- Email: [email protected]
Response phase of the flicker electroretinogram (ERG) is influenced by cone excitation strength
Vision Research 38 (1998) 3247 – 3251
Response phase of the flicker electroretinogram (ERG) is influenced by cone excitation strength Tomoaki Usui a,b, Jan Kremers a,*, Lindsay T. Sharpe a, Eberhart Zrenner a a
Department of Experimental Ophthalmology, Uni6ersity Eye Hospital, Ro¨ntgenweg 11, 72076 Tu¨bingen, Germany b Department of Ophthalmology, Niigata Uni6ersity School of Medicine, Niigata, Japan Received 10 July 1997; received in revised form 27 July 1997; accepted 23 December 1997
Abstract We measured electroretinogram (ERG) response phases at different cone contrasts in trichromats and dichromats to investigate the dynamics of the long-wavelength-sensitive (L-) and middle-wavelength-sensitive (M-) cone pathways. ERG responses to stimuli, temporally modulated at 30 Hz, were recorded. The stimuli were generated on a computer controlled colour monitor. Thirty-two different combinations of L- and M-cone excitation strength, expressed as cone contrasts, were presented. The short-wavelength-sensitive (S-) cones were not stimulated (S-cone contrast =0%). The response phase of the fundamental stimulus component was obtained from Fourier analysis. The ERG response phase lags decreased with increasing cone contrast. This was observed in all subjects with a normal appearing fundus. In dichromats and trichromats at low and intermediate contrasts, the phase lags to M-cone isolating conditions were smaller than those to L-cone isolating stimuli. In one dichromat with extreme myopia and cupping of the optic disc, the ERG phase lags increased with increasing cone contrast. The ERG response phase may be potentially useful for detecting retinal abnormalities. © 1998 Elsevier Science Ltd. All rights reserved. Keywords: Electroretinogram (ERG); Response phase; Cone contrast; Dichromacy; Trichromacy
1. Introduction It is controversial whether the dynamics of the primate long-wavelength-sensitive (L-) and middle-wavelength-sensitive (M-) cones differ or not. The time to peak response and the integration time of single macaque M-cones recorded with suction electrodes [1] appear to be shorter than those of L-cones. Yet the differences are not significant. However, data from several other studies, involving additional contributions from post-receptoral mechanisms, suggest that the Mcone pathway responds faster than the L-cone pathway. The data include cone electroretinogram (ERG) recordings in monkey [2] and in human [3,4], psychophysical measurements in humans [5] and ganglion cell recordings in the macaque retina [6 – 8]. In this paper, we add new data about L- and M-cone dynamics obtained from flicker ERG recordings in human subjects.
* Corresponding author: Tel: +49 7071 2985031; fax: + 49 7071 295777; e-mail: [email protected]. 0042-6989/98/$19.00 © 1998 Elsevier Science Ltd. All rights reserved. PII: S0042-6989(98)00046-7
We modified the ‘silent substitution’ technique (see review in Ref. [9]) and the classic heterochromatic flicker photometric procedure (see review in Ref. [10], for the implementation in ERG recordings) so that we could stimulate the L- or M-cones in a predefined manner. This procedure has been used to discriminate protanopes from deuteranopes and to demonstrate individual variations in the ERG amplitudes of L- and M-cone signals in normal trichromats [11]. Here we report the dependence of the ERG phase on photoreceptor contrast in dichromats and trichromats. Results about the dependence of response amplitude on photoreceptor contrast are reported elsewhere [11]. Part of this study has been presented in abstract form [12].
2. Subjects Fifteen dichromats (all males; six protanopes, and nine deuteranopes;) and eight trichromats (six males and two females) took part in this study. The classifica-
T. Usui et al. / Vision Research 38 (1998) 3247–3251
3248
tion of dichromacy was based upon Rayleigh matches obtained on a Nagel type I anomaloscope (Schmidt and Haensch, Germany) and was consistent with molecular genetical analysis of blood samples (by J. Nathans, personal communication). Of the protanopes, four had multiple genes and two had single genes. All had the alanine amino acid at position 180. Of the deuteranopes, four had multiple genes and four had single genes; all had serine at position 180. Normal ophthalmological findings were evident in all subjects, except for one protanope (subject AO) who had high myopia ( − 8.0 diopter) bilaterally with a typical fundus appearance, including vascular narrowing and optic disc cupping. Informed consent was obtained from all subjects after explanation of the purpose of the study.
