Use of the Hermann Grid Illusion in the Measurement of Contrast Perception in Dyslexia

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Vision Research 45 (2005) 1–8 www.elsevier.com/locate/visres Rapid communications Use of the Hermann grid illusion in the measurement of contrast perception in dyslexia

James M. Gilchrist a,*, Barbara K. Pierscionek b, William M. Mann a

a Department of Optometry, University of Bradford, Richmond Road, Bradford, BD7 1DP West Yorkshire, UK b School of Biomedical Sciences, University of Ulster, Cromore Road, Coleraine, BT52 1SA Northern Ireland, UK

Received 12 March 2004; received in revised form 28 July 2004

Abstract

We measured contrast thresholds for perception of the Hermann grid illusion, using different contrast polarities and mean luminances, in dyslexics and non-dyslexics. Both groups of subjects gave significantly lower thresholds with grids having dark squares and light paths, but there was no significant threshold difference between groups. Perceived strength of illusion was also measured in grids at suprathreshold contrast levels. Dyslexics perceived the illusion to be significantly stronger than non-dyslexics when the grid had light paths and low luminance. 2004 Elsevier Ltd. All rights reserved.

Keywords: Dyslexia; Contrast perception; Hermann grid; Magnocellular

1. Introduction (Livingstone, Rosen, Drislane, & Galaburda, 1991) and psychophysical studies which indicate that some Dyslexia has been defined as a specific reading diffi- dyslexics suffer reduced sensitivity to contrast, flicker culty, inexplicable by any deficit in intelligence, learning and/or motion (Cornelissen, Richardson, Mason, & opportunity, motivation or sensory acuity (Critchley, Stein, 1995; Demb, Boynton, Best, & Heeger, 1998; 1970). The signs of dyslexia are varied and include Everatt, Bradshaw, & Hibbard, 1999; Lovegrove, abnormal phonological awareness, difficulties with writ- 1991; Lovegrove, Bowling, Badcock, & Blackwood, ing, spelling, auditory discrimination and visual process- 1980). However, the magnocellular deficit theory of dys- ing (Habib, 2000). Aspects of vision that appear lexia is by no means universally accepted, particularly correlated with dyslexia are oculomotor instability, im- with regard to contrast perception. In a review of the paired magnocellular processing and increased visual evidence from contrast sensitivity, Skottun (2000) re- discomfort/disturbance due to pattern glare effects ported that, while some studies have found contrast sen- (Evans, 2001). The last two of these may be associated sitivity loss in dyslexics consistent with a magnocellular with changes in contrast perception at threshold and/ deficit, these are outnumbered by studies that show or suprathreshold levels. either no loss of sensitivity, or loss of sensitivity incon- There is evidence that a deficit of the magnocellular sistent with a magnocellular deficit (see also Stuart, system is a significant correlate of dyslexia (Stein & McAnally, & Castles, 2001). More recent studies, too, Walsh, 1997). Support for this comes from anatomical cast doubt on the basic claim for contrast sensitivity loss in dyslexia, with Williams, Stuart, Castles, and McA- * Corresponding author. Tel.: +44 1274 234630. nally (2003) finding no significant difference in con- E-mail address: [email protected] (J.M. Gilchrist). trast sensitivity between dyslexics and controls, while

