Visual Discomfort and Reading

Wilkins et al (1980) found that paroxysmal activity in photosensitive epileptic patients can be triggered by binocular viewing of a large repetitive striped pattern with a spatial frequency of 1-4 cycles deg-1, a contrast of 10% or more, and a duty cycle of 50%. Sensitivity to this visual pattern varies in the non-epileptic population (Wilkins et al, 1984) but can produce unpleasant somatic (e.g., eye-strain, headache) and perceptual effects (e.g., flicker, color illusions). This condition has been termed visual discomfort (Wilkins et al., 1984) or visual stress (Wilkins, 1995).

Reading can be especially difficult for those with visual discomfort because most printed text is a repetitive striped pattern with the same characteristics that cause visual discomfort; that is, spatial frequencies between 1-6 cycles deg-1 (Legge et al., 1985), high contrast, and 50% duty cycle (Wilkins \& Nimmo-Smith, 1984; 1987). Those with visual discomfort report a variety of symptoms when they read, such as, movement of print, blurring, letter distortion (size, doubling, fading or darkening), patterns appearing in white spaces, color illusions as letter highlights or moving blobs, nausea, dizziness, headache, eye-strain, and glare discomfort (Wilkins, 2003). One study (Wilkins \& Nimmo-Smith, 1987) found that decreasing the duty cycle (double or triple spacing the lines of text) was sufficient to reduce or eliminate most of the unpleasant symptoms.

Symptom frequency while reading appears to vary with a person's sensitivity to the striped pattern (Conlon et al., 1998) and is more common among those who suffer from headaches or migraines (Wilkins et al, 1984; Conlon et al., 1999). Effective reading time is considerably reduced by visual discomfort symptoms (Conlon et al, 1998; Conlon et al, 1999, Wilkins, 2003). Most college students with visual discomfort cannot read for more than an hour at a time without taking a break (Conlon et al., 1999).

Very little is known about the symptoms and frequency of reading problems associated with visual discomfort. Conlon et al (1999) has conducted the only study that has used a symptom questionnaire with published validity and reliability. The scale consists of 23 items that sample a wide variety of visual symptoms associated with reading using a four point rating scale. The internal consistency of the scale was high (Cronbach alpha = .91), and several experiments confirmed the scale could predict unpleasant symptom frequency and performance efficiency while reading.

The scale was normed on a sample of 515 college student volunteers from Australia. Individuals were rated as having low, moderate, or severe discomfort based on total rating scores below 35%, 35-69.5%, and above 70%, respectively. Twelve percent were classified as severe and 35% as moderate.

Experiments with a sample of the normative group were conducted to validate the scale. They checked for normal or corrected normal visual acuity but did not screen for other concomitant conditions, such as oculomotor dysfunction or learning disabilities. The most frequently reported symptoms included text movement, the appearance of color, and eye-strain. Headache, dizziness and visual distortions were reported infrequently. Oral reading rate was highly correlated (r = .71) with visual discomfort scores. Conlon (personal communication) has found that individuals with visual discomfort read silently at a normal rate but comprehension suffers, whereas oral reading rate is impaired but comprehension is good.

Mechanisms of Visual Discomfort.

Three different mechanisms have been proposed that could induce visual discomfort while reading.

1. Noisy visual system. Working from a model of visual epilepsy, Wilkins (1986) has proposed that visual discomfort is caused by hypersensitivity in some cortical cells that share sensitivity for a particular spatial frequency and orientation. Several studies have examined the contrast sensitivity functions (CSF) of high visual discomfort individuals and found weaker sensitivity at intermediate (Wilkins et al., 1984) and sometimes at higher spatial frequencies (Conlon, Lovegrove, Barker, \& Chekaluk, 2001). A recent pilot study of ten children with visual discomfort found differences in visual evoked potential (VEP) amplitudes among those with a history of headache and migraines (Riddell, Wilkins, \& Hainline, 2006).

These results have been interpreted as evidence for greater noise in the visual system. Prolonged exposure to a single spatial frequency is thought to increase localized cortical excitation that spreads to adjacent cells in a different spatial frequency channel. This spreading of cortical activation is what causes the unpleasant symptoms. Studies by Wilkins and colleagues (1984) showed the most common spatial frequency at which unpleasant symptoms occurred was at 4 cycles deg-1 - the peak of the contrast sensitivity function. Conlon et al. (1999) presented a similar cortical model in which visual discomfort is caused by 'noise' due to poor inhibitory connections between spatial frequency channels.