3. Methods The methods, including visual stimuli and recording techniques, have been described in detail elsewhere [11]. Briefly, the stimuli were presented on a computer controlled monitor (BARCO, frame rate: 100 Hz) with a VSG 2/2 graphics card (Cambridge Research System). The monitor subtended 124 ×108° at the 10 cm viewing distance. The red, green and blue phosphors of the monitor were temporally modulated at 30 Hz for predefined Michelson contrasts. Contrast was defined as 100%×
(Lmax −Lmin) (Lmax +Lmin)
where Lmax and Lmin are the maximum and minimum luminances, respectively. A power spectrum analysis revealed that the main frequency components in the stimulus occur at 0 Hz (the DC level), the 100 Hz frame rate and the 30 Hz modulation frequency. Frequency components arising from interference between the frame rate and the modulation frequency were very small in magnitude (i.e. less than 8% of the power at 30 Hz). The time averaged luminance of the monitor was 66 cd/m2 (40 cd/m2 for the green phosphor, 20 cd/m2 for the red phosphor and 6 cd/m2 for the blue phosphor). The time averaged chromaticity in ([13]) large field coordinates was: x = 0.33, y =0.32. The mean quantal catches were very similar for the L- and Mcones (assuming a pupil diameter of 8 mm and a foveal cone collecting area of 2.92 mm2 [14]): the mean foveal L-cone quantal catch was 4.42 log quanta s − 1 cone − 1 and the mean foveal M-cone quantal catch was 4.31 log quanta s − 1 cone − 1. The spectral characteristics of the monitor phosphors were measured with a spectroradiometer (Instrument Systems). The monitor calibration and the gamma correction were performed using the internal luminance device of the BARCO monitor and the VSG software. The sensitivities of each cone type to the phosphors were calculated by multiplying
the emission spectra with psychophysically-derived estimates of the corneal sensitivity spectra of the cones [15]. The sensitivities were then used to calculate the modulation contrasts of each phosphor in order to obtain the desired L- or M-cone contrast. S-cone contrast was 0% in all conditions. We measured the ERGphases in the dichromats for a series of contrasts, with a maximum L- or M-cone contrast of 78.7%. For trichromats, there were fewer conditions resulting in pure L- and pure M-cone modulation. The maximal M-cone contrast was 28.6% and the maximal L-cone contrast 22.9%. One eye was dilated with a mydriatic (0.5% tropicamide) and kept light-adapted at least 10 min before the ERG recordings began. Corneal ERG responses were measured with a DTL fiber electrode (UniMed Electrode Supplies). The reference and ground skin electrodes were attached to the ipsilateral temple and the forehead, respectively. Signals were amplified and band-pass filtered with 10 and 100 Hz corner frequencies. The signal was sampled at 1000 Hz with a CED 1401 on-line computer. Forty-eight ERG measurements, each lasting 1 s, were averaged. ERG responses were Fourier analysed and the ERG response amplitude and phase were defined as the amplitude and phase of the fundamental stimulus components. Power spectra of the responses revealed that, at the 30 Hz temporal frequency, the response is mainly determined by the fundamental component. It is difficult to speculate about the cellular origin of this component, but the b-wave is traceable up to 33 Hz and the a-wave only up to 20 Hz [16], suggesting that the direct origin is not in the photoreceptors themselves. Stimulus artefacts were measured with the monitor covered by black cardboard. Although the stimulus artefacts were normally very small, ERG signals were corrected by vector subtracting the stimulus artefact from the measured signal. More detailed information about correcting for stimulus artefact is given elsewhere [11]. The use of a CRT screen has inherent temporal limitations which in principle can interfere with ERG measurements [17]. We were, however, careful to verify that the response phase effects, described in the present paper, had in fact a physiological origin and were not determined by the properties of the CRT screen. One possible source of error is that the three different phosphors of the monitor have different times to maximal excitations and different decay times. The overall excitation and decay time of the phosphors is determined by the amount of excitations of the individual phosphors, which might influence the response phase. We excluded this possible source of error by measuring the response amplitudes and phases in dichromats using different stimulus conditions, in which the phosphor excitations differed but the cone contrasts did not. We
T. Usui et al. / Vision Research 38 (1998) 3247–3251
found that the response amplitudes and phases were not significantly different between these conditions, indicating that the response phase is indeed determined by the cone contrast and not by the excitation strength of the phosphors. A second possible source of error might arise from the sequential excitations of the individual monitor pixels. The pixels in the lower part of the monitor are excited later than pixels in the upper part of the monitor. This introduces a substantial phase shift at the 30 Hz stimulus frequency. Thus, any difference in the L- and M-cone distribution between the upper and lower part of the retina could produce a measured phase difference between the responses of the two cone types. However, in that case, it would be expected that the phase difference is reversed by rotating the monitor 180°. A control experiment with the monitor positioned upside down, however, did not result in a sign reversal of the phase differences.
3249
than the response amplitude to L-cone isolating stimuli at similar cone contrasts [11], resulting in less reliable M-cone ERG-phases. The fits through the data obtained in the dichromats are replotted in Fig. 1(B), enabling a comparison between the dichromatic and trichromatic data. Clearly, a similar relationship between response phase and L-cone contrast is found in deuteranopes and trichromats. The similarity between the response phases of protanopes and trichromats to M-cone stimuli is less satisfactory, probably because of the less reliable trichromatic data arising from the
4. Results In Fig. 1(A) the phases of the fundamental components of the ERG responses in the protanopes (filled circles) and deuteranopes (open circles) are plotted as a function of the M- or L-cone contrast. Since the Fourier analysis only gives phases within a 360° range, we estimated the absolute response phases at 30 Hz by comparing them with response phases at other temporal frequencies (unpublished data) and from explicit times which have been reported as ranging between 25 and 30 ms for 30 Hz flicker ERGs [18]. By definition, a negative response phase indicates a phase lag. The response phase increases with increasing contrast, indicating that the phase lag decreases. But a plateau is reached between 20 and 40% contrast. The relation between cone contrast and ERG phase can be described by an exponential function: P(C) = k1 + k2(1− e − k3 · C) where P(C) is the response phase, C is the cone contrast, and k1, k2 and k3 are constants. The function was separately fitted to the data points of the protanopes and deuteranopes using the solver routine in the Excel 4.0 for Windows program. The values for the constants in the fits are: k1 = −320.6°, k2 =43.6° and k3 = 0.09 for the protanopes; and k1 = −364.2°, k2 =73.0° and k3 = 0.07 for the deuteranopes. The scatter in the data is to a large extent caused by individual differences, which were highly reproducible. The response phases as a function of cone contrast in five trichromats are plotted in Fig. 1(B). There were fewer conditions for trichromats than for dichromats in which the response of a single cone type could be isolated. Further the response amplitudes in trichromats to M-cone isolating stimuli are normally smaller
Fig. 1. (A) Response phase of the fundamental ERG component (in degrees) as a function of cone contrast (in percent) for M-cones measured in protanopes (filled circles) and for L-cones measured in deuteranopes (open circles). The curves represent the best fitting exponential function (see equation in text; drawn curve for M-cones, dashed curved for L-cones). (B) Response phase as a function of cone contrast measured in trichromats to L- (open diamonds) and M-cone (closed diamonds) isolating stimuli. For trichromats, there are fewer conditions in which a single cone is selectively stimulated. Further, the maximal achievable cone contrast is smaller than in dichromats. The response amplitudes to M-cone isolating stimuli is generally smaller than those to L-cone isolating stimuli, resulting in less reliable M-cone response phases. The fits to the dichromatic data are replotted for comparison.
