur J Ophthalmol. 2001;11(suppl)2:S41–S49.

13. Aldebasi YH, Drasdo N, Morgan JE, North RV. Cortical OFF-poten-tials from the S-cone pathway reveal neural damage in early glau-coma. Vision Res. 2003;43:221–226.

14. Crognale MA, Switkes E, Rabin J, Schneck ME, Haegerstromport-noy G, Adams AJ. Application of the spatiochromatic visual-evoked


potential to detection of congenital and acquired color-vision de-ficiencies. J Opt Soc Am A-Opt Image Sci Vis. 1993;10:1818–1825.

15. Sartucci F, Murri L, Orsini C, Porciatti V. Equiluminant red-greenand blue-yellow VEPs in multiple sclerosis. J Clin Neurophysiol.2001;18:583–591.

16. Drance SM, Lakowski R, Schulzer M, Douglas GR. Acquired colorvision changes in glaucoma: use of 100-hue test and Pickfordanomaloscope as predictors of glaucomatous field change. ArchOphthalmol. 1981;99:829–831.

17. Curcio CA, Allen KA, Sloan KR, et al. Distribution and morphologyof human cone photoreceptors stained with anti-blue opsin.J Comp Neurol. 1991;312:610–624.

18. Craft CM, Whitmore DH, Wiechmann AF. Cone arrestin identifiedby targeting expression of a functional family. J Biol Chem. 1994;269:4613–4619.

19. Swanson WH, Birch DG, Anderson JL. S-Cone function in patientswith retinitis-pigmentosa. Invest Ophthalmol Vis Sci. 1993;34:3045–3055.

20. Nork TM, Mccormick SA, Chao GM, Odom JV. Distribution ofcarbonic-anhydrase among human photoreceptors. Invest Oph-thalmol Vis Sci. 1990;31:1451–1458.

21. Gupta N, Ang LC, de Tilly LN, Bidaisee L, Yucel YH. Humanglaucoma and neural degeneration in intracranial optic nerve,lateral geniculate nucleus, and visual cortex. Br J Ophthalmol.2006;90:674–678.

22. Yucel YH, Zhang QA, Weinreb RN, Kaufman PL, Gupta N. Effectsof retinal ganglion cell loss on magno-, parvo-, koniocellular path-ways in the lateral geniculate nucleus and visual cortex in glau-coma. Prog Retin Eye Res. 2003;22:465–481.

23. Mortlock KE, Chiti Z, Drasdo N, Owens DR, North RV. Silentsubstitution S-cone electroretinogram in subjects with diabetesmellitus. Ophthalmic Physiol Opt. 2005;25:392–399.

24. Drasdo N, Aldebasi YH, Chiti Z, Mortlock KE, Morgan JE, NorthRV. The S-cone PhNR and pattern ERG in primary open angleglaucoma. Invest Ophthalmol Vis Sci. 2001;42:1266–1272.

25. Kellner U, Foerster MH. Electroretinography with color stimuli toseparate cone dystrophies. Klin Monatsbl Augenheilkd. 1992;201:102–106.

26. Stiles WS, Wyszecki G. Color-matching data and spectral absorp-tion curves of visual pigments. Vision Res. 1974;14:195–207.

27. Boynton RM, Hayhoe MM, Macleod DIA. Gap effect: chromatic andachromatic visual-discrimination as affected by field separation.Opt Act. 1977;24:159–177.

28. Donner KO, Rushton WAH. Retinal stimulation by light substitu-tion. J Physiol (Lond). 1959;149:288–302.

29. Estevez O, Spekreijse H. A spectral compensation method fordetermining the flicker characteristics of the human colour mech-anisms. Vision Res. 1974;14:823–830.

30. Macleod DIA, Boynton RM. Chromaticity diagram showing coneexcitation by stimuli of equal luminance. J Opt Soc Am. 1979;69:1183–1186.

31. Smith VC, Pokorny J. Spectral sensitivity of foveal cone photopig-ments between 400 and 500 Nm. Vision Res. 1975;15:161–171.

32. Estevez O, Spekreijse H. The “silent substitution” method in visualresearch. Vision Res. 1982;22:681–691.

33. Lang H. Farbwiedergabe in den Medien. Fernsehen, Film, DruckGottingen: Muster-Schmidt; 1995.

34. Hunt RWG. Measuring colour. Ellis Horwood Series in AppliedScience and Industrial Technology. London: Ellis Horwood; 1995.

35. Stockman A, Sharpe LT. Cone spectral sensitivities and colormatching. In: Gegenfurtner KR, Sharpe LT, eds. Color Vision.Cambridge, UK: Cambridge University Press; 1999:52–87.

36. Albrecht J, Jagle H, Hood DC, Sharpe LT. The multifocal electro-retinogram (mfERG) and cone isolating stimuli: variation in L- andM-cone driven signals across the retina. J Vis. 2002;2:543–558.

