Graefe’s Arch Clin Exp Ophthalmol (2005) 243:26–32 CLINICAL INVESTIGATION DOI 10.1007/s00417-004-0893-z Herbert Jägle Markus Pirzer Lindsay T. Sharpe Received: 18 December 2003 Revised: 3 February 2004 Accepted: 11 February 2004 Published online: 31 July 2004
Springer-Verlag 2004 H. Jgle ()) · M. Pirzer Department of Pathophysiology of Vision & Neuro-Ophthalmology, University Eye Hospital, Schleichstrasse 12–16, 72076 Tbingen, Germany e-mail: herbert.jaegle@uni-tuebingen.de Tel.: +49-7071-2980744 Fax: +49-7071-294678 L. T. Sharpe Department of Psychology, University of Newcastle upon Tyne, Newcastle upon Tyne, UK The Nagel anomaloscope: its calibration and recommendations for diagnosis and research Abstract Background: The Nagel anomaloscope Model I is the definitive clinical instrument for classifying phenotypic variations in X-linked color-vision disorders. Its system of classification is based on the Rayleigh equation: the relative amounts of red and green primary lights required to match a yellow primary. Our aim was to characterize how changes in mains voltage and ambient temperature influence the wavelength and intensity of each primary and alter the Rayleigh matches of normal and anomalous trichromats. Methods: A Nagel Model I anomaloscope was calibrated in wavelength and intensity while varying the temperature of its prism housing and the mains voltage. Three normal, three protanomalous and three deuteranomalous trichromats made Rayleigh matches at various temperatures and voltages. Results: The intensities of the green and red primaries show an Introduction The Nagel anomaloscope Model I, manufactured by Schmidt & Haensch to a standard design until 1983, is still the definitive clinical instrument for classifying the phenotypic variation in X-chromosome linked color vision disorders [12]. Optically similar to an instrument first described by Nagel [9], it is used to determine the Rayleigh equation; specifically, the match of a spectral yellow light or primary to a mixture of spectral red and green lights (the R-G match). It consists of a light source, a separate internal adaptation light, an ocular, a compound exponential growth with mains voltage. Additionally, the wavelengths and intensities of all three primaries change with prism housing temperature. As a result, the R-G match midpoints of normal and anomalous trichromats shift with increasing mains voltage, and more markedly with increasing prism housing temperature, to higher R-G settings. Conclusions: Rayleigh matches obtained with the Nagel I anomaloscope are sensitive to changes in voltage supply and prism housing temperature, arising largely from thermal effects of the internal light sources. However, the instrument may still be safely used for diagnostic and research purposes provided that: (1) a stable voltage supply is used; (2) it is kept at a constant temperature; and (3) the match midpoint of the reference population has been established under identical conditions. direct-vision prism and three entrance slits (see Fig. 1). The entrance slits are carefully calibrated to define the red, green and yellow primaries in terms of wavelength and intensity, according to a norm (DIN 6160) of the German Institute for Standardization [4]. However, the accuracy of its calibrated settings is influenced by several environmental factors, including temperature changes and fluctuations in mains voltage, which have never been fully investigated. 27 Fig. 1 Schematic diagram of the Nagel Model I anomaloscope used in this study. The diagram has been modified from the original, which appears in the manufacturer’s handbook Mains voltage It has long been known that the stability and reliability of Nagel anomalocope matches are influenced by the mains voltage [11]. Cavonius [2] found an increase of the R-G match midpoints of normal trichromats with the mains voltage of ca. 0.05 (Nagel) units per volt and concluded that it is unlikely that these variations lead to clinical misdiagnosis. However, although the nominal mains voltage was raised in 1983 from 220 V to 230 V in Germany, no adjustment was made in the design or operation of the instrument, which was constructed to run at 220 V. Further, for reasons of harmonization within the European Union, the permitted voltage range will be further extended to 195.5–253 V in 2004 [5, 6]. These changes could potentially influence the Rayleigh matches and diagnosis of anomalous trichromats. Temperature Richter [10] was the first to report a seasonal variation in the amount of red and green required to match the yellow primary on the Nagel anomaloscope. Jordan & Mollon [8] replicated his results on three anomaloscopes, including two different Model I instruments (Schmidt & Haensch) and an antique Model II instrument (manufactured by Spindler & Hoyer). Expressing the matches in terms of the anomalous quotient, a normalizing technique introduced by Trendelenburg [13] to compensate for minor changes in line voltage and bulb aging and to allow comparisons between different instruments, they concluded that the variation