Increasing reading speed by using colours: issues concerning reliability and specificity, and their theoretical and practical implications.

By using techniques for precision ophthalmic tinting, individuals who report perceptual distortion of text can often find a colour of illumination that eliminates the distortions and increases reading speed. Most individuals choose green or blue hues, but there is considerable variability. We investigated how specific the colour has to be to obtain optimal reading speed. Individuals who habitually wear coloured filters for reading were asked to read text illuminated by coloured light (without using their filters). Reading speed was measured repeatedly with light of different colours. The colour (chromaticity) at which reading was fastest was consistent from one test session to the next. It was different from one individual to another, but highly specific for each individual: departures of colour from optimum by about 6 JNDs eliminated most of the speed advantage conferred by the optimal colour. It was difficult to attribute the consistency and specificity simply to familiarity with the tint or immediate memory for the colour of illumination. A consecutive sample of 1000 tint prescriptions was analysed numerically. For most prescriptions the variation in chromaticity with different types of lighting was not such as to remove all the potential benefit of the tint, as judged from a model of the effect of chromaticity on reading speed. The exceptions were the few tints that were weakly saturated or purple in colour. Across participants, reading speed was not consistently related to the scotopic energy, to the energy captured by any cone class, or to opponent colour processes. The reading was generally slowest with white light, and not with the colour complementary to the optimum. Explanations in terms of magnocellular deficits and cortical hyperexcitability are briefly discussed.