Bednarek and Grabowska (2002) report increased con- achieved faster visual search performance at low contrast sensitivity in subjects with poor reading and writ- trast, consistent with observations by Meares (1980), ing skills (see also Lovegrove et al., 1982). but OÕBrien, Mansfield, and Legge (2000) found that Differences in contrast perception may also be impli- the dependence of reading speed on text contrast was cated in the occurrence of visual discomfort and distur- the same for dyslexics and controls. bance that is often reported by dyslexics. Some Overall, therefore, we have no definitive picture of individuals are susceptible to such effects when viewing how the contrast responses of dyslexics differ from those repetitive patterns, including text (Wilkins, 1991, 1993, of non-dyslexics, with or without the effects of pattern 1995; Wilkins & Nimmo-Smith, 1987; Wilkins et al., glare. In this study we explore an alternative approach 1984). Wilkins et al. (1984) found headaches, visual to direct psychophysical measurement of contrast re- discomfort and illusory effects in observers viewing sponses in dyslexics and non-dyslexics using the Her- grating patterns, and demonstrated that their occur- mann grid illusion. This is the appearance of illusory rence was influenced by a range of parameters includ- grey spots at the intersections of pathways in a two ing spatial frequency, contrast, location in the visual dimensional array of squares. Baumgartner (1960, cited field, and whether viewed with one or both eyes. Visual by Spillmann & Levine, 1971) explained the illusion in discomfort and disturbance associated with repetitive, terms of receptive fields of on- or off-centre visual neu- high contrast stimuli has been called Ôpattern glareÕ rons, depending on the polarity of the grid. For a grid (Wilkins, 1993), and its effects may be significant, with dark squares and light paths (Fig. 1), it was sug- reducing the clarity and comfort of printed text (Wil- gested that an on-centre receptive field centred on a path kins & Nimmo-Smith, 1987), and the speed and accu- intersection would receive about twice as much inhibi- racy of visual search (Conlon et al., 1998). Meares tion from the field surround as an on-centre receptive (1980) was the first to report that pattern glare effects field illuminated by a light bar. Hence the firing rate are common in children with reading disabilities, and of the neuron viewing the Hermann grid would be re- she found that the reading performance of such chil- duced and the path intersection should appear darker. dren may be improved with smaller print, coloured The opposite effect would be produced in grids with paper or reduced contrast. Alleviation of pattern glare dark paths due to the responses of off-centre neurons lo- by coloured filters was proposed by Irlen (1983) who cated at path intersections. The disappearance of the also claimed that the condition is particularly common illusion on fixation at grid intersections has been attrib- in dyslexia, affecting 12% of the general population but uted to the small size of foveal receptive field centres 65% of dyslexics (Irlen, 1991). Evans, Cook, Richards, (Spillmann, 1994). Grid contrast influences detection and Drasdo (1994) also suggest that susceptibility to of the illusion (Spillmann & Levine, 1971) and sensitiv- pattern glare may be more prevalent in dyslexics, par- ity to it can be reduced in certain disease states that dis- ticularly those who also have an oculomotor deficit turb contrast perception, such as diabetes (Davies & (Evans, Busby, Jeanes, & Wilkins, 1995; Evans, Morland, 2002). Drasdo, & Richards, 1994). Like gratings, Hermann grid patterns enable a ÔpureÕ The pattern glare effect, or Meares–Irlen Syndrome measure of contrast perception, avoiding the effects of (MIS) (Evans et al., 1995), has been described as a form typography, as well as the other sensory and cognitive of physiological hyperexcitability (Wilkins, 1993) which factors involved in text reading. In addition, the grid might be demonstrated by differences in pattern contrast stimulus has general characteristics that are most likely sensitivity. However, an investigation into the aetiology to elicit a pattern glare response; a high contrast, repet- of MIS found no significant difference in static contrast itive pattern structure containing predominantly low to sensitivity between MIS subjects and controls over a medium spatial frequencies. Also, since the illusory grid range of spatial frequencies from 1 to 12cdegÀ1 (Evans response derives from the characteristics of visual et al., 1996). On the other hand, Conlon, Lovegrove, receptive fields, it is possible to selectively stimulate Barker, and Chekaluk (2001) measured perceived distor- receptive fields of different sizes by manipulating the tion and contrast sensitivity in three groups of subjects dimensions of grid paths and squares. If dyslexics are classified according to their visual discomfort level. They indeed more sensitive than non-dyslexics to the struc- found that those who reported high visual discomfort ture and contrast of printed text and other repetitive saw greater distortion with low frequency gratings, patterns, then there may be some manifestation of this and gave reduced static contrast sensitivity in the range in their responses to the Hermann grid illusion. This 1–12cdegÀ1 but no loss of contrast sensitivity for 6Hz work therefore sought to investigate whether dyslexic temporally modulated gratings. and non-dyslexic subjects differ in their contrast sensi- Studies using suprathreshold contrast levels have also tivity for perception of the Hermann grid illusion, reported results that are difficult to interpret consist- and to determine whether there is any difference in ently. Williams, May, Solman, and Zhou (1995), for the perceived strength of illusion between the two example, found that subjects with reading difficulty groups. J.M. Gilchrist et al. / Vision Research 45 (2005) 1–8 3

Fig. 1. (a) Hermann grid patterns used for Experiment 1 (threshold measurement). (b) Hermann grid patterns used for Experiment 2 (contrast matching).

