The optometric correlates of migraine


  • Deacon E. Harle,

    1. The Institute of Optometry, 56-62 Newington Causeway, London
    2. Department of Optometry and Visual Science, City University, Northampton Square, London, UK
    Search for more papers by this author
  • Bruce J. W. Evans

    1. The Institute of Optometry, 56-62 Newington Causeway, London
    2. Department of Optometry and Visual Science, City University, Northampton Square, London, UK
    Search for more papers by this author

Correspondence and reprint requests to: Deacon Harle.
Tel.: 020 7407 4183; Fax: 020 7403 8007.
E-mail address:


Migraine is a common, chronic, multi-factorial, neuro-vascular disorder typically characterised by recurrent attacks of unilateral, pulsating headache and autonomic nervous system dysfunction. Migraine may additionally be associated with aura; those focal neurological symptoms that may precede or sometimes accompany the headache. This review describes the optometric aspects of migraine headache. There have been claims of a relationship between migraine headaches and errors of refraction, binocular vision anomalies, pupil anomalies, visual field changes and pattern glare. The quality of the evidence for a relationship between errors of refraction and binocular vision and migraine is poor. The quality of the evidence to suggest a relationship between migraine headache and pupil anomalies, visual field defects and pattern glare is stronger. In particular the link between migraine headache and pattern glare is striking. The therapeutic use of precision-tinted spectacles to reduce pattern glare (visual stress) and to help some migraine sufferers is described.

Brief historical overview of explanations of migraine

From 3000 bc, vision has been linked to migraine headache (Alvarez, 1945; Pearce, 1986). Hippocrates himself alluded to the visual prodrome of migraine (Allory, 1859). Migraine has been described in other ancient writings, too numerous to review here. Particularly relevant to the present review, Celsus (ad 30, cited by Thomas, 1887) listed sunlight among the triggers of migraine. The severity of migraine, and its association with photophobia, was highlighted by Araetaeus (ad 81, translated by Adams, 1856):

For they flee the light; the darkness soothes the disease; nor can they bear readily to look upon or hear anything pleasant… The patients are weary of life and wish to die.

Gowers (1886) referred to the two main theories of migraine, vascular and neural; an observation which is equally valid today. The 1920s saw allergic theories come and go, as did the psychosomatic theories of the 1950s (Pearce, 1986). Nowadays migraine headache can be considered to be a reaction or biological adaption determined by a primary disorder of brain threshold in combination with a variety of external precipitating factors. Together, these lower this threshold to a point when a migraine attack will occur.

Pathophysiology of migraine

Goadsby et al. (2002) have reviewed migraine pathophysiology from a medical perspective, but in a broad sense, migraine can be thought of as a tendency to have headache that is characterised by certain associated symptoms. The basis of this predisposition has been attributed to a lack of stability in the control of pain, the control of sensory information coming from the pain producing intracranial structures and sensitivity to cyclic changes in the central nervous system (Lance and Goadsby, 1998).

The migraine brain has a reduced threshold to a variety of stimuli, and this has been described as cortical hyperexcitability. The factors that set this threshold are genetic (Ophoff et al., 1996; Ducros et al., 2001), and involve magnesium deficiency, excitory amino acids, sensitivity of the dopamine system and the hypothalamus, reduced habituation to visual and auditory stimuli, and vascular reactivity. Because of this reduced threshold, migraines can be initiated by ‘triggers’. Such triggers can be divided into internal and external. One example of an internal trigger might be hormonal factors, whilst external triggers could be flickering lights, certain patterns or strong smells. External triggers have the potential to cause, and therefore to prevent, migraine and will be outlined in more detail later.

Once triggered, a migraine has two main consequences: spreading depression (which may or not be perceived as aura) and pain. Leão (1944) described ‘spreading depression’ as a progressive shutdown of cortical function and suggested that it may be related to the fortification spectra of migraine. Waves of cortical inhibition, sometimes preceded by transient excitation, move slowly over the cortex (2–3 mm min−1), suppressing normal activity, and take 5–60 min before recovery takes place.