3250
T. Usui et al. / Vision Research 38 (1998) 3247–3251
5. Discussion
Fig. 2. Comparison of the ERG response phases determined from Mand L-cone isolating stimuli presented to trichromats, protanopes and deuteranopes. Cone contrast is between 20 and 30% for all subjects. Data points are averages of at least two separate measurements. The trichromatic data, obtained in the same individual, are connected by a line. The data from one protanope (AO) are not included in this figure.
smaller response amplitudes. As for the dichromats, highly reproducible individual differences in the absolute phases of the trichromats are observed. Fig. 2 presents a comparison of the phase data of all trichromats, protanopes and deuteranopes. Only data measured at M- and L-cone contrasts between 20 and 30% are included. These contrasts were large enough to obtain reliable responses, but low enough to show clear L- and M-cone ERG phase differences between the dichromats. The phases of M- and L-cone ERGs in the trichromats (which are connected by lines for the individual subjects) correspond to those in the dichromats, indicating that the receptoral and postreceptoral pathways are similar for the trichromats and the dichromats. The phase differences, therefore, probably reflect physiological differences between the pathways mediating the L- and M-cone ERGs. Interestingly, the phases in one protanope (AO) with high myopia are distinctly different. His phases, which were measured three times in separate sessions, decrease with increasing cone contrast (Fig. 3; same scale as Fig. 1). Especially at the high M-cone contrasts, where large response amplitudes can be obtained, the phases are highly reproducible. The constants for the fit are: k1 = − 284.6°, k2 = − 81.6° and k3 =0.02. Note that k2 has a negative value.
The ERG phases increase with increasing cone contrast up until contrasts of 20–40% (see Fig. 1). It is not clear what mechanism is responsible for the phase shifts, but they clearly signify a nonlinearity in the ERG pathway. In retinal ganglion cells of the cat [19], macaque [20] and marmoset [21] and in LGN cells of the marmoset [22] similar phase shifts have been reported and described as a contrast gain control mechanism. Possibly, similar processes play a role in the ERG nonlinearities. But it is unlikely that the ERG nonlinearities have the same cellular origins as the nonlinearities in the single cell responses, because the phase shifts in the single cell recordings occur at frequencies between 4 and 15 Hz. The 30 Hz stimulus used in our experiments is clearly outside this range. Further, the contrast gain control found in the retinal ganglion cells occurs in combination with response saturation; whereas the ERG signals do not saturate. At low and intermediate contrasts, the M-cone ERG phases (measured in protanopes and trichromats) are larger than the L-cone ERG phases (measured in deuteranopes and trichromats) by about 40° (Fig. 2), indicating that M-cone signals, measured with the ERG, are transmitted about 3.7 ms faster than L-cone signals. At high contrasts, however, the phase differences are smaller (about 10°). Phase data obtained at other temporal frequencies are required to quantify the delay difference accurately, and to establish whether a
Fig. 3. Phase of the fundamental components of the ERG response (in degrees) as a function of M-cone contrast (in percent) in the protanope (AO) with high myopia. The measurements were repeated four times at each cone contrast (filled circles). The curve represents the best fitting exponential function. Note that the constant k2, for the exponential fit, has a negative value (see text for details).
T. Usui et al. / Vision Research 38 (1998) 3247–3251
delay difference suffices to explain the phase differences. We found phase differences between 10 and 105° in individual trichromats (Fig. 2), suggesting delay differences ranging between 1.0 and 9.7 ms. But, the phases of the M-cone ERGs are less reliable than those of the L-cone ERGs in the trichromats. This will necessarily influence the variability in phase difference. Whitmore and Bowmaker also reported that M-cone ERG signals are phase advanced relative to L-cone ERG signals [4]. But the difference they obtained (12 ms) was much larger than the values reported here. Yeh et al. reported that parvocellular projecting ganglion cells with M-cone input to the centres responded about 3.8 ms faster than cells with L-cone input [8]. Although it is interesting to speculate that the latency differences result wholly or mainly from differences in the M- and L-cones kinetics, differences in the post-receptoral mechanisms cannot be ignored. In recordings from single cones to short flashes, Schnapf et al. found an average time to peak response of about 51 ms in M-cones and 55 ms in L-cones, although this difference was not statistically significant [1]. However, the psychophysical data of Hamer and Tyler [5] also suggest that the pathways subserving the M-cones respond faster than those subserving the Lcones. On the other hand, Lutze et al. [23] reported that deuteranopes, for whom the L-cones are mediating detection, have a higher critical flicker fusion than protanopes, for whom the M-cones are mediating detection. This is seemingly in contradiction with our results. But, a change in delay between L- and M-cone responses does not necessarily change the critical flicker fusion. Differences between L- and M-cones in critical flicker fusion frequency, whether measured in normal trichromats or dichromats, might involve other mechanisms than a simple time delay. Interestingly, in contrast to all other dichromatic subjects, one protanope showed decreasing ERG phase with increasing cone contrast. His high myopia with typical fundus change might have contributed to this unusual phase behaviour. Further studies are required to establish whether such ERG phase anomalies can be used in the diagnosis of retinal disorders.