1. Accommodative and Vergence Insufficiency. Visual discomfort symptoms also have been association with convergence and accommodative insufficiency in children (Evans et al, 1995; Evans et al, 1996; Borsting et al., 1999; 2003, 2004). Convergence Insufficiency (CI), or difficulty maintaining binocular fusion of an approaching object, is defined by one or more clinical signs that include: exophoria at near that is greater than at distance, decreased positive fusional vergence, and a near point of convergence that has receded. The prevalence of CI in school-aged children varies from about 4-8% depending on the presence of two or three CI signs (Rouse et al., 1997; Rouse et al., 1999). Accommodative insufficiency (AI), or difficulty maintaining focus on an approaching target, is often associated with CI. Borsting et al. (2003) found that 40% of school-aged children with AI also had CI, and almost 80% of those with all three signs of CI had AI. Borsting et al (2003) found children with normal binocular vision had an average symptom score of 3.8, whereas the three-sign CI and AI children reported significantly more symptoms, averaging 6.7 and 6.4, respectively.

Recently, Rouse et al (2004) administered a revised version of the Convergence Insufficiency Symptom Survey (CISS) to 46 adults with CI and 46 with normal binocular vision to assess reliability and validity. The CISS is a 15 item 4-point Likert survey that quantifies the severity symptoms with a score range from 0 to 60. The mean CISS score for the CI and normal groups were 37.3 and 11.0, respectively. The interclass correlation coefficient was 0.89, and the 95% limits of agreement were -9.0 to 7.6 indicating good to excellent reliability. These results raise questions about the CI and AI of adults with visual discomfort symptoms and suggest oculomotor dysfunction could be a possible cause of this disorder.

A comparison of several items between Conlon's Visual Discomfort Survey and the CISS shows striking similarities and suggest the two surveys are measuring similar symptomatology.

1. Longitudinal Chromatic Aberration. The focal characteristics of the accommodative response also are affected by longitudinal chromatic aberration (Kruger et al., 1995; Lee et al., 1999; Stark et al., 2002; Rucker \& Kruger, 2004; Rucker \& Kruger 2006). When the L/M-cone contrast ratio of a stimulus is high, the mean accommodation level decreases, biasing the response towards far; S-cone stimulus contrast biases the response towards near (Stark et al., 2002; Rucker \& Kruger, 2004). A low-pass filter reduces accommodation demand by 0.50 D (Kroger \& Binder, 2000), and reading performance tends to be better with blue transmitting filters (Solman, Cho, \& Dain, 1991; Williams, Lecluyse, \& Rock-Fauchex, 1992; Solan, 1998; Iovino et al., 1998; Edwards \& Hogben, 1999; Ray, Fowler, \& Stein, 2005), whereas diffuse red light impairs reading (Chase et al., 2003).

Children with higher L/M ratio sensitivity are poorer readers (Chase et al., in press). This L/M ratio variation may affect their accommodative response and interfere with text perception. Wilkins et al. (2005) found that individuals who benefit from color filters while reading preferred very specific hues and saturations, perhaps to compensate for individual accommodative bias.

Ray et al. (2005) reported a high proportion (39%) of reading-impaired children had poor accommodation. The use of yellow filters significantly may have normalized their accommodative response by balancing L- and M-cone input (Stromeyer et al., 1997; Stromeyer et al., 2000), and reading skills improved more in those who used yellow filters than a placebo. Simmers et al. (2001) found significantly higher variability in the low frequency component of accommodative response functions among five subjects with visual discomfort symptoms. This variability decreased to normal levels when color filters were used or luminance was reduced.

To explore these alternative hypotheses, we will administer the Visual Discomfort Scale to a large sample of students and measure the prevalence, severity and symptom patterns of this problem in a college population. A stratified, random sample of 100 students will receive a vision and psycho-educational assessment to screen for concomitant conditions that might hamper reading performance or contribute symptoms of visual discomfort. Vision testing includes an optometric exam to assess accommodation and vergence, CSF to test for noisy SF channels, and electroretinograms (ERG) and visual evoked potentials (VEP) to measure L/M cone sensitivity variation.