Spreading depression is associated with vascular changes (Lauritzen et al., 1982; Piper et al., 1991; Goadsby, 1992). One such vascular change that has been suggested in patients with migraine with aura is a ‘spreading oligaemia’ (Olesen et al., 1981; Dreier et al., 2001). Dreier et al. (2002) have suggested that the link between the vascular oligaemia and the neurological spreading depression may be that endothelial irritation triggers cortical spreading depression. Hadjikhani et al. (2001) showed vasoconstriction and then vasodilation followed the cortical spreading depression using an imaging study. The oligaemic waves of reduced blood flow progress over the cortex at the same rate of 2–3 mm min−1 as cortical spreading depression. They start in the visual cortex and advance forward without respecting arteriolar territories. These vascular changes can last several hours and are followed by delayed hyperaemia (Andersen et al., 1988). As the spreading oligaemia reaches sensory motor areas of the brain, the patient experiences the focal neurological aura symptoms. The neurological changes during aura parallel what is seen if the brain is directly stimulated (Penfield and Perot, 1963; Brindley and Lewin, 1968) and are also remarkably similar to the changes that would be predicted if ocular dominance columns (Hubel and Weisel, 1968) in the cortex were serially activated.

Woods et al. (1994) demonstrated a spreading oligaemia directly with a positron emission tomography study. Interestingly, the patient in this study did not perceive aura in any traditional sense, suggesting that the oligaemia can traverse the whole cortex without the patient experiencing symptoms. Indeed, Lance and Anthony (1966) claimed that only 10% of migraine patients perceive the fortification spectra but 25% of patients perceive less specific symptoms of ‘spots before the eyes’ or ‘shimmering vision’ covering the entire visual field.

Other neuro-vascular interactions can occur with migraine. Kruit et al. (2004) found that some patients with migraine were at risk of subclinical lesions in certain brain areas and suggested that the cerebellar region of the posterior circulation territory was an area where migraine sufferers had a greater number of infarcts than controls. Lipton and Pan (2004) considered that this might be evidence that migraine is a progressive brain disease as this area had been previously implicated in persons with stroke and migraine (De Benedittis et al., 1995; Hoekstra-van Dalen et al., 1996).

There is some pathophysiological evidence linking the aura phase of migraine and the pain phase of migraine. Moskowitz (1984) considered that the spreading depression of the cortex might depolarise trigeminal nerve fibres and initiate pain. However, if this hypothesis were true then the headache would always develop on the side of the head responsible for the aura symptoms (e.g. a left sided headache would arise from a right field aura). Olesen et al. (1990) showed that in 38 patients with migraine with aura, three experienced headache on the ‘wrong’ side and Jensen et al. (1986) showed that aura symptoms were ipsilateral to the headache in 19 patients and contralateral in 18 patients. Thus, there must be some ‘central link’ which can trigger pain on either side of the head for one-sided aura symptoms. Bolay et al. (2002) have suggested that cortical spreading depression activates trigeminal vascular afferents to evoke meningeal and brainstem events that potentially lead to the development of headache.

An alternative explanation to the link between pain and aura was provided by May et al. (2001) who examined neural influences on the cranial circulation by studying healthy volunteers’ responses to injection of the pain-producing compound ‘capsaicin’ using magnetic resonance angiographic techniques. They concluded that their data was consistent with the notion that pain drives changes in vessel calibre in migraine, not vice versa.

Migraine classification

Headache is an extremely common symptom presenting to primary health care professionals, and an accurate diagnosis is essential to ensure both the correct management of benign conditions and to ensure that when headache presents as a symptom of serious disease then it is dealt with appropriately. The International Headache Society (IHS) published the second edition of The International Classification of Headache Disorders recently (Headache Classification Sub-Committee of the International Headache Society, 2004). The IHS classification is lengthy and is briefly summarised in Table 1. The first edition has been summarised, from a clinical optometric viewpoint, by Patel et al. (2003).