Acknowledgements The authors wish to thank Sabine Meierkord and Eva Burkhardt for technical assistance, and Dr Stefan Weiss for computer programming assistance. This study was supported by fortu¨ne-Grant F.1222147.2 (Tu¨bingen, Germany) and Deutsche Akademisches Austauschdienst (Bonn, Germany) to TU, DFG grant Zr 1/9-3 (Bonn, Germany) to JK and EZ, and a Hermann-undLilly-Schilling-Stiftung fellowship to LTS.
3251
References [1] Schnapf JL, Nunn BJ, Meister M, Baylor DA. Visual transduction in cones of the monkey Macaca fascicularis. J Physiol 1990;427:681 – 713. [2] Yamamoto S, Gouras P, Lopez R. The focal cone electroretinogram. Vis Res 1995;35:1641 – 9. [3] Spileers W, Falcao-Reis F, Hogg C, Arden GB. Evidence from human electroretinogram a and off responses that color processing occurs in the cones. Invest Ophthalmol Vis Sci 1993;34:2079– 91. [4] Whitmore AV, Bowmaker JK. Differences in the temporal properties of human longwave- and middlewave-sensitive cones. Eur J Neurosci 1995;7:1420 – 3. [5] Hamer RD, Tyler CW. Analysis of visual modulation sensitivity. V. Faster visual response for G- than for R-cone pathway? J Opt Soc Am 1992;A9:1889 – 904. [6] Smith VC, Lee BB, Pokorny J, Martin PR, Valberg A. Responses of macaque ganglion cells to the relative phase of heterochromatically modulated lights. J Physiol 1992;458:191 – 221. [7] Lee BB, Martin PR, Valberg A, Kremers J. Physiological mechanisms underlying psychophysical sensitivity to combined luminance and chromatic modulation. J Opt Soc Am 1993;A10:1403– 12. [8] Yeh T, Lee BB, Kremers J. Temporal responses of ganglion cells of the macaque retina to cone-specific modulation. J Opt Soc Am 1995;A12:456 – 64. [9] Este´vez O, Spekreijse H. The ‘silent substitution’ method in visual research. Vis Res 1982;22:681 – 91. [10] Jacobs GH, Neitz J, Krogh K. Electroretinogram flicker photometry and its applications. J Opt Soc Am 1996;A13:641–8. [11] Usui T, Kremers J, Sharpe LT, Zrenner E. Flicker cone ERG in dichromats and trichromats. Vis Res 1998;38:3391 – 6. [12] Usui T, Kremers J, Zrenner E. Phase differences between L- and M-cone ERG. Invest Ophthalmol Vis Sci 1997;38 Suppl:885 (abstract). [13] CIE Proceedings, Bureau Central de la CIE, Paris, 1964. [14] Pugh EN. Vision: physics and retinal physiology. In: Atkinson RC, editor. Stevens Handbook of Experimental Psychology. New York: Wiley, 1988:75 – 163. [15] Stockman A, MacLeod DIA, Johnson NE. Spectral sensitivities of the human cones. J Opt Soc Am 1993;A10:2491– 521. [16] Kim S, Bush RA, Sieving PA. Monkey photopic flicker ERG to sine wave and square wave stimuli: component identification using drug isolation. Invest Ophthalmol Vis Sci 1997; 38 Suppl:885 (abstract). [17] Mollon JD, Baker MR. The use of CRT displays in research on colour vision. In: Drum B, editor. Colour Vision Deficiencies XII. Dordrecht: Kluwer Academic, 1995:423 – 44. [18] Birch DG. Flicker electroretinography. In: Heckenlively JR, Arden GB, editors. Principles and Practice of Clinical Electrophysiology of Vision. St. Louis: Mosby Year Book, 1991:348–51. [19] Shapley RM, Victor JD. The effect of contrast on the transfer properties of cat retinal ganglion cells. J Physiol 1978;285:299– 310. [20] Lee BB, Pokorny J, Smith VC, Kremers J. Responses to pulses and sinusoids in macaque ganglion cells. Vis Res 1994;34:3081– 95. [21] Yeh T, Lee BB, Kremers J, Cowing JA, Hunt DM, Martin PR, Troy JB. Visual responses in the lateral geniculate nucleus of dichromatic and trichromatic marmosets (Callithrix jacchus). J Neurosci 1995;15:7892 – 904. [22] Kremers J, Weiss S, Zrenner E. Temporal properties of marmoset lateral geniculate cells. Vis Res 1997;37:2649 – 60. [23] Lutze M, Smith VC, Pokorny J. Critical flicker frequency in X-chromosome linked dichromats. Doc Ophthalmol Proc Ser 1989;52:69 – 77.