2. Methods (WAIS-III UK) (Wechsler, 1998). All dyslexics had reading performance significantly poorer than expected for 2.1. Subjects age and general intelligence. Only one dyslexic subject reported any form of visual discomfort or disturbance Twenty five dyslexic and twenty five non-dyslexic when reading normal text, and this was described as ÔcontrolÕ subjects were recruited at the University of mild. No such effects were reported by non-dyslexics. Bradford. All were of above average intelligence, had normal visual acuity, and were assessed by their own 2.2. Stimuli optometrist to exclude ocular pathology and oculomotor deficits. Each group comprised 13 males and 12 fe- Stimuli were generated on a Cambridge Research males, with ages from 17 to 42years (mean = 23). Systems VSG2/3 and displayed on a 20in. Pronitron Control group subjects, none of whom reported any his- 80.20 monitor with resolution 1024 · 768 pixels and tory of reading, spelling or specific learning difficulties, frame refresh rate 70Hz. The non-linear luminance re- were matched to dyslexics by sex and age. sponse of the monitor was measured with a Minolta Dyslexic subjects were registered with Disability Of- CS-100 photometer, and the display luminance linea- fice of the University and diagnosed by an educational rized using the inverse of the response function. psychologist. Reading and spelling were assessed using All stimuli were grid patterns comprising 64 (8 · 8) the Wide Range Achievement Test, Revision III squares of one luminance level, separated by horizontal (WRAT3) (Wilkinson, 1993), Graded Word Reading and vertical paths of higher or lower luminance (Fig. 1). Test (GWRT) (Raban, 1985) and Wechsler Objective This configuration produces 49 path intersections, which Reading Dimensions (WORD) Test (Rust, Golombek, may be expected to maximize the strength of the illusory & Trickey, 1993). Verbal and non-verbal intelligence response (Wolfe, 1984). Grids were viewed at a distance was assessed using the Wechsler Adult Intelligence Scale of 1m and subtended 15deg horizontally and vertically. 4 J.M. Gilchrist et al. / Vision Research 45 (2005) 1–8