Table 1.  A summary of the classification of migraine
Migraine without aura
Migraine with aura
 Typical aura with migraine headache
 Typical aura with non-migraine headache
 Typical aura without headache
 Family hemiplegic migraine
 Sporadic hemiplegic migraine
 Basilar-type migraine
Retinal migraine
Childhood periodic syndromes that are commonly precursors of migraine
 Benign paroxysmal vertigo of childhood
 Abdominal migraine
 Cyclical vomiting
Complications of migraine
 Chronic migraine
 Status migrainosus
 Persistent aura without infarction
 Migrainous infarction
 Migraine-triggered seizure
Probable migraine
 Probable migraine without aura
 Probable migraine with aura
 Probable chronic migraine

Migraine is described in section 1 of the IHS classification. Section 11 of the classification describes headache or facial pain associated with disorders of the cranium, neck, eyes, ears, nose, sinuses, teeth, mouth or other facial or cranial structures. The ‘eyes’ section is further subdivided into acute glaucoma, refractive errors, heterophoria or heterotropia and ocular inflammatory disorder. Section 13 of the classification describes cranial neuralgias and central causes of facial pain: ophthalmoplegic migraine, optic neuritis and ocular diabetic neuropathy are included in this section. Since the present review is concerned only with migraine, these other types of eye headache will not be discussed. However, it should be noted that, to an optometrist, these sections of the IHS classification would appear weak.

Notwithstanding these comments, it must be recognised that the IHS classification is a useful framework for classifying headaches (Leone et al., 1994). However, others have suggested that the classification is more useful for research than for clinical practice (Cady and Dodick, 2002).

The visual disturbances of migraine

Visual aura

The cornerstone to visual aura in migraine are fortification spectra or ‘teichopsia’, although this may present in only 10% of migraine patients (Lance and Anthony, 1966). Originally described by Airy (1870), the term ‘teichopsia’ was coined from the Greek terms ‘teckhos’ meaning fortification and ‘opsis’ meaning seeing, alluding to the zig-zag design of early Italian military fortifications with which Airy was familiar. The symptoms of scintillating scotoma and a marching fortification figure that gradually expands and then breaks up is characteristic of migraine with aura. Wilkinson (2004) has reviewed migraine visual aura in the context of other visual hallucinations and suggested how these might relate to the neural mechanism of aura.

Queiroz et al. (1997) showed that visual aura accompanied the patient's first headache in 39% of patients but only 19% had visual aura with every attack. The free period between visual aura and head pain was <30 min in 75% of cases. The symptoms were described as ‘small bright dots’ (42%), ‘flashes of light’ (39%), ‘blind spots’ (32%) and ‘foggy vision’ (27%). Fortification spectra were reported by only 20%.

Usually migraine aura are binocular but rarely migraine can affect the anterior visual pathway and produce monocular symptoms. These retinal migraines produce monocular scotomas, and are caused if any of the circulation of the anterior visual pathway becomes involved in the angio-spastic disturbances of migraine. Often the visual loss is described as a black-out or grey-out which can last from seconds to hours, the vast majority lasting <30 min (Hupp et al., 1989).

Migraine aura can occur without headache. The Framingham Study (Wijman et al., 1998) demonstrated that these migrainous visual accompaniments occur in just over 1% of the population aged between 30 and 62 years. This study showed that the mean age of onset of these symptoms was 56 years and in 58% of subjects no headache was reported. Indeed, 42% had no headache history at all.

A variety of ophthalmic conditions may produce visual-aura-like symptoms and need to be differentially diagnosed. Table 2 contrasts the signs and symptoms of these conditions.

Table 2.  A variety of ophthalmic conditions may produce visual-aura-like symptoms and need to be differentially diagnosed. The characteristic signs and symptoms of these conditions are shown
DiagnosisMonocular or binocular disturbancesOnset of symptomsUsual duration of symptomsScotomaPhotopsiaeBuild up of scotomaMigration of scotoma
Migraine with auraBinocularGradual15–30 minYesYesYesYes
Retinal migraineMonocularGradual15–30 minYesNoNoNo
Amaurosis fugaxMonocularSuddenMinutesYesNoNoNo
Occipital transient ischaemic attackBinocularSuddenMinutesYesYesNoNo
Posterior vitreous detachmentMonocularSudden1 monthNoYesNoNo
Retinal break or detachmentMonocularSudden1 month to continuousYesYesYesNo

Photophobia and glare

Most migraine sufferers avoid bright light during headache (Selby and Lance, 1960) and many migraine sufferers feel the need to use sunglasses even in between attacks (Drummond, 1986). Wolff (1963) argued that true photophobia is pain induced and exacerbated by bright light, for example in corneal disease or anterior uveitis, and is derived from stimulation of the trigeminal nerve. He argued that glare or dazzle, on the other hand, is uncomfortable but not painful. Glare can be caused by stray light scattering into the eye from ocular structures (such as cataract) or environmental factors (such as a poorly placed lamp). Glare might also be caused by a general excitability of the senses in migraine sufferers, and they have been shown to be more susceptible to glare than controls (Drummond, 1986).