Response phase of the flicker electroretinogram (ERG) is influenced by cone excitation strength Tomoaki Usui a,b, Jan Kremers a,*, Lindsay T. Sharpe a, Eberhart Zrenner a a
Department of Experimental Ophthalmology, Uni6ersity Eye Hospital, Ro¨ntgenweg 11, 72076 Tu¨bingen, Germany b Department of Ophthalmology, Niigata Uni6ersity School of Medicine, Niigata, Japan Received 10 July 1997; received in revised form 27 July 1997; accepted 23 December 1997
Abstract We measured electroretinogram (ERG) response phases at different cone contrasts in trichromats and dichromats to investigate the dynamics of the long-wavelength-sensitive (L-) and middle-wavelength-sensitive (M-) cone pathways. ERG responses to stimuli, temporally modulated at 30 Hz, were recorded. The stimuli were generated on a computer controlled colour monitor. Thirty-two different combinations of L- and M-cone excitation strength, expressed as cone contrasts, were presented. The short-wavelength-sensitive (S-) cones were not stimulated (S-cone contrast =0%). The response phase of the fundamental stimulus component was obtained from Fourier analysis. The ERG response phase lags decreased with increasing cone contrast. This was observed in all subjects with a normal appearing fundus. In dichromats and trichromats at low and intermediate contrasts, the phase lags to M-cone isolating conditions were smaller than those to L-cone isolating stimuli. In one dichromat with extreme myopia and cupping of the optic disc, the ERG phase lags increased with increasing cone contrast. The ERG response phase may be potentially useful for detecting retinal abnormalities. © 1998 Elsevier Science Ltd. All rights reserved. Keywords: Electroretinogram (ERG); Response phase; Cone contrast; Dichromacy; Trichromacy
1. Introduction It is controversial whether the dynamics of the primate long-wavelength-sensitive (L-) and middle-wavelength-sensitive (M-) cones differ or not. The time to peak response and the integration time of single macaque M-cones recorded with suction electrodes [1] appear to be shorter than those of L-cones. Yet the differences are not significant. However, data from several other studies, involving additional contributions from post-receptoral mechanisms, suggest that the Mcone pathway responds faster than the L-cone pathway. The data include cone electroretinogram (ERG) recordings in monkey [2] and in human [3,4], psychophysical measurements in humans [5] and ganglion cell recordings in the macaque retina [6 – 8]. In this paper, we add new data about L- and M-cone dynamics obtained from flicker ERG recordings in human subjects.
* Corresponding author: Tel: +49 7071 2985031; fax: + 49 7071 295777; e-mail: [email protected]. 0042-6989/98/$19.00 © 1998 Elsevier Science Ltd. All rights reserved. PII: S0042-6989(98)00046-7
We modified the ‘silent substitution’ technique (see review in Ref. [9]) and the classic heterochromatic flicker photometric procedure (see review in Ref. [10], for the implementation in ERG recordings) so that we could stimulate the L- or M-cones in a predefined manner. This procedure has been used to discriminate protanopes from deuteranopes and to demonstrate individual variations in the ERG amplitudes of L- and M-cone signals in normal trichromats [11]. Here we report the dependence of the ERG phase on photoreceptor contrast in dichromats and trichromats. Results about the dependence of response amplitude on photoreceptor contrast are reported elsewhere [11]. Part of this study has been presented in abstract form [12].
2. Subjects Fifteen dichromats (all males; six protanopes, and nine deuteranopes;) and eight trichromats (six males and two females) took part in this study. The classifica-
T. Usui et al. / Vision Research 38 (1998) 3247–3251
3248
tion of dichromacy was based upon Rayleigh matches obtained on a Nagel type I anomaloscope (Schmidt and Haensch, Germany) and was consistent with molecular genetical analysis of blood samples (by J. Nathans, personal communication). Of the protanopes, four had multiple genes and two had single genes. All had the alanine amino acid at position 180. Of the deuteranopes, four had multiple genes and four had single genes; all had serine at position 180. Normal ophthalmological findings were evident in all subjects, except for one protanope (subject AO) who had high myopia ( − 8.0 diopter) bilaterally with a typical fundus appearance, including vascular narrowing and optic disc cupping. Informed consent was obtained from all subjects after explanation of the purpose of the study.
