Colour vision standards: Past, present, and future

Physiology of the human visual system

Human visual content and perception have changed. Earlier, humans spent long times in the darkness of caves and jungles with no artificial light. The primary mode of vision was scotopic, using the rods which form 95% of the photoreceptors (100–120 million). By contrast, now almost all our vision is cone (photopic) based even though the cones make up only 5% of the photoreceptors in the retina (3–5 million cones). The cones are densely packed in the fovea, while rods are located almost exclusively in the periphery of the retina. The foveola is rod free and the visual acuity relies on as few as 100,000 cones (or 0.1% of total number of photoreceptors).^[12,13]

The three kinds of cones active in the human visual system are the short-wavelength (S) cones with a maximum response at 450 nm (violet-blue), medium-wavelength (M) cones with a maximum response at 530 nm (green), and long-wavelength (L) cones with a maximum response at 560 nm (greenish-yellow).^[14] In the center – most section of the fovea, the foveola, S cones are virtually absent, reducing our sensitivity to small blue objects.

The photoreceptors transmit neural responses through the bipolar, horizontal, and amacrine cells, which combine and compare the responses from individual photoreceptors before transmitting the signals to the retinal ganglia cells. For approximately 120 million photoreceptors, there are only about 1.6 million ganglion cells, that is, in a ratio of 1:100. While 60–100 rods converge onto a single ganglion cell, only two cones do so, improving the visual acuity of cones. This ratio is very high in the periphery and less near the fovea. Close to the fovea the rods and cones also become more slender. These effects progressively increase the acuity of vision in the central retina. In the central fovea, there are only about 35,000 slender cones and no rods, with each cone being represented in a ratio of 1:1, resulting in the maximum possible visual acuity in the central retina.^[14-16]

Influence of the opponent theory on design of colour vision tests

Stimulation of cones by electromagnetic energy in the visual range generates signals that are perceived as colour vision. Three cones that respond best to yellowish-green (L), green (M), and indigo (S) can perceive 16 million colours, and hue, saturation, and brightness for each. This seemingly impossible feat is explained by the opponent-process theory.^[15,17] This is achieved through four interpretation paradigms; (a) brightness is interpreted using the combined stimulation of the M and L cones. This is also known as the white-black channel. (b) Yellowness is interpreted by the difference between “Brightness” and S cones, that is, if both M and L are strong and S is weak, it is interpreted as yellow, while if all three are strong, it is interpreted as white. (c) If the S cones are relatively stronger than the M and L, it is interpreted as blue. (d) The difference between M and L is interpreted as green-red.^[11] The above would indicate that light with the same blue hue, with a slight change in the brightness could result in “colours” looking yellow or blue. Similarly, a slight change in S cones would look like blue or dark or any other bizarre combination. This would translate specifically to tests used in routine, as under:

Evolution of colour vision testing

Signal lights were introduced in shipping and later in rail transport in the early 19^th century. The colours used were red, green, and white.^[19,20] A train accident occurred at Lagerlunda in Sweden on November 15, 1875, where two trains collided on a single-line track leaving nine dead.^[21] Holmgren, a professor of physiology, suggested the possibility of the accident having resulted from colour blindness. His tests on all employees indicated colour deficiency in 4.8% of the personnel, including a stationmaster and an engineer. This spurred the development of colour vision tests between 1870 and 1930.^[7,19,20,22] In aviation, according to Tredici (1972), the United States Air Force (USAF) devised colour standards in 1918.^[7]

Scientists were aware of colour vision defects since much earlier. One major development came in 1837, when Seebeck introduced matching of 300 hues of paper, rather than naming them,^[23] thus obviating lexical semantics. Seebeck’s work allowed differentiation of red-green blindness and also the development of a severity scale. Holmgren adopted and modified this test in 1877 using skeins of wool. The Holmgren wool test was widely used and even commercially available more than 100 years later.^[19,23]

Two major developments in colour vision testing came in the last quarter of the 19^th century. Stilling, a German ophthalmologist, published the first painted set of PIP in 1878, which filled an important void, and was immediately commercialized. It had 14 successive editions until the author’s death in 1915, after which it continued with newer editions modified by successive authors.^[7,19,22,23] The second major development was by Rayleigh (1842–1919) who demonstrated the peculiarity of human trichromatic/ opponent vision that forms the basis of all advanced colour vision testing. He demonstrated that mixing of red-green lights could be not be differentiated from yellow. This match known as Rayleigh match is basis of all modern anomaloscopes. This also helped define dichromatism and anomalous trichromatism.^[7,19,22,23]