Stimulation of the trigeminal nerve during a migraine attack probably accounts for photophobia. Drummond and Woodhouse (1993) stimulated the trigeminal nerve with ice on the forehead and measured discomfort thresholds for migraine sufferers and controls. They showed that trigeminal discharge contributes to photophobia in migraine sufferers and that this trigeminal discharge continued during headache-free periods. However, Drummond (1997) has shown that it is glare, rather than true photophobia, that probably accounts for the light sensitivity experienced by migraine sufferers between attacks. This heightened sensitivity to light is consistent with the heightened sensitivity found to other visual stimuli in migraine sufferers, such as pattern glare, which is discussed later.

Visual migraine triggers

Migraine triggers are the internal or external factors that excite the migraine brain above its genetically reduced threshold and in so doing, precipitate the chain of neurovascular events that produce a migraine headache. It has been suggested that common triggers include certain foods, stress, smells, hormonal changes, irregular meals, changes in sleep pattern and environmental factors such as excessive heat, light or noise (Peatfield and Olesen, 1993). It should be noted that some authors suggest that migraines occur spontaneously and that the triggers that patients associate with their migraine headache are actually due to the fact that in the ‘prodrome’ phase of a migraine attack some migraine sufferers have a craving for certain foods or drinks (Dowson and Cady, 2002). These then may be blamed for the attack when in fact they are a consequence. Nevertheless, it is generally considered that by making lifestyle changes, the frequency and severity of migraine headache can be reduced.

Migraine patients are sensitive to light during and between headaches (Drummond, 1986). It has also been stated that migraine, as compared with other headaches, is worse during midnight-sun summer than during the polar night (Salvesen and Bekkelund, 2000). These visual stimuli do not have to be strong. Jacome (1998) described a patient who, on multiple occasions, could trigger his typical headache within 30 min just by rubbing his eyes gently and inducing bilateral photopsias. Liveing (1873) described falling snow as a migraine trigger. Debney (1984) produced a thorough review of the literature relating to visual stimuli as migraine trigger factors. She showed that visual stimuli were quoted by at least 10 other authors and ranked visual triggers as similarly important to other more obvious triggers such as stress and hormonal factors.

Debney (1984) reviewed the medical notes of 344 migraine patients and showed that 62% had ‘glare’ as a precipitating factor, 53% had ‘flicker’ as a precipitating factor and 1% had ‘colour’ as a precipitating factor. Debney analysed these findings further and sought to correlate other precipitating factors to those patients who claimed their migraines were induced by visual stimuli. She found significance only with two factors: 1, ‘other sensory and environmental factors’; 2, ‘dietary factors’. Debney (1984) suggested:

…that it would be interesting if the aberrant biochemistry underlying dietary triggers of migraine also affected the sensitivity of the sufferer to visual triggers and to other sensory and environmental triggers.

Debney (1984) then analysed her data further and split them into two groups, one detailing visual tasks quoted to have induced migraine because of glare, and one detailing visual tasks quoted to have induced migraine because they involve flicker. In the glare group, she found that the following situations had all been implicated in precipitating migraine: sun reflections; rippling water or sea; at the beach; snow; paper; chrome trim on a car; microscopy; facing bright windows; fluorescent lighting. In the flicker group, she found that the following situations had all been implicated in precipitating migraine: television; cinema; faulty fluorescent lighting; lighting in vehicular tunnels; flashlights; headlights; stroboscope; travelling past railings, telegraph poles and fences (by train).

Debney (1984) listed many visual stimuli reported to induce migraine. This list was lengthy but can be summarised by splitting visual triggers into four simple groups; glare, flicker, patterns and colours. Glare could be explained by the trigeminal nerve sensitivity demonstrated by migraine sufferers (Drummond, 1986) or, with flicker, patterns and colour, by cortical hypersensitivity theories (Wilkins, 1995). Both these aspects will be discussed later.

Traditional clinical advice is to avoid trigger factors. Interestingly however, Martin (2000) showed that in patients with visually triggered headaches, there is a desensitisation period such that the visual triggers become less likely to produce headache symptoms with continued exposure. This finding could conceivably alter the way headaches are managed, with exposure to triggers to produce desensitisation as a possible approach, rather than avoidance.