was almost certainly owing to an instrumental sensitivity to ambient temperature—in particular, thermal changes in the prism—rather than to a seasonal change in human physiology [8]. They estimated the magnitude of the increase of the Rayleigh match midpoint per degree Celsius of ambient room temperature to be 0.175 Nagel units. However, they did not explore the relevant sources of instrument temperature variation in detail. These include not only thermal fluctuations in external air temperature, but also changes in the heat being emitted by the internal incandescent source (see Fig. 1), located near the prism in the same housing and used to provide a neutral adaptation light. The contribution of both of these temperature effects to the R-G matches of normal and anomalous trichromats is yet unknown. Anomalous trichromacy The classification of anomalous trichromacy is based on the observation that protanomalous and deuteranomalous observers require more of the red (protanomaly) or more of the green (deuteranomaly) primary in the red–green mixture to match the yellow primary than do normal observers. The amount of red or green shift is a measure of the severity of the anomaly. Given that the relative amounts of red and green required in the match depend upon the wavelength and intensity of each primary, and that these, in turn, critically depend upon the lamp voltage and the temperature of the instrument’s prism housing, it is possible that voltage and temperature changes may have a significant effect upon phenotype–genotype correlations of anomalous red–green color vision. Therefore, the principal aim of this study was to characterize the wavelength and intensity changes of the Nagel Model I anomaloscope primaries, arising from mains voltage and temperature changes, and to estimate their effects on R-G matches in normals and anomalous trichromats. 28 Methods Subjects Nine subjects were selected from a population of normal and anomalous trichromats who were previously diagnosed by their anomaloscope matches on the Rayleigh Model I instrument and on the Oculus anomaloscope. They were classified as normal (MB, JA, TG), deuteranomalous (RS, AS, SJ) or protanomalous trichromats (TT, TZ, MP). The study was conducted in accordance with the tenets of the Declaration of Helsinki and with the approval of the ethics committee on human experimentation of the University of Tbingen. Apparatus and calibration The Rayleigh equation settings were made with a Schmidt & Haensch Nagel anomaloscope Model I constructed in 1979. All physical (calibration) data were confirmed with a second instrument constructed in 1980. Both instruments are equipped with a 220-V, 100-W cine-projection lamp for producing the red, green and yellow primaries and a 220-V, 60-W incandescent bulb for providing neutral adaptation. The instruments were connected to the mains power supply through a variable transformer so that their voltages could be adjusted between 100 and 250 V or kept constant at 230 V. All voltages were controlled by a digital meter and manually readjusted, resulting in residual fluctuations of less than €0.5 V. The emission spectra of the red, green and yellow primaries were measured with a compact array spectroradiometer (CAS 140, Instrument Systems, Munich, Germany). The half-field intensities—one half-field produces the yellow primary light, the other, the mixture of the red and green primary lights—were measured over the full range of R/G settings with a Pin-10 diode (United Detector Technology, Santa Monica, CA) and a calibrated radiometer (Model 80X Optometer, United Detector Technology). To control the effect of the temperature on the R/G settings a thermal sensor was mounted directly inside the prism housing, and a second sensor was positioned 1 m away from the instrument on the measuring table to monitor the ambient room temperature. To reduce the instrument temperature below room temperature, the prism housing was slowly cooled down to 15C by the appropriate placing of Coolpacks (3 M “Nexcare” Coldhot Pack, 3 M Health Care, D-41453 Neuss, Germany). Lower temperatures were avoided to prevent damage from water condensation within the prism housing. found with the R/G half-field. In the event that the half-fields were seen as identical in color and intensity, the R/G setting and the actual prism housing temperature were recorded. This procedure was repeated at higher prism housing temperatures, ranging from 16C to 39C. Results Physical data The physical calibration data, obtained at a prism housing temperature of approximately 36C (see Methods), reveal that the wavelengths of the green, red and yellow primaries do not change with mains voltage. Further, at any given temperature, the Nagel anomaloscope conforms to the German DIN 6160, insofar as the calibrated primary wavelengths and luminances fall within the standard, accepted range of values. Nevertheless, the primary wavelengths as well as the luminances vary systematically with mains voltage and prism housing temperature so as to cause definable shifts in the Rayleigh