Psychophysical Influences on the Validity of Anomaloscopic Assessments of Color Vision

Article excerpt

Anomaloscopes are used in clinical and research applications involving the assessment of color vision. Output data include the matching range (MR), the midpoint, and the anomaly quotient (AQ). The latter is commonly used to compare data obtained using different instruments. However, the midpoint and AQ ultimately depend on the MR, for which there is no universal operational definition. In this study, 5 volunteers with normal color vision each completed 510 trials in a rating task with an anomaloscope employing the Moreland equation. The aim was to investigate the nature of the perceptual transition into, and out of, a metameric match. This was found to be less than abrupt, providing ample opportunity for the influence of individual response biases. Consequently, an agreed-upon definition of the MR, associated with an appropriate psychophysical technique, would facilitate the identification and tracking over time of acquired color vision deficits and would improve the validity of interstudy comparisons.

A number of anomaloscopes are commercially available for clinical and research applications in which the assessment of color vision is required. Some instruments utilize only the red-green Rayleigh equation, whereas others also employ the blue-green Moreland equation. These instruments facilitate the diagnosis of congenital color vision deficits; however, the extent to which they allow valid comparisons among studies in which different instruments are used, or provide sufficient accuracy to track subtle changes in color vision occurring over time, is less certain. Nousiainen, Kälviäinen, and Mäntyjärvi (2000) reported that, in their investigation of color vision deficiencies associated with treatment for epilepsy, an anomaloscope (Color Vision Meter 712) did not seem efficient in detecting patients with acquired color vision abnormalities. Even if it is assumed that different instruments employ comparable optics and instantiate the appropriate equations equivalently, the reported outcomes may be affected by the variety of psychophysical procedures and operational definitions used to determine the parameters of interest. The output data generally include the matching range (MR), the midpoint, and the anomaly quotient (AQ). The AQ divides the ratio of the fixed and variable wavelengths for the current observer by that obtained for a normative group at the midpoint of the MR (National Academy of Sciences [NAS], 1981). This measure is commonly used to compare data obtained from studies in which different instruments have been used. However, the midpoint and, therefore, the AQ ultimately depend on the MR, which, for the reasons mentioned above, may vary not only between instruments, but also between sessions with the same observer.

Concerns of this nature arose during an investigation of color vision and contrast sensitivity in dental professionals, where it was considered necessary to examine the validity of the range of blue-green metameric matches reported by an anomaloscope utilizing the Moreland equation (Stillman, 2007). From the observer's point of view, the display, created with different light mixtures on either side of a circular field, is experienced either as a single-hued circle or as possessing different hues on either side of center. With the Roland anomaloscope, in an automatic diagnostic mode, stimuli on each trial are chosen by the software and are presented to the observers for simple same or different judgments, and in other modes (manual or dynamic), manual adjustments are made by the researcher or the observer to establish a match. The method of adjustment is commonly used by other instruments as well-for example, the Nagel anomaloscope (see Squire, Rodriguez-Carmona, Evans, & Barbur, 2005) and the Oculus Heidelberg-Multi-Color (HMC) anomaloscope. The procedure and scoring with the Color Vision Meter 712 anomaloscope (Interzeag, Schlieren, Switzerland) and the Pickford-Nicholson anomaloscope (NAS, 1981) are more sophisticated. …