Pseudoisochromatic tests

PIP tests were developed based on observation wherein the colours were selected randomly and plates showing diagnostic efficiency were retained. The plate design is complicated by factors such as form, size, glossiness, texture, and glitter. Stilling applied information obtained from two colour-defective assistants (one red-green blind, the other blue-yellow blind). He used this information ingenuously, in conjunction with the opponent theory to hand paint a series of plates in which figures composed of one set of variegated colour dots appeared on a background of corresponding confusion colours. For instance, he used red-orange dots on a dull yellow background and yellow-green dots on an orange-brown background to detect red-green deficiencies, and pale blue dots on a pale yellow-green background to detect yellow-blue deficiencies.^{[11,19,22,23]}

Soon after the outbreak of WW II, Japanese Ophthalmologist Dr. Shinobu Ishihara was asked by the Japanese Army to develop a colour vision test for its conscripts. He expanded Stilling’s work to achieve more technical results. Using standard camouflage technique, he used a configuration of dots to disintegrate characters, if they could not be identified on the basis of hue.^[5,10,23,24] He was the first to create a hierarchy of plates that could be divided as follows: Demonstration plates that were read by all subjects, normal or pathological, vanishing plates that could be read by normal subjects but not read colour defective subjects, transformation plates that are read differently by normal and colour defective subjects, hidden plates that could not be read by normal subjects, but read by colour defective subjects, qualitative plates that indicated the type of the dyschromatopsia, and quantitative diagnostic plates that indicated the severity of the dyschromatopsia.

Ishihara hand painted each plate in watercolour with Japanese hiragana characters and completed his work in 1916. The test was reconfigured in 1921 for use outside Japan by replacing hiragana characters with numerical designs and sent to universities and ophthalmologists across the world.^[3] This led to a wide acceptance and its adoption for assessing merchant seamen, railway workers, and later navy and air force personnel. Despite its inability to test for blue blindness, it remains the most widely used test in the field.^[3]

At present, there are many other plate tests such as Stilling-Velhagen, Dvorine, Bostrom and Kugelberg, and Hardy, Rand, and Rittler. All make use of the four basic plate designs invented by Ishihara, namely, demonstrative, vanishing, transformation, and hidden.^[5,19] Some modern tests such as the Tokyo Medical College Test, test for blue blindness, can quantify the defect better.^[25] However, as has been brought out earlier in the discussion on the opponent theory, a change in illumination can change the hues perceived on a plate. This may not be understood well enough even by experienced examiners and may be a frequent cause for misinterpretation of the tests.^[19,25]

The low cost, mass production, and rapid administration of PIP tests make them an ideal screening tool. However, the accuracy of the results depends on the interpretation. Furthermore, printing differences between different lots of plates may lead to errors, as can changes in illumination. These plates can be memorized, leading to cheating in enthusiastic applicants.^[26] A consequence of these difficulties is that there are no international pass-fail criteria for PIP tests.

Lantern tests

The signals for both shipping and railways in the late 19^th and early 20^th centuries consisted of lanterns with coloured filters. Hence, it was logical to use similar, standardized lanterns to assess colour vision defects in operators. The original Board of Trade Lanterns (1912) showed “flashed ruby” and “signal green” lights to the observer, either singly or in pairs from a fixed distance of 20 ft. The construction was simple.

There were small and large apertures (controlled by circular diaphragms), the light was furnished by an oil lamp and the colour filters consisted of specimens of glasses with other neutral filters.^[4] Over the years and with the availability of electric power, the oil lamp was substituted with an electric light by Martin (1939).^[6]

Several different models of lanterns are available: Giles-Archer, Holmes-Wright, Royal Canadian Air Force, Optec-900, Edridge-Green, Martin’s, Sloan Colour Threshold Tester, and Farnsworth Lantern. Lantern tests are easy to administer. Their value lies in their simulation of the working condition. Lantern tests do not specifically screen for colour defect, although it is expected that colour-defective observers will not perform as observers with normal colour vision.^{[1,2,7,19,27]}

In view of this, it would be worth reiterating that lantern tests are quite inadequate for any testing other than for signal colours. Furthermore, they have little to recommend themselves above the PIP tests. Since modern-day cockpits employ multiple colours different from signal colours, lantern tests have no place in colour vision testing for modern aviation.^[28]