matches of both normal and anomalous trichromats. Voltage data The intensities (calculated for a 2-deg visual field) of the green and red primaries (Fig. 2) grow approximately exponentially with the mains voltage. The significantly higher increase in green primary intensity reflects the color temperature change of the light source from 1821.5 K, at the lowest calibrated mains voltage of 100 V, to 2292.2 K and 2512.9 K, at mains voltages of 196 V and 253 V, respectively (corresponding to the range permitted in DIN EN 50160). Procedure Anomaloscope matches were determined in the preferred eye of each subject. To avoid learning effects, all subjects were trained in performing Rayleigh matches prior to the first day of experimentation. In the mains voltage experiments, only normal trichromat matches were made. They were averages of at least five settings at each of 14 voltages from 120 V to 250 V. The order of voltages was randomly chosen by the investigator. During the experiments, the neutral adaptation light was turned on and the prism housing temperature was maintained between 34C and 36.6C, with lower temperatures corresponding to lower mains voltage. In the temperature variation experiments, all matches were performed at a constant mains voltage of 230 V. All experiments were conducted with the internal neutral adaptation light switched on. To determine the full range of acceptable matches the investigator presented a series of R/G settings starting with the lowest temperature. The subject was then allowed to adjust the brightness of the yellow half-field, until an acceptable match was found or not Fig. 2 The exponential growth of the red and green primary intensities, in retinal illuminance (calculated for a two degree visual field) with increasing mains voltage 29 Fig. 4 Dependency of the centroid wavelengths of the red and green primaries upon the prism housing temperature. The wavelength of the green primary shifts ca. 0.6 nm to longer wavelengths per 10C temperature difference; that of the red primary, ca. 1.0 nm Fig. 3 Time-course of the change in room and prism housing temperature, with the internal neutral adaptation light turned off (A) or on (B) respectively, at 25.3C to 550.6 nm and 668.6 nm at 39.3C. These spectral shifts of 0.9 nm for the green and 1.6 nm for the red primary are clearly visible to both normals and anomalous trichromats. Temperature data Rayleigh matches The wavelengths of all three primaries change with prism housing temperature. Two operational modes have to be distinguished. When the internal neutral adaptation light is switched off, the temperature stays approximately constant at 24.5C for 15 min, then increases slowly until it reaches its final temperature of 29C (Fig. 3A). Alternatively, when the internal neutral adaptation light is switched on, after an initial 10-min period of nearly constant temperature, the prism housing temperature increases by about 12C, reaching its final temperature of approximately 36.4C after about 90 min (Fig. 3B). Because the maximum temperature also depends on the ambient room temperature, it may reach even higher values. These data clearly show that the major source of prism temperature variation is thermal variation (warm-up effects) of the neutral adaptation light. The temperature-dependent changes have two potential consequences for Rayleigh matches. First, the position and width of the entrance slits defining the red and green primaries (see Fig. 1) may be altered. Thus, the wavelengths of the primaries and the relative proportions of the green and red primaries in the Rayleigh match may change. Second, the refractive index of the prism and thus the wavelengths of the green and red primaries may be altered. As expected, our data (Fig. 4) show a shift of the peak wavelength with increasing temperature from 549.7 nm and 667.0 nm, for the green and red primaries, Voltage data As a result of the changes in the retinal illuminances of the three primaries, the R-G match midpoints of all three normal trichromats show a shift with increasing mains voltage to higher R-G settings (i.e. requiring less green and more of the red primary, see Fig. 5). Cavonius [2] showed similar results obtained from two normal trichromats and assumed a nearly linear dependency at higher main voltage. We find, however, an approximately exponential rise within the range from 190 V to 250 V. From our data, we estimate the match midpoint of normal trichromats to vary between 45 and 48 with a change of the mains voltage from 190 V to 250 V. At the given temperature of 36C, all values for the match midpoints within this mains voltage range are higher than those permitted by the DIN 6160 (36.5–43.8 units). Temperature data Figure 6 reveals, for two normal trichromats, three deuteranomalous and three protanomalous observers, how R-G mixture settings on a Nagel Model I anomaloscope vary with prism housing temperature. For each subject, a series of matches were obtained within temperature ranges centered at approximately 18C, 25C and 36C. Only 30 Fig. 5 The variation in