Grid path width was 0.3deg, and square width 1.5deg. Gaussian spot with maximum visible diameter of The path width corresponds to that which has been 0.3deg, the peak luminance of which was adjustable to found to produce maximum sensitivity to the Hermann match the perceived luminance of illusory spots seen grid illusion (Davies & Morland, 2002). The structural elsewhere in the grid. The size of the central square differences between grid and grating stimuli make it dif- (1.5 · 1.5deg) ensured that no illusory confounding spot ficult to relate our stimuli directly to those which have would be seen within this region. been shown to produce pattern glare and visual discomfort (Conlon et al., 2001; Wilkins et al., 1984). However, 2.3. Procedure if we assume that the active region for the grid illusion (path intersection) has a width equivalent to one bar Subjects viewed the display at eye level and were opti- of a grating stimulus, then a grid path width of 0.3deg cally corrected for the 1m viewing distance. Room illumi- will correspond approximately to a grating with cycle nance was maintained below 50lux, with no light falling width 0.6deg, or spatial frequency 1.67cdegÀ1. This re- directly on the display. Subjects were encouraged to let lates well to the grating frequency that gives maximum their fixation roam over the grid stimuli, advised not to contrast sensitivity in subjects who experience visual maintain fixation at one location and cautioned that fix- pattern discomfort, particularly when the stimuli are ing on an illusory spot would cause the spot to disappear. temporally modulated (Conlon et al., 2001). The dimen- All subjects were given a period of free viewing of the sions of our grid stimuli also reflect the characteristics of stimuli at maximum contrast to ensure familiarity with retinal magnocellular (M) ganglion cell receptive fields. the illusory effect, and were given practice runs on meas- The overall grid size meant that subjects viewing with urement procedures before data collection. All subjects unrestricted eye movements would use receptive fields were naı¨ve with respect to psychophysical methods. lying between 0 and 20deg of retinal eccentricity (grid Experiment 1 measured contrast for threshold corner to corner). In this range, the average centre ra- perception of illusory spots in three grids (L45C+, dius of retinal M ganglion cells is between 0.10 and L45CÀ and L21CÀ), using four runs of a yes–no stair- 0.18deg (Croner & Kaplan, 1995) and our path width case procedure in which contrast was adjusted by chang- of 0.3deg corresponds to a receptive field centre radius ing grid square and path luminances in proportion to of 0.15deg. Likewise, the average surround radius of their relative areas, so mean luminance remained con- M ganglion cells in this retinal area is between 0.72 stant. Each staircase run incorporated direction reversal and 1.19deg (Croner & Kaplan, 1995), so a square when subjects reported a change from seeing to not see- width of 1.5deg ensures that an entire receptive field ing the illusion, or vice-versa. After three reversals at the centred on one path intersection should not encounter minimum contrast step size (0.01 log unit), the proce- any other intersection within its surround. dure was terminated and the final value taken as a Grids having different mean luminance and contrast threshold estimate. First and third staircase runs began specifications were constructed for: (1) measurement of at a contrast where the illusion was clearly visible, grid contrast for threshold perception of illusory spots, second and fourth began at zero contrast. The mean (2) suprathreshold matching of illusory spot contrast. of the four staircase estimates was taken as the final esti- Grid mean luminance levels were calculated from lumi- mate of threshold for each subject. nances of squares and paths weighted by their relative Experiment 2 presented suprathreshold grids areas. Contrast measures were of Michelson form: (L61C+0.4, L12CÀ0.8), and measured the perceived C =(Lmax À Lmin)/(Lmax + Lmin), where Lmax was the contrast of illusory spots using a method-of-adjustments higher of the square and path luminances, Lmin the procedure in which subjects adjusted the contrast of the lower. In Experiment 1, three grids were used. Two Gaussian matching-spot until it appeared equal to illu- had mean luminance 45cdmÀ2 but opposite contrast sory spots seen elsewhere on the grid. No time limit polarity: L45C+ and L45CÀ, where L45 indicates mean was imposed for selection of the matching contrast by luminance, C+ positive contrast (squares lighter than this technique. Four estimates were obtained for each paths) and CÀ negative contrast (squares darker than subject. Matching-spot contrast started at zero for the paths). The third grid (L21CÀ) had mean luminance first and third estimates, and at a level exceeding illusory 21cdmÀ2 with negative polarity. Thus, one pair of grids spot contrast for the second and fourth estimates. The presented equal luminance with opposite contrast polar- mean of the four estimates was taken as the final con- ity, another presented the same polarity at different trast match for each subject. luminance levels. Experiment 2 used two grids; one with mean luminance 61cdmÀ2 and positive contrast 0.4 (L61C+0.4), the other with mean luminance 12cdmÀ2 3. Results and negative contrast 0.8 (L12CÀ0.8). Each of these incorporated a central square region having the same Thresholds (log contrast) from Experiment 1 are luminance as the paths (Fig. 1b), containing a circular plotted in Fig. 2. Analysis by two-way ANOVA J.M. Gilchrist et al. / Vision Research 45 (2005) 1–8 5

Fig. 2. Mean log contrast threshold on three Hermann grids: circles = controls, squares = dyslexics. Error bars indicate one standard error of the mean.