37. Odom JV, Bach M, Barber C, et al. Visual evoked potentials stan-dard. Doc Ophthalmol. 2004;108:115–123.

38. Sporckmann G. Signalerfassung, Signalverarbeitung und Mus-tererkennung bei visuell evozierten Potentialen zur verbessertenobjektiven Diagnostik der menschlichen Sehleistung. Aachen,Germany: RWTH Aachen University; 1996. Thesis.

39. Gratkowski M, Haueisen J, Arendt-Nielsen L, Chen AC, Zanow F.Decomposition of biomedical signals in spatial and time-frequencymodes. Methods Inform Med. 2008;47:26–37.

40. Porciatti V, Sartucci F. Normative data for onset VEPs to red-greenand blue-yellow chromatic contrast. Clin Neurophysiol. 1999;110:772–781.

41. Crognale MA, Kelly JP, Weiss AH, Teller DY. Development of thespatio-chromatic visual evoked potential (VEP): a longitudinalstudy. Vision Res. 1998;38:3283–3292.

42. Rabin J, Switkes E, Crognale M, Schneck ME, Adams AJ. Visualevoked potentials in three-dimensional color space: correlates ofspatio-chromatic processing. Vision Res. 1994;34:2657–2671.

43. Stockman A, Sharpe LT. Into the twilight zone: the complexities ofmesopic vision and luminous efficiency. Ophthalmic Physiol Opt.2006;26:225–239.

44. Holopigian K, Greenstein VC, Seiple W, Hood DC, Ritch R. Elec-trophysiologic assessment of photoreceptor function in patientswith primary open-angle glaucoma. J Glaucoma. 2000;9:163–168.

45. Dejong LAMS, Snepvangers CEJ, Vandenberg TJTP, LangerhorstCT. Blue-yellow perimetry in the detection of early glaucomatousdamage. Doc Ophthalmol. 1990;75:303–314.

46. Maeda H, Tanaka Y, Nakamura M, Yamamoto M. Blue-on-yellowperimetry using an Armaly glaucoma screening program. Ophthal-mologica. 1999;213:71–75.

47. Hood DC, Greenstein VC, Holopigian K, et al. An attempt to detectglaucomatous damage to the inner retina with the multifocal ERG.Invest Ophthalmol Vis Sci. 2000;41:1570–1579.

48. Bach M, Unsoeld AS, Philippin H, et al. Pattern ERG as an earlyglaucoma indicator in ocular hypertension: a long-term, prospec-tive study. Invest Ophthalmol Vis Sci. 2006;47:4881–4887.

49. Bach M, Hoffmann MB. Update on the pattern electroretinogram inglaucoma. Optom Vis Sci. 2008;85:386–395.

50. Hood DC, Zhang X, Greenstein C, et al. An interocular comparisonof the multifocal VEP: a possible technique for detecting localdamage to the optic nerve. Invest Ophthalmol Vis Sci. 2000;41:1580–1587.

51. Robson AG, Kulikowski JJ. Objective specification of tritanopicconfusion lines using visual evoked potentials. Vision Res. 1998;38:3499–3503.

52. McKeefry DJ, Parry NR, Murray IJ. Simple reaction times in colorspace: the influence of chromaticity, contrast, and cone oppo-nency. Invest Ophthalmol Vis Sci. 2003;44:2267–2276.

53. Gegenfurtner KR. Cortical mechanisms of colour vision. Nat RevNeurosci. 2003;4:563–572.

54. Hendry SHC, Reid RC. The koniocellular pathway in primate vi-sion. Ann Rev Neurosci. 2000;23:127–153.

55. Dacey DM. Primate retina: cell types, circuits and color oppo-nency. Prog Retin Eye Res. 1999;18:737–763.

56. Sumner P, Anderson EJ, Sylvester R, Haynes JD, Rees G. Combinedorientation and colour information in human V1 for both L-M andS-cone chromatic axes. Neuroimage. 2008;39:814–824.

57. Gouras P, Mackay CJ. Electroretinographic responses of the short-wavelength-sensitive cones. Invest Ophthalmol Vis Sci. 1990;31:1203–1209.

58. Sawusch M, Pokorny J, Smith VC. Clinical electroretinography forshort wavelength sensitive cones. Invest Ophthalmol Vis Sci.1987;28:966–974.

59. Onofrj M, Thomas A, Iacono D, D’Andreamatteo G, Paci C. Age-related changes of evoked potentials. Clin Neurophysiol. 2001;31:83–103.

60. Fiorentini A, Porciatti V, Morrone MC, Burr DC. Visual ageing:unspecific decline of the responses to luminance and colour.Vision Res. 1996;36:3557–3566.


glaucoma silent substitution stimulation of s-cone pathway ... · silent substitution stimulation...

Documents