R-G match midpoints for three normal trichromats with increasing mains voltage. The solid lines represent the best-fitting function (exponential rise to maximum) to the individual data Fig. 6 Rayleigh-matches of two normal, three deuteranomalous and three protanomalous observers on a Nagel anomaloscope as a function of prism housing temperature. The continuous and dashed curves represent the best-fitting linear regressions. The anomalous quotient shown is based upon a normal match midpoint of 44 corresponding to a temperature of approximately 25C the Rayleigh match determinations made at the two lower temperatures by normal observers fall within the accepted DIN 6160 ranges. In the case of the anomalous trichromats, the width of the matching range could be roughly estimated (see Table 1). For all six, only a small range variation, between ca. 1.5 and 3.0 Nagel units, was found, which did not significantly change between low (18C) and high (36C) temperature. (For the deuteranomalous subjects at 25C, no range could be estimated because only a few matches were obtained.) The anomalous quotients (AQ) have been calculated using the average match midpoint of the normal trichromats obtained at the center of the given temperature ranges. Based on these instrument- and temperature-specific match midpoints, only small changes of the AQ are observed for the three temperature ranges. How- ever, usually the AQ is calculated using a fixed normal match midpoint of 40 Nagel units, as suggested by Zrenner [14]. If this is done, a significantly larger shift of the anomalous quotient (AQ40) is found. For deuteranomalous observers, the AQ worsens (i.e., shifts away from the normal value) with lower temperature; whereas for the protanomalous observers, it improves (i.e., shifts towards the normal value). Table 1 Rayleigh-matches of normal and anomalous trichromats made at prism housing temperatures of the Nagel Model I anomaloscope within one of three ranges centered at 18C, 25C and 36C. Listed are the match midpoints (MMPs), ranges and anomalous quotients (AQs) calculated for each subject based on a normal Subject Eye Normal MB OD JA OS Protanomal TT OD TZ OS MP OD Deuteranomal RS OD AS OD SJ OS Nagel 18C MMP estimated from the average normal matches at each temperature. For comparison, the anomalous quotient for the highest temperature range (36C), calculated for a normal MMP of 40, is given as AQ40 Nagel 25C Nagel 36C MMP Range AQ MMP Range AQ MMP Range AQ AQ40 41.20 40.83 0.5 0.5 0.99 1.01 44.40 45.21 0.5 1.0 1.02 0.98 46.43 47.60 1.0 0.8 1.04 0.97 0.69 0.65 63.25 62.75 58.50 1.5 2.5 2.0 0.16 0.16 0.28 63.75 64.75 63.00 1.5 1.5 2.0 0.20 0.17 0.22 67.00 67.25 65.00 2.0 2.5 2.0 0.13 0.10 0.19 0.09 0.07 0.13 24.00 22.75 23.75 3.0 2.5 2.5 2.79 2.97 2.79 27.50 26.50 27.75 – – – 2.71 2.87 2.71 29.75 30.50 33.00 2.5 3.0 3.0 2.91 2.75 2.32 1.95 1.84 1.55 31 Discussion We find a significant shift in Rayleigh match midpoints in normal and anomalous trichromats with mains voltage as well as with prism housing temperature, which should be taken into account when establishing normal match midpoints and determining anomalous quotients for redgreen color blind observers. Main sources of voltage fluctuations in modern environments are elevators and generators connected to the same circuit. In regions with heavy industry or far from the energy source, the average voltage may be smaller than in our laboratory, in which a mains voltage range of 221 to 237 Volt was measured. However, even the small voltage changes measured in our laboratory result in intensity changes in the Nagel primaries that are clearly visible to a sensitive observer and correspond to a shift of ca. 0.5 Nagel units. These shortcomings could easily be avoided by using a stabilized power supply. Furthermore our results show a correlation between the R–G mixture required to match the yellow primary and the prism housing temperature. The shift is towards the red primary (i.e., higher values) with increasing temperature and is estimated as ca. 0.5 Nagel units per degree Celsius. The total shift may be up to 10 Nagel units with significant seasonal variations in room and instrumental temperature. In terms of Nagel units, compared with a normal match midpoint of 40, deuteranomalous observers improve (i.e., shift towards the normal value) with higher temperature; whereas protanomalous observers worsen (i.e., shift away from the normal value). Our data also indicate that the original recommendation of Trendelenburg [13] to use a anomalous quotient to reduce variability between instruments is still satisfactory from a clinical point of view. However, the unavoidable temperature-dependent shifts make it impossible to have normative data of the match midpoint available for all the temperatures within the possible range of prism housing temperature. Additionally, the temperature reached after ca. 90 min (steady state) will also vary significantly with ambient room temperature from the lowest value in winter (ca. 18–20C) to the highest value in summer (ca. 36–40C), in rooms that are not climate controlled. The prism housing temperature will be far less variable if the built-in neutral adaptation light is switched off and a separate light source is used for neutral adaptation. Although changes in wavelength and intensity of the primaries are mainly responsible for the shift of the match midpoints, other physiological factors may contribute as well. A reduction of the retinal illumination of the stimulus, in addition to direct voltage and prism temperature related changes, may affect the optical density of the cone pigments and the R–G mixture required to match the yellow primary. The size of such effects may vary among subjects and may be genetically influenced [12]. Given the dependency of the Rayleigh match upon prism temperature and mains voltage, it will have to be determined in further studies whether the strict limits for anomalous quotients set by regulatory authorities are fair, consistent and reliable when using the Nagel Model I and other mechano-optical anomaloscopes. Conclusions The severity of congenital red-green color-vision defects is routinely determined using Rayleigh matches on traditional mechano-optical anomaloscopes. The most widely used instrument is the Nagel Model I anomaloscope. The matches obtained with this instrument are sensitive to changes in voltage supply and prism housing temperature, arising largely from thermal effects produced by the internal light sources. Additionally, these instruments were originally designed for a voltage of 220 V and have not been modified, even though the standard voltage in Germany and Europe has been 230 V since 1983. While this small voltage change results in a shift of the match midpoint of normal subjects by ca. 1 Nagel unit, a difference in prism housing temperature of 20C—as occurs under the usual conditions of clinical investigations—results in a shift of the match midpoint by ca. 10 Nagel units in the direction of the red primary. Modern electronic anomaloscopes, which are not yet as widely used as the Nagel Model I anomaloscope, overcome these problems by using stabilized power supplies and color LEDs as sources for the three primaries. Its shortcomings notwithstanding, the Nagel Model I anomaloscope may still be safely used for diagnostic and research purposes, especially those involving the determination of the AQ in anomalous trichromats, provided that: (1) a stable voltage supply is used (a modern UPS for a single personal computer may be sufficient); (2) the instrument is kept at a constant temperature (e.g., the power is turned on at least 60 min prior to an investigation); and (3) the match midpoint of the reference population has been established under identical conditions. However, it will not fail to diagnose inherited red-green color-vision defects when operated at any of the temperatures or mains voltages within the ranges investigated in this study. Nevertheless, special care has to be taken if an expert opinion about the eligibility of a patient’s color vision for certain occupations (including driver’s and pilot’s licenses) is based on Rayleigh matches obtained with such anomaloscopes. Acknowledgements This work was supported by a grant from the Deutsche Forschungsgemeinschaft SFB 430/A6 awarded to Herbert Jgle. 32 References 1. Birch J (1993) Classification of anomalous trichromatism with the Nagel anomaloscope. In: Drum B (ed) Colour vision deficiencies XI. Kluwer Academic Press, Netherlands, pp 19–24 2. Cavonius CR (1995) Effect of lamp voltage on Nagel anomaloscope settings. In: Drum B (ed) Colour vision deficiencies XII. Kluwer Academic Press, Netherlands, pp 489–494 3. Cole BL (1993) Does defective colour vision really matter? In: Drum B (ed) Colour vision deficiencies XI. Kluwer Academic Press, Netherlands, pp 67–86 4. DIN 6160. Anomaloskope zur Diagnose von Rot-Grn-Farbenfehlsichtigkeiten. Ausgabe 1996 5. DIN EN 50160. Merkmale der Spannung in ffentlichen Elektrizittsversorgungsnetzen. Ausgabe 2000-03 6. EN 50160. Voltage characteristics of electricity supplied by public distribution systems. Version 1999-01-01 7. Gramberg-Danielsen B (1976) Rechtssicherheit bei der Bestimmung des Anomalquotienten? Klin Monatsbl Augenheilkd 168:429–433 8. Jordan G, Mollon JD (1993) The Nagel anomaloscope and a seasonal variation of colour vision. Nature 363:546–549 9. Nagel W (1907) Zwei Apparate fr die augenrztliche Funktionsprfung. Z Augenheilkd 17:201–222 10. Richter M (1948) Vom gegenwrtigen Stand der Farbenlehre. Z Wiss Photogr 43:209–237 11. Schmidt I (1955) Some problems related to testing color vision on the Nagel anomaloscope. J Opt Soc Am 45:514–521 12. Sharpe LT, Stockman A, Jgle H, Nathans J (1999) Opsin genes, cone photopigments and colorblindness. In: Gegenfurtner K, Sharpe LT (eds) Color vision: from genes to perception. Cambridge University Press, Cambridge, pp 3–52 13. Trendelenburg W (1929) Zur Diagnostik des abnormen Farbensinnes. Klin Monatsbl Augenheilkd 83:721–732 14. Zrenner E (1984) A proposal for the documentation of clinically applied color vision tests. Dev Ophthalmol 9:157–170