(between-within) showed no significant overall differ- 4. Discussion ence between groups (F(1,48) = 0.467, p = 0.497), and no interaction between groups and grid (F(2,96) = 0.565, Contrast thresholds for perception of the Hermann p = 0.570). Simple comparisons on individual grids also grid illusion revealed no statistically significant differ- failed to show any significant difference between dyslexic ence between dyslexics and non-dyslexics with any grid and non-dyslexic thresholds. A significant difference be- configuration. Significant threshold differences were ob- tween grids was confirmed, however, (F(2,96) = 8.705, tained, for both subject groups, between grids having p < 0.001), and pairwise comparisons incorporating different contrast polarity and mean luminance. The Bonferonni correction showed significant differences be- finding of lower thresholds with negative contrast (dark tween L45C+/L45CÀ (p < 0.001) and L45CÀ/L12CÀ squares, light paths) is consistent with previous findings (p = 0.049), though not between L45C+/L12CÀ (p = that subjects report illusory spots more often within 0.264). Overall, both dyslexic and non-dyslexic subjects black-on-white grids (Spillmann & Levine, 1971). This were least sensitive to the illusion when the grid had may occur because decrements are detected more easily light squares (L45C+), and most sensitive when the grid than increments (Bowen, Pokorny, & Smith, 1989) and/ had dark squares and higher mean luminance (L45CÀ). or due to functional asymmetries in the responses of ON Matching values (log contrast) from Experiment 2 are and OFF retinal ganglion cells (Schiller, 1992). The find- plotted in Fig. 3. Dyslexics perceived higher illusion ing that lower thresholds on negative contrast grids are contrasts, particularly with low mean grid luminance obtained with higher mean luminance is likely to reflect and high negative contrast. Analysis by t-test showed the well-established effect that contrast sensitivity gener- no significant difference between dyslexics and non- ally increases with mean stimulus luminance. dyslexics for grid L61C+0.4 (t(48) = 0.518, p = 0.607), The second experiment provided suprathreshold esti- but a significant difference for grid L12CÀ0.8 mates of illusion strength. Once again the effect is con- (t(48) = 2.400, p = 0.020). Alternatively, comparison sistent across grid conditions, with dyslexics giving between grids revealed no significant difference in non- higher mean values than non-dyslexics (Fig. 3). Differ- dyslexics (t(24) = 0.128, p = 0.899), but a significant ences between the two grid conditions here are marked difference in dyslexics (t(24) = 2.214, p = 0.037). and results confirm that dyslexics do perceive illusory 6 J.M. Gilchrist et al. / Vision Research 45 (2005) 1–8

Fig. 3. Mean log contrast for matching illusory spots on two Hermann grids: circles = controls, squares = dyslexics. Error bars indicate one standard error of the mean. effects significantly more strongly than non-dyslexics Shapley (1990) has shown that high contrast stimuli do when the grid involves dark squares and light paths, generate greater responses in M neurons; that is, M neu- with high contrast and low mean luminance. These find- rons have higher contrast gain than Pneurons ( Croner ings are consistent with Meares (1980) who reported & Kaplan, 1995) and the difference can be traced to dif- that children with reading difficulties who complained ferences in the responses of retinal midget and parasol of visual disturbance with black print on a white page cells (Kaplan & Shapley, 1986). In addition, there is evi- had no such problems when the print contrast was re- dence that low levels of mean luminance emphasise M versed. These children repeatedly stated that they would pathway inputs to the visual cortex (Purpura, Kaplan, prefer text to be lighter than background. & Shapley, 1988) which would also be likely to favour If the lateral inhibition hypothesis of the Hermann M pathway involvement. Furthermore, subjectsÕ use of grid illusion (Baumgartner, 1960, cited by Spillmann & free eye movements during the experiments could also Levine, 1971) is accepted, then the darker illusory spots favour M pathway activity by introducing temporal var- in dyslexics compared to controls may suggest a greater iation of grid stimuli on the retina. A relative increase in level of lateral inhibition in dyslexics that becomes activity in the magnocellular system may be caused by a apparent at suprathreshold levels of contrast. Assuming deficit in processing by the parvocellular system, which that both dyslexic and control subjects viewing the same has been suggested as a possible explanation for a lack grid will use receptive fields with the same centre radius, of inhibition across spatial frequency channels leading determined by grid path width, this result might indicate to experience of visual discomfort in vulnerable individ- that the ratio of inhibitory surround/excitatory centre is uals (Conlon et al., 2001). Those who experienced high larger in dyslexics. Another possibility is that dyslexics levels of visual discomfort, when viewing striped or experience more global integration of the contrast re- repetitive patterns, also reported greater pattern distor- sponse across the grid (Wolfe, 1984). tions at low spatial frequencies (Conlon et al., 2001). The significantly higher illusory response in dyslexics However, this study was not concerned with neural viewing the dark square grid is intriguing, because the processing pathways per se, and the results cannot be spatial dimensions of all the grids used in our study were used to conclude that dyslexics generally have a height- consistent with the dimensions of retinal magnocellular ened magnocellular response compared to controls. Dif- receptive fields (Croner & Kaplan, 1995). This may sug- ferences in eye movements of dyslexics compared to gest greater magnocellular (M) over parvocellular (P) normal readers, when scanning text and viewing the activity in some dyslexics viewing high contrast stimuli. Hermann grid, may be another contributory factor. J.M. Gilchrist et al. / Vision Research 45 (2005) 1–8 7