3. Methods The methods, including visual stimuli and recording techniques, have been described in detail elsewhere [11]. Briefly, the stimuli were presented on a computer controlled monitor (BARCO, frame rate: 100 Hz) with a VSG 2/2 graphics card (Cambridge Research System). The monitor subtended 124 ×108° at the 10 cm viewing distance. The red, green and blue phosphors of the monitor were temporally modulated at 30 Hz for predefined Michelson contrasts. Contrast was defined as 100%×
(Lmax −Lmin) (Lmax +Lmin)
where Lmax and Lmin are the maximum and minimum luminances, respectively. A power spectrum analysis revealed that the main frequency components in the stimulus occur at 0 Hz (the DC level), the 100 Hz frame rate and the 30 Hz modulation frequency. Frequency components arising from interference between the frame rate and the modulation frequency were very small in magnitude (i.e. less than 8% of the power at 30 Hz). The time averaged luminance of the monitor was 66 cd/m2 (40 cd/m2 for the green phosphor, 20 cd/m2 for the red phosphor and 6 cd/m2 for the blue phosphor). The time averaged chromaticity in ([13]) large field coordinates was: x = 0.33, y =0.32. The mean quantal catches were very similar for the L- and Mcones (assuming a pupil diameter of 8 mm and a foveal cone collecting area of 2.92 mm2 [14]): the mean foveal L-cone quantal catch was 4.42 log quanta s − 1 cone − 1 and the mean foveal M-cone quantal catch was 4.31 log quanta s − 1 cone − 1. The spectral characteristics of the monitor phosphors were measured with a spectroradiometer (Instrument Systems). The monitor calibration and the gamma correction were performed using the internal luminance device of the BARCO monitor and the VSG software. The sensitivities of each cone type to the phosphors were calculated by multiplying
the emission spectra with psychophysically-derived estimates of the corneal sensitivity spectra of the cones [15]. The sensitivities were then used to calculate the modulation contrasts of each phosphor in order to obtain the desired L- or M-cone contrast. S-cone contrast was 0% in all conditions. We measured the ERGphases in the dichromats for a series of contrasts, with a maximum L- or M-cone contrast of 78.7%. For trichromats, there were fewer conditions resulting in pure L- and pure M-cone modulation. The maximal M-cone contrast was 28.6% and the maximal L-cone contrast 22.9%. One eye was dilated with a mydriatic (0.5% tropicamide) and kept light-adapted at least 10 min before the ERG recordings began. Corneal ERG responses were measured with a DTL fiber electrode (UniMed Electrode Supplies). The reference and ground skin electrodes were attached to the ipsilateral temple and the forehead, respectively. Signals were amplified and band-pass filtered with 10 and 100 Hz corner frequencies. The signal was sampled at 1000 Hz with a CED 1401 on-line computer. Forty-eight ERG measurements, each lasting 1 s, were averaged. ERG responses were Fourier analysed and the ERG response amplitude and phase were defined as the amplitude and phase of the fundamental stimulus components. Power spectra of the responses revealed that, at the 30 Hz temporal frequency, the response is mainly determined by the fundamental component. It is difficult to speculate about the cellular origin of this component, but the b-wave is traceable up to 33 Hz and the a-wave only up to 20 Hz [16], suggesting that the direct origin is not in the photoreceptors themselves. Stimulus artefacts were measured with the monitor covered by black cardboard. Although the stimulus artefacts were normally very small, ERG signals were corrected by vector subtracting the stimulus artefact from the measured signal. More detailed information about correcting for stimulus artefact is given elsewhere [11]. The use of a CRT screen has inherent temporal limitations which in principle can interfere with ERG measurements [17]. We were, however, careful to verify that the response phase effects, described in the present paper, had in fact a physiological origin and were not determined by the properties of the CRT screen. One possible source of error is that the three different phosphors of the monitor have different times to maximal excitations and different decay times. The overall excitation and decay time of the phosphors is determined by the amount of excitations of the individual phosphors, which might influence the response phase. We excluded this possible source of error by measuring the response amplitudes and phases in dichromats using different stimulus conditions, in which the phosphor excitations differed but the cone contrasts did not. We
T. Usui et al. / Vision Research 38 (1998) 3247–3251
found that the response amplitudes and phases were not significantly different between these conditions, indicating that the response phase is indeed determined by the cone contrast and not by the excitation strength of the phosphors. A second possible source of error might arise from the sequential excitations of the individual monitor pixels. The pixels in the lower part of the monitor are excited later than pixels in the upper part of the monitor. This introduces a substantial phase shift at the 30 Hz stimulus frequency. Thus, any difference in the L- and M-cone distribution between the upper and lower part of the retina could produce a measured phase difference between the responses of the two cone types. However, in that case, it would be expected that the phase difference is reversed by rotating the monitor 180°. A control experiment with the monitor positioned upside down, however, did not result in a sign reversal of the phase differences.
3249
than the response amplitude to L-cone isolating stimuli at similar cone contrasts [11], resulting in less reliable M-cone ERG-phases. The fits through the data obtained in the dichromats are replotted in Fig. 1(B), enabling a comparison between the dichromatic and trichromatic data. Clearly, a similar relationship between response phase and L-cone contrast is found in deuteranopes and trichromats. The similarity between the response phases of protanopes and trichromats to M-cone stimuli is less satisfactory, probably because of the less reliable trichromatic data arising from the
4. Results In Fig. 1(A) the phases of the fundamental components of the ERG responses in the protanopes (filled circles) and deuteranopes (open circles) are plotted as a function of the M- or L-cone contrast. Since the Fourier analysis only gives phases within a 360° range, we estimated the absolute response phases at 30 Hz by comparing them with response phases at other temporal frequencies (unpublished data) and from explicit times which have been reported as ranging between 25 and 30 ms for 30 Hz flicker ERGs [18]. By definition, a negative response phase indicates a phase lag. The response phase increases with increasing contrast, indicating that the phase lag decreases. But a plateau is reached between 20 and 40% contrast. The relation between cone contrast and ERG phase can be described by an exponential function: P(C) = k1 + k2(1− e − k3 · C) where P(C) is the response phase, C is the cone contrast, and k1, k2 and k3 are constants. The function was separately fitted to the data points of the protanopes and deuteranopes using the solver routine in the Excel 4.0 for Windows program. The values for the constants in the fits are: k1 = −320.6°, k2 =43.6° and k3 = 0.09 for the protanopes; and k1 = −364.2°, k2 =73.0° and k3 = 0.07 for the deuteranopes. The scatter in the data is to a large extent caused by individual differences, which were highly reproducible. The response phases as a function of cone contrast in five trichromats are plotted in Fig. 1(B). There were fewer conditions for trichromats than for dichromats in which the response of a single cone type could be isolated. Further the response amplitudes in trichromats to M-cone isolating stimuli are normally smaller
Fig. 1. (A) Response phase of the fundamental ERG component (in degrees) as a function of cone contrast (in percent) for M-cones measured in protanopes (filled circles) and for L-cones measured in deuteranopes (open circles). The curves represent the best fitting exponential function (see equation in text; drawn curve for M-cones, dashed curved for L-cones). (B) Response phase as a function of cone contrast measured in trichromats to L- (open diamonds) and M-cone (closed diamonds) isolating stimuli. For trichromats, there are fewer conditions in which a single cone is selectively stimulated. Further, the maximal achievable cone contrast is smaller than in dichromats. The response amplitudes to M-cone isolating stimuli is generally smaller than those to L-cone isolating stimuli, resulting in less reliable M-cone response phases. The fits to the dichromatic data are replotted for comparison.