In conclusion, visual stimuli are common and potent migraine triggers. This is emphasised by the fact that some experimenters have used an alternating red and green checkerboard as a strong visual stimulus to cause migraine headache for experimental purposes (Cao et al., 1999).

Refractive errors and migraine

In the early 1900s, uncontrolled studies by Gould (1904) and Snell (1904) argued that low refractive errors, particularly astigmatism, are associated with migraine.

Turville (1934) claimed that uncorrected errors of refraction were a major cause, or at least an important precipitating factor, in cases of migraine. He also claimed that the conventional methods of the time used to provide correction for refractive errors were inadequate. In his opinion, the investigation of refractive errors must include both manifest and latent errors. He defined a latent error not just as latent hyperopia but also as heterophorias, accommodative anomalies ‘and in fact any departure from normal visual activity, physiologically, optically, functionally, mentally and psychologically.’

Turville stated that even an inequality of refractive error of 0.25 dioptres was important in many cases and noted that the difference was rarely more than 0.75 dioptres. Turville's study lacked a control group and lacked any form of statistical analysis. It is unclear whether it was the correction of refractive errors, the correction of any decompensated phorias, or placebo effects that were relieving symptoms.

Wilmut (1956), although mostly concerned with the effect of binocular vision on migraine (described below), considered the refractive errors of 116 cases of migraine and compared them to a non-migraine group. He found a similar prevalence of refractive errors in migraine and a non-migraine control group.

Several other authors have argued that headaches or migraine are associated with uncorrected refractive errors, but these studies will not be described in detail because there were no control groups or statistical analyses (Gordon, 1966; Lanche, 1966; Vaithilingham and Khare, 1967; Cameron, 1976; Hedges, 1979; Worthen, 1980).

Waters (1970) identified by a questionnaire, in a random sample of a general population, groups of individuals with; headache, unilateral headache, migraine, and a fourth group who had not had a headache for a year. A masked assessment of the visual acuity and ocular-motor balance was then performed on each group. Visual acuity was measured unaided and aided if spectacles were worn. Waters found that there was no significant difference between the unaided vision, or visual acuity with spectacles if normally used, of either men or women in the four groups. In addition, he found no significant difference between groups in the number of individuals wearing spectacles for either distance or near vision. He concluded by suggesting that these data showed that in the general population headaches are seldom caused by a visual defect. However, Waters did not assess refractive error at all and so doubt must be raised over his conclusions.

Vincent et al. (1989) determined the prevalence of visual symptoms and eyestrain factors in a group of chronic headache sufferers as compared with age- and sex-matched controls and found near visual tasks to be one of the many visual triggers to chronic headache. However, this questionnaire survey did not take account of whether the near visual tasks were carried out with corrected or uncorrected refractive errors. Nevertheless, it has been suggested (Gordon et al., 2001) that Vincent's data could suggest a relationship between headache, refractive error, accommodation and convergence.

Gordon et al. (2001) reviewed the experimental and clinical evidence on possible links between refractive errors and headaches and listed several issues that were still to be resolved. This review did not relate specifically to migraine, so will not be described in detail. Evans et al. (2002), in a study described in the next section, found no significant difference between a group of migraine and a group of control patients in the subjective refractive error or in the proportion of participants who wore spectacles.

To conclude, the association between uncorrected refractive errors and migraine seems to be equivocal. Early studies reported much anecdotal evidence but the few modern studies, which included masked control groups and statistical analyses, have found little evidence. Often researchers have failed to classify headaches correctly and so data relating specifically to migraine is rare. In addition, little or no evidence appears to relate to any possible pathogenic link between refractive errors and migraine.

Binocular vision (orthoptic anomalies) and migraine

Snell (1904) argued that heterophoria is a cause of headache, especially esophoria when found in conjunction with myopia. Turville (1934) suggested that low convergent and divergent fusional reserves are correlates of migraine and that base in prisms are an effective treatment for many cases of severe classical migraine. Turville describes his first successful case of relief of migraine with base in prisms and it is interesting to note that this patient was esophoric rather than exophoric as might have been expected. The prism power was determined in an unconventional way: as one-third of the recovery point from the measurement of the divergent fusional reserves. He described a migraine sample of 123 cases, but there was no control group or placebo treatment. As recently as 2000, the use of base in prism to relieve migraine headache was still advocated (Bush, 2000).