Scan paths were not taken into account in this study. It Evans, B. J. W., Wilkins, A. J., Brown, J., Busby, A., Wingfield, A., is highly likely, given the nature of the illusion which re- Jeanes, R., & Bald, J. (1996). A preliminary investigation into the quired subjectsÕ eyes to roam, and the lack of a fixation aetiology of Meares–Irlen Syndrome. Ophthalmic and Physiological Optics, 16, 286–296. target, that eye movement patterns would have varied Everatt, J., Bradshaw, M. F., & Hibbard, P. B. (1999). Visual widely between subjects. processing and dyslexia. Perception, 28, 243–254. The results of this study show that sensitivity of both Habib, M. (2000). The neurological basis of developmental dyslexia: dyslexic and non-dyslexic subjects to the Hermann grid an overview and working hypothesis. Brain, 123, 2373–2399. illusion is greater with stimuli having dark squares and Irlen, H. (1983). Successful treatment of learning difficulties. In Annual Convention of the American Psychological Association. Anaheim, light paths (Spillmann & Levine, 1971). The illusory ef- CA. fect of such a grid at high contrast is significantly stron- Irlen, H. (1991). Reading by the colors: overcoming dyslexia and other ger in dyslexics than in controls. Overall, there is a weak reading disabilities by the Irlen method. Avery, New York. but consistent effect, in both illusion contrast sensitivity Kaplan, E., & Shapley, R. M. (1986). The primate retina contains two and perceived strength of illusion at suprathreshold con- types of ganglion cells with high and low contrast sensitivity. Proceedings of the National Academy of Science, 83, trast, that provides some general evidence of a height- 2755–2757. ened response of dyslexics to Hermann grid patterns. Livingstone, M., Rosen, G., Drislane, F., & Galaburda, A. (1991). These results may contribute to an understanding of Physiological and anatomical evidence for a magnocellular defect why the configuration of black text on white paper is of- in developmental dyslexia. Proceedings of the National Academy of ten associated with visual disturbance in dyslexia, and Science, 88, 7943–7947. Lovegrove, W. J. (1991). Spatial frequency processing in dyslexic and the use of the Hermann grid as an exploratory tool normal readers. In J. F. Stein (Ed.), Vision and visual dyslexia may hold promise for further quantitative investigation (pp. 148–154). Boca Raton, FL: CRC Press. of pattern contrast effects. Lovegrove, W. J., Bowling, A., Badcock, D., & Blackwood, M. (1980). Specific reading disability: differences in contrast sensitivity as a function of spatial frequency. Science, 210, 439–440. Lovegrove, W. J., Martin, F., Bowling, A., Blackwood, M., Badcock, D., & Paxton, S. (1982). Contrast sensitivity functions and specific References reading disability. Neuropsychologia, 20, 309–315. Meares, O. (1980). Figure/ground, brightness contrast and reading Bednarek, D. B., & Grabowska, A. (2002). Luminance and chromatic disabilities. Visible Language, 14, 13–29. contrast sensitivity in dyslexia: the magnocellular hypothesis OÕBrien, B. A., Mansfield, J. S., & Legge, G. E. (2000). The effect of revisited. Neuroreport, 13, 2521–2525. contrast on reading speed in dyslexia. Vision Research, 40, Bowen, R. W., Pokorny, J., & Smith, V. C. (1989). Sawtooth contrast 1921–1935. sensitivity: decrements have the edge. Vision Research, 29, Purpura, K., Kaplan, E., & Shapley, R. M. (1988). Background light 1501–1509. and the contrast gain of primate Pand M retinal ganglion cells. Conlon, E., Lovegrove, W., Barker, S., & Chekaluk, E. (2001). Visual Proceedings of the National Academy of Science, 85, discomfort: the influence of spatial frequency. Perception, 30, 4534–4537. 