Arrangement tests

Arrangement tests involve arrangement of a set of coloured objects according to their similarity in colour. Examples are the Farnsworth–Munsell 100 hue test and its shortened version, the Farnsworth–Munsell dichotomous D-15 test, that consists of a series of coloured disks and colour vision defects are diagnosed by incorrect order of disks. A numeric score can be calculated and displayed on a polar graph. Age norms have been prepared giving the range of the expected total error scores for an unselected population as well as the expected intereye variability.^[28,29] Fineness of discrimination can be added using more hues as in the 100 hue test, as against the D-15. Similar tests may be used to test lightness discrimination (Verriest’s lightness discrimination test) or to test saturation discrimination (Sahlgren’s saturation test and Lanthony new colour test).^[29,30]

Observers with congenital colour vision deficiencies make characteristic errors on arrangement tests because their chromatic discrimination ability is lost on particular axes in chromaticity space. Discrimination loss in acquired colour vision defects is more variable. Different hues assess defects of S, M, and L cones. The partitioned scores can be examined to determine whether an acquired colour vision defect causes a particular type of discrimination loss, that is, the S or L/M systems.^[29,30]

Anomaloscope

The principle of the anomaloscope is the fact shown by Rayleigh that the human eye cannot distinguish between pure yellow (from the sodium line, 589 nm), and the mixture of red (from lithium line, 537 nm) and a green (from thallium line, 665 nm), in a certain proportion.^[31] If the brightness of yellow is now changed, the normal eye cannot adjust the red/green to achieve the new brightness. However, a defective eye can still make the match (as it cannot see the relative red-green).^[24,31]

Rayleigh’s initial equation became the gold standard congenital red-green deficiencies. However, more complex deficiencies affecting the blue-yellow axis spawned other colour equations as well, for example, Engelking (1925) allowed the equation blue 470 nm + green 517 nm = cyan 490 nm. Several other instruments based on various equations followed (Trendelenburg, Pickford, Jack Moreland).^[24,31] However, their calibration and use require a high degree of skill.^[19] Interpretation of results is time consuming and requires expertise. These issues preclude mass use of anomaloscopes, despite them being the gold standard for colour vision defect detection and grading.

Cone contrast test (CCT)

The CCT by Dr. Jeff C. Rabin of the USAFSAM is unique in that it pays as much attention to the background as the letters to be identified. Second, the CCT does not test for all the myriad colours, but concentrates on stimulating individual cone type populations, that is, L-cone, M-cone, and S-cone.^[32,33]

The opponent theory has been explained in detail above. The CCT has been designed to isolate the cone-opponent mechanisms by designing test stimuli that change the contrast to the cone-opponent mechanisms with minimal change in achromatic contrast or brightness. In addition, random achromatic contrast is often added to the background to further limit detection by the achromatic mechanism.^[32-34] The cone contrast is computed as the difference between letter and background divided by background multiplied by 100: Cone contrast $\left(\%\right)=\left[\left(E_{letter}-E_{background}\right)/E_{background}\right]\times 100.$ The test is conducted monocularly in a dark room at 1 m after adequate refraction. The test begins with a random series of 20 reddish letters (L), followed by greenish letters (M), and then violet letters (S) progressing from most visible down to least visible. The letter appears for a duration of 1.0–1.6 s (duration increases as contrast decreases) followed by an intertrial interval (grey field) of equal duration.^[32,33] The log contrast sensitivity scores are normalized to an intuitive 100 point scale so that each letter that is read correctly counts as 5 points, making the max score 100 and the minimum score 0.

Peculiarities of the Indian Scenario

The Indian scenario is different from other countries, which make designing of colour vision tests easier to some extent. These factors are Common Regulator: The Director General of Medical Services for the Air Force is the advisor for medical standards for all flying in India, military, as well as civil. Common Airfields: Civil flying in India happens from defense airfields and vice versa. This requires common standards for all safety critical issues, including colour vision. Common Airspace: The Govt. of India has removed restrictions from civil flights in military airspace, requiring commonality between standards. Expansion of Civil Flying: Civil flying in India is expanding exponentially. As new aircraft is being acquired, it is possible to influence manufacturers to adhere to regulatory requirements specific to India. Indigenous Designing of Aircraft: The Indian aircraft industry stands at a cusp where it is about to design and fly more and more indigenous aircraft, both military and civil. There is thus the need for a common standard that the industry can follow, to assure safety.

Characteristics of an ideal colour vision test

As has been brought out above, a colour vision test introduced today will need to be compatible with today’s technology and future ready. The following are the envisaged characteristics for an ideal colour vision test:

Sensitivity

The need for high sensitivity in the aviation scenario is easily understood. It is felt that if the above principles are applied, an extremely sensitive test may be possible.

Specificity

Any test or testing standard is always a trade-off between sensitivity and specificity. Since the said test would be an entry-level elect out standard for aviation, a lower specificity may be acceptable.