Wilmut (1956), using a polarised version of the Turville Infinity Balance, found that 91% of patients with migraine had ‘excessive exophoria’ and had previous argued that 56% of his cases were cured with base in prism (Wilmut, 1951). Wilmut's (1956) study was of a clinical sample and may have suffered from referral bias, and does not appear to have been a randomised control trial. However the results were compared with an unspecified control group in which exophoria occurred in only 25%.

Waters’ (1970) questionnaire regarding headache and migraine sufferers, discussed in the previous section, not only looked at visual acuity but also ocular-motor balance. The ocular-motor balance was assessed by the cover test and a Maddox hand frame with habitual spectacle correction, if worn. Thus, the total dissociated strabismus or phoria was assessed. Waters stated that there was no evidence that the proportion of subjects with esophoria or exophoria for either distance or near vision differed in the four groups in either sex. Unfortunately, the data for esophoria and exophoria were combined and so data on this aspect are not meaningful. The only statistically significant finding Waters made was that the migraine group had a higher proportion of individuals who had hyperphoria at near. Waters stressed however that this result was only meaningful when the male and female groups were analysed together and over 20 chi-squared tests had been completed. He concluded by suggesting that these data showed that in the general population headaches are seldom caused by a visual defect. He also noted that the beneficial effect of any treatment, if applied in an uncontrolled manner, could not be considered as evidence relevant to the aetiology of headache.

Friedman (1977) claimed that ‘fusional stress’ could accompany ‘dynamic binocular seeing’ and that this could be a cause of migraine. He advocated a specific instrument for intensive visual training. Friedman presented no data to back up his claims, only case study reports.

Worthen (1980) studied the effects of stimulating extraocular muscles in patients on whom operations for strabismus were performed under local anaesthesia. The muscles were exposed under light anaesthesia and then stimulated in various ways. Pinching, pricking or cutting the recti muscles caused no sensations, but traction produced prompt exclamations of pain. The pain was always described as an aching sensation localised deep in the eye/orbit on the side of the stimulated muscle. Worthen went on to describe two case studies where the reproduction of extraocular muscle imbalances produced consistent results of headache and aesthenopic symptoms. Electromyographic recording of these patients suggested that the symptoms arose from increased tension in the muscles of the head and neck. Nevertheless, Worthen claimed that the headaches caused by muscle imbalance (heterophoria) could be eliminated by proper alignment of the visual axes and stated that prisms, orthoptic training, or even surgery may be necessary. He suggested that occlusion could be used to diagnose headaches associated with binocular anomalies. Although Worthen (1980) used an interesting approach, his small number of subjects limits the strength of his conclusions.

Sucher (1994) related the symptoms of headache to the ‘monocular blur effect’: a consistent blur of one eye when viewing the 6/18 letters on a letter chart during the Turville infinity balance test, whilst the patient raises and lowers their chin. Sucher found a statistical relationship between this monocular blur effect and patients who have three or more headaches a month. He also found that the monocular blur occurred on the same side as lateralized headaches in 94%, and then in 93%, of two cohorts of patients tested. Sucher speculated that the monocular blur effect could be corrected by prisms, and that this correction would then relieve tension on the ocular motor system and so remove a source of headache. However, Sucher's study did not look at the effect of treatment.

Evans et al. (2002) compared 21 migraine sufferers to 11 controls and found no difference between the groups in relation to strabismus or hyperphoria. The main purpose of this study was to investigate the effect of coloured filters (Wilkins et al., 2002), so the migraine sufferers were selected as those who found a coloured filter to be helpful. They therefore did not represent a ‘normal’ group of migraine sufferers. Evans et al. (2002) did find using one test method, that the migraine group tended to have a marginally decompensated exophoria at near; however, other test methods suggested that the migraine group were as able to compensate for their exophoria as the control group.

Decompensated heterophoria, the diagnosis of which is discussed by Evans (2002), has been linked to headaches by many authors (e.g. Jenkins et al., 1989; Yekta et al., 1989; Evans, 2002). However, these authors do not specifically discuss migraine.

The association between anomalies of binocular vision and migraine seems to be equivocal. Early studies have suggested anecdotal evidence but the few modern studