Metropsis: macular pigment optical density

The test makes a psychophysical estimate of macular pigment optical density (MPOD) at different eccentricities across the retina, using a CRT monitor.

The module incorporates a number of significant improvements over other methods of subjective MPOD estimation. The system incorporates three key innovations:

• The test presentation system incorporates a unique optical filter that significantly improves the accuracy and reliability of the pigment density estimate by eliminating CRT phosphor problems that normally confound MPOD estimates.
• An undemanding test stimulus radically improves test performance and measurement repeatability especially with naive subjects.
• Integrated eye tracking ensures that the correct retinal locations are tested.

This test was first demostrated at the ARVO meeting (the Association for Research in Vision and Ophthalmology) in Fort Lauderdale, Florida, May 1-5 2005.

Links to Macular Pigment Research Topics

The Metropsis MPOD test was presented at the July 2005 ICVS (International Colour Vision Society) meeting in a talk entitled 'An innovative instrument for the psychophysical measurement of Macular Pigment Optical Density using a CRT display' by P. West, Cambridge Research Systems Ltd, UK and J. Mellerio, London, UK.

Eat Your Greens - Mother Knows Best or Using Visual Illusions to Measure Macular Pigment.

Luminance matching

Psychophysical techniques for estimating MPOD use two luminance matches of monochromatic or narrow band lights, one of short wavelength that is absorbed by the MP, and one of long wavelength that is not. The initial luminance match is made at a location that is sufficiently eccentric for it to be assumed that there is no MP present. This serves as a reference for subsequent matches. Luminance matches are then made at retinal locations across the macula. These values are then compared to the reference match and differences are assumed to be due to the differential absorbance of the test lights by the MP. The density is expressed as the log of the ratio of the differences:

Ef_λs is short wavelength luminance match at the a central location within the macula.
Ef_λl is long wavelength luminance match at the a central location within the macula.
Ep_λs is short wavelength luminance match at the a peripheral location outside the macula.
Ep_λl is the short wavelength luminance match at the a peripheral location outside the macula.

Test presentation

The CRT monitor is an ideal way to display a structured test stimulus, for example, to easily test a series of points across the retina. A number of studies have been conducted in this way, using a CRT for MPOD measurements. However there are severe limitations in using a conventional CRT in this application and the measured MPOD values are a fraction of the true value. Whilst in theory this can be corrected for using model based predictions, the extreme under estimate makes correction prone to error due to signal to noise issues. The problem arises from the broad spectral output of the CRT phosphors. Whilst perceptually distinct and clearly red, blue and green there is substantial overlap between their spectra. The problem is further confounded by the absorption spectra of the macular pigment which shows that it absorbs light from all three phosphors, even to some extent, the red.

The Metropsis MPOD module employs a new and novel technique to largely overcome these short comings. In theory two narrow-band interference filters could be used that pass one wavelength that is absorbed by the MP and one that is not. However, this is not a practical solution, the amount of light from the broad-band phosphors that is absorbed by such filters means that the test stimuli are not of sufficient luminance. We have solved this problem by using a specially designed multi-layer thin film optical filter which is inserted into the viewing path. This is a band blocking filter that has been designed to separate spectrally the output of the blue and red guns. The pass band of the filter is, however, sufficient to allow the test to be conducted at practical luminances. With the filter in place, the residual output of the blue gun is absorbed by the MP whilst that of the red largely is not. This allows us to make MPOD measurements that are much closer to the true value.

Stimulus

A difficulty when using any psychophysical method is conveying to the subject the judgement that they are required to make. Both flicker and motion nulling paradigms have been previously employed in psychophysical estimates of MPOD. These methods require the subject either to adjust the stimulus to minimise flicker or motion, or judge the relative magnitude of flicker or motion in sequentially presented stimuli. In practice these methods are difficult to use, especially with naive subjects, where lengthy training is usually required for robust test results. In neither of the methods is there an absolute null; some residual flicker or motion is present or the range of any null and uncertain.

We employ the novel method of Cavanagh and Anstis in which the ambiguous judgment of relative luminance, or minimum flicker, is replaced by an unambiguous decision that can be expressed in simple conceptual terms. The subject is simply asked whether a pattern is rotating clockwise or anticlockwise. The task is entirely unambiguous, a grating is perceived as rotating in one direction or another or at equiluminance, to be stationary. This stimulus is ideally suited to a forced choice psychophysical paradigm.

The image above gives a rough impression of the stimulus used. It consists of 4 consecutively presented square wave gratings, each 90 deg out of phase with the next. The first grating is a chromatic grating of red and blue bars. The luminance of the blue is fixed whilst the red luminance can be varied.

The second grating is a purely luminance modulated grating, modulated around the mean luminance of the blue/red chromatic grating. If the luminance of the red component in the chromatic grating is greater than the blue, the observer correlates that with the brighter of the bars of the luminance grating when it is presented. However if the luminance of the red is less, then it is correlated with the darker bar in the luminance grating. This continues in successive grating presentations, so that the sequence of gratings appears to move in on direction or the other, the direction being solely dependent upon the relative luminance of the two components in the chromatic gratings.

In our stimulus, when the red luminance is greater the grating appears to drift upwards, when the blue is greater id drifts downwards. The subject is simply required to decide whether the grating is drifting upwards or downwards in a 2AFC weighted up/down staircase procedure.

A second advantage to the technique is that motion (first order) is encoded only in the luminance channels of the visual system. This means that the adaptive measures taken to suppress blue cone input that have previously been used in heterochromatic flicker photometry are unnecessary, the blue cones having no input to the luminance channel.

Eye tracking

The quality of any psychophysical MPOD estimate relies on knowing the exact retinal eccentricity that is being tested. In a test paradigm such as this, where central fixation must be maintained whist test target are presented, it is a natural reflex to look to the test target when it is presented. A simple video based gaze tracking system is used to ensure that correct central fixation is maintained. The stimulus presentation sequence is inhibited unless the subject is accurately maintaining central fixation, and subject feed back is provided by a visual cue when fixation is correct.