3250
T. Usui et al. / Vision Research 38 (1998) 3247–3251
5. Discussion
Fig. 2. Comparison of the ERG response phases determined from Mand L-cone isolating stimuli presented to trichromats, protanopes and deuteranopes. Cone contrast is between 20 and 30% for all subjects. Data points are averages of at least two separate measurements. The trichromatic data, obtained in the same individual, are connected by a line. The data from one protanope (AO) are not included in this figure.
smaller response amplitudes. As for the dichromats, highly reproducible individual differences in the absolute phases of the trichromats are observed. Fig. 2 presents a comparison of the phase data of all trichromats, protanopes and deuteranopes. Only data measured at M- and L-cone contrasts between 20 and 30% are included. These contrasts were large enough to obtain reliable responses, but low enough to show clear L- and M-cone ERG phase differences between the dichromats. The phases of M- and L-cone ERGs in the trichromats (which are connected by lines for the individual subjects) correspond to those in the dichromats, indicating that the receptoral and postreceptoral pathways are similar for the trichromats and the dichromats. The phase differences, therefore, probably reflect physiological differences between the pathways mediating the L- and M-cone ERGs. Interestingly, the phases in one protanope (AO) with high myopia are distinctly different. His phases, which were measured three times in separate sessions, decrease with increasing cone contrast (Fig. 3; same scale as Fig. 1). Especially at the high M-cone contrasts, where large response amplitudes can be obtained, the phases are highly reproducible. The constants for the fit are: k1 = − 284.6°, k2 = − 81.6° and k3 =0.02. Note that k2 has a negative value.
The ERG phases increase with increasing cone contrast up until contrasts of 20–40% (see Fig. 1). It is not clear what mechanism is responsible for the phase shifts, but they clearly signify a nonlinearity in the ERG pathway. In retinal ganglion cells of the cat [19], macaque [20] and marmoset [21] and in LGN cells of the marmoset [22] similar phase shifts have been reported and described as a contrast gain control mechanism. Possibly, similar processes play a role in the ERG nonlinearities. But it is unlikely that the ERG nonlinearities have the same cellular origins as the nonlinearities in the single cell responses, because the phase shifts in the single cell recordings occur at frequencies between 4 and 15 Hz. The 30 Hz stimulus used in our experiments is clearly outside this range. Further, the contrast gain control found in the retinal ganglion cells occurs in combination with response saturation; whereas the ERG signals do not saturate. At low and intermediate contrasts, the M-cone ERG phases (measured in protanopes and trichromats) are larger than the L-cone ERG phases (measured in deuteranopes and trichromats) by about 40° (Fig. 2), indicating that M-cone signals, measured with the ERG, are transmitted about 3.7 ms faster than L-cone signals. At high contrasts, however, the phase differences are smaller (about 10°). Phase data obtained at other temporal frequencies are required to quantify the delay difference accurately, and to establish whether a
Fig. 3. Phase of the fundamental components of the ERG response (in degrees) as a function of M-cone contrast (in percent) in the protanope (AO) with high myopia. The measurements were repeated four times at each cone contrast (filled circles). The curve represents the best fitting exponential function. Note that the constant k2, for the exponential fit, has a negative value (see text for details).
T. Usui et al. / Vision Research 38 (1998) 3247–3251
delay difference suffices to explain the phase differences. We found phase differences between 10 and 105° in individual trichromats (Fig. 2), suggesting delay differences ranging between 1.0 and 9.7 ms. But, the phases of the M-cone ERGs are less reliable than those of the L-cone ERGs in the trichromats. This will necessarily influence the variability in phase difference. Whitmore and Bowmaker also reported that M-cone ERG signals are phase advanced relative to L-cone ERG signals [4]. But the difference they obtained (12 ms) was much larger than the values reported here. Yeh et al. reported that parvocellular projecting ganglion cells with M-cone input to the centres responded about 3.8 ms faster than cells with L-cone input [8]. Although it is interesting to speculate that the latency differences result wholly or mainly from differences in the M- and L-cones kinetics, differences in the post-receptoral mechanisms cannot be ignored. In recordings from single cones to short flashes, Schnapf et al. found an average time to peak response of about 51 ms in M-cones and 55 ms in L-cones, although this difference was not statistically significant [1]. However, the psychophysical data of Hamer and Tyler [5] also suggest that the pathways subserving the M-cones respond faster than those subserving the Lcones. On the other hand, Lutze et al. [23] reported that deuteranopes, for whom the L-cones are mediating detection, have a higher critical flicker fusion than protanopes, for whom the M-cones are mediating detection. This is seemingly in contradiction with our results. But, a change in delay between L- and M-cone responses does not necessarily change the critical flicker fusion. Differences between L- and M-cones in critical flicker fusion frequency, whether measured in normal trichromats or dichromats, might involve other mechanisms than a simple time delay. Interestingly, in contrast to all other dichromatic subjects, one protanope showed decreasing ERG phase with increasing cone contrast. His high myopia with typical fundus change might have contributed to this unusual phase behaviour. Further studies are required to establish whether such ERG phase anomalies can be used in the diagnosis of retinal disorders.