571–581. Raban, B. (1985). Graded word reading test. NFER Nelson. Conlon, E., Lovegrove, W., Hine, T., Chekaluk, E., Piatek, K., & Rust, J., Golombek, S., & Trickey, G. (1993). Wechsler objective Hayes-Williams, K. (1998). The effects of visual discomfort and reading dimensions. Sidcup, UK: The Psychological Corporation. pattern structure on visual search. Perception, 28, 21–33. Schiller, P. H. (1992). The ON and OFF channels of the visual system. Cornelissen, P. L., Richardson, A. J., Mason, A. J., & Stein, J. F. Trends Neuroscience, 15, 86–92. (1995). Contrast sensitivity and coherent motion detection meas- Shapley, R. M. (1990). Visual sensitivity and parallel retinocortical ured at photopic luminance levels in dyslexics and controls. Vision channels. Annual Review of Psychology, 41, 635–658. Research, 35, 1483–1494. Skottun, B. C. (2000). The magnocellular deficit theory of dyslexia: the Critchley, M. (1970). The dyslexic child. London: Heinemann Medical. evidence from contrast sensitivity. Vision Research, 40, 111–127. Croner, L. J., & Kaplan, E. (1995). Receptive fields of Pand M Spillmann, L. (1994). The Hermann grid illusion: a tool for studying ganglion cells across the primate retina. Vision Research, 35, 7–24. human perceptive field organization. Perception, 23, 691–708. Davies, N. P., & Morland, A. B. (2002). The Hermann–Hering grid Spillmann, L., & Levine, J. (1971). Contrast enhancement in a illusion demonstrates disruption of lateral inhibition processing in Hermann grid with variable figure-ground ratio. Experimental diabetes mellitus. British Journal of Ophthalmology, 86, 203–208. Brain Research, 13, 547–559. Demb, J., Boynton, G., Best, M., & Heeger, D. (1998). Psychophysical Stein, J., & Walsh, V. (1997). To see but not to read; the magnocellular evidence for a magnocellular pathway deficit in dyslexia. Vision theory of dyslexia. Trends in Neurosciences, 20, 147–152. Research, 38, 1555–1559. Stuart, G. W., McAnally, K. I., & Castles, A. (2001). Can contrast Evans, B. J. W. (2001). Dyslexic and vision. London: Whurr Publishers. sensitivity functions in dyslexia be explained by inattention rather Evans, B. J. W., Busby, A., Jeanes, R., & Wilkins, A. J. (1995). than a magnocellular deficit. Vision Research, 41, 3205–3211. Optometric correlates of Meares–Irlen Syndrome: a matched group Wechsler, D. (1998). The Wechsler adult intelligence scale (3rd ed.). study. Ophthalmic and Physiological Optics, 15, 481–487. London, UK: The Psychological Corporation. Evans, B. J. W., Cook, A., Richards, I. L., & Drasdo, N. (1994). Effect Wilkins, A. J. (1991). Visual discomfort and reading. In J. F. Stein of pattern glare and colored overlays on a simulated-reading task (Ed.). Vision and visual dysfunction (vol. 13, pp. 155–170). Basing- in dyslexics and normal readers. Optometry and Vision Science, 71, stoke: Macmillan. 619–628. Wilkins, A. J. (1993). Reading and visual discomfort. In D. M. Evans, B. J. W., Drasdo, N., & Richards, I. L. (1994). Investigation of Willows, R. S. Kruk, & E. Corcos (Eds.), Visual process in reading accommodative and binocular function in dyslexia. Ophthalmic and and reading disabilities (pp. 435–456). Hillsdale, NJ: Lawrence Physiological Optics, 14, 5–19. Erlbaum Associates. 8 J.M. Gilchrist et al. / Vision Research 45 (2005) 1–8

Wilkins, A. J. (1995). Visual stress. Oxford, UK: Oxford University Williams, M. C., May, J. G., Solman, R., & Zhou, H. (1995). Press. The eﬀects of spatial ﬁltering and contrast reduction on visual Wilkins, A. J., & Nimmo-Smith, I. (1987). The clarity and comfort of search times in good and poor readers. Vision Research, 35, printed text. Ergonomics, 30, 1705–1720. 285–291. Wilkins, A. J., Nimmo-Smith, I., Tait, A., McManus, C., Della Sala, Williams, M. J., Stuart, G. W., Castles, A., & McAnally, K. I. (2003). S., Tilley, A., et al. (1984). A neurological basis for visual Contrast sensitivity in subgroups of developmental dyslexia. Vision discomfort. Brain, 107, 989–1017. Research, 43, 467–477. Wilkinson, G. S. (1993). Wide range achievement test 3. Wilmington, Wolfe, J. M. (1984). Global factors in the Hermann grid illusion. DE: Wide Range. Perception, 13, 33–40.