Acknowledgements The authors wish to thank Sabine Meierkord and Eva Burkhardt for technical assistance, and Dr Stefan Weiss for computer programming assistance. This study was supported by fortu¨ne-Grant F.1222147.2 (Tu¨bingen, Germany) and Deutsche Akademisches Austauschdienst (Bonn, Germany) to TU, DFG grant Zr 1/9-3 (Bonn, Germany) to JK and EZ, and a Hermann-undLilly-Schilling-Stiftung fellowship to LTS.
3251
References [1] Schnapf JL, Nunn BJ, Meister M, Baylor DA. Visual transduction in cones of the monkey Macaca fascicularis. J Physiol 1990;427:681 – 713. [2] Yamamoto S, Gouras P, Lopez R. The focal cone electroretinogram. Vis Res 1995;35:1641 – 9. [3] Spileers W, Falcao-Reis F, Hogg C, Arden GB. Evidence from human electroretinogram a and off responses that color processing occurs in the cones. Invest Ophthalmol Vis Sci 1993;34:2079– 91. [4] Whitmore AV, Bowmaker JK. Differences in the temporal properties of human longwave- and middlewave-sensitive cones. Eur J Neurosci 1995;7:1420 – 3. [5] Hamer RD, Tyler CW. Analysis of visual modulation sensitivity. V. Faster visual response for G- than for R-cone pathway? J Opt Soc Am 1992;A9:1889 – 904. [6] Smith VC, Lee BB, Pokorny J, Martin PR, Valberg A. Responses of macaque ganglion cells to the relative phase of heterochromatically modulated lights. J Physiol 1992;458:191 – 221. [7] Lee BB, Martin PR, Valberg A, Kremers J. Physiological mechanisms underlying psychophysical sensitivity to combined luminance and chromatic modulation. J Opt Soc Am 1993;A10:1403– 12. [8] Yeh T, Lee BB, Kremers J. Temporal responses of ganglion cells of the macaque retina to cone-specific modulation. J Opt Soc Am 1995;A12:456 – 64. [9] Este´vez O, Spekreijse H. The ‘silent substitution’ method in visual research. Vis Res 1982;22:681 – 91. [10] Jacobs GH, Neitz J, Krogh K. Electroretinogram flicker photometry and its applications. J Opt Soc Am 1996;A13:641–8. [11] Usui T, Kremers J, Sharpe LT, Zrenner E. Flicker cone ERG in dichromats and trichromats. Vis Res 1998;38:3391 – 6. [12] Usui T, Kremers J, Zrenner E. Phase differences between L- and M-cone ERG. Invest Ophthalmol Vis Sci 1997;38 Suppl:885 (abstract). [13] CIE Proceedings, Bureau Central de la CIE, Paris, 1964. [14] Pugh EN. Vision: physics and retinal physiology. In: Atkinson RC, editor. Stevens Handbook of Experimental Psychology. New York: Wiley, 1988:75 – 163. [15] Stockman A, MacLeod DIA, Johnson NE. Spectral sensitivities of the human cones. J Opt Soc Am 1993;A10:2491– 521. [16] Kim S, Bush RA, Sieving PA. Monkey photopic flicker ERG to sine wave and square wave stimuli: component identification using drug isolation. Invest Ophthalmol Vis Sci 1997; 38 Suppl:885 (abstract). [17] Mollon JD, Baker MR. The use of CRT displays in research on colour vision. In: Drum B, editor. Colour Vision Deficiencies XII. Dordrecht: Kluwer Academic, 1995:423 – 44. [18] Birch DG. Flicker electroretinography. In: Heckenlively JR, Arden GB, editors. Principles and Practice of Clinical Electrophysiology of Vision. St. Louis: Mosby Year Book, 1991:348–51. [19] Shapley RM, Victor JD. The effect of contrast on the transfer properties of cat retinal ganglion cells. J Physiol 1978;285:299– 310. [20] Lee BB, Pokorny J, Smith VC, Kremers J. Responses to pulses and sinusoids in macaque ganglion cells. Vis Res 1994;34:3081– 95. [21] Yeh T, Lee BB, Kremers J, Cowing JA, Hunt DM, Martin PR, Troy JB. Visual responses in the lateral geniculate nucleus of dichromatic and trichromatic marmosets (Callithrix jacchus). J Neurosci 1995;15:7892 – 904. [22] Kremers J, Weiss S, Zrenner E. Temporal properties of marmoset lateral geniculate cells. Vis Res 1997;37:2649 – 60. [23] Lutze M, Smith VC, Pokorny J. Critical flicker frequency in X-chromosome linked dichromats. Doc Ophthalmol Proc Ser 1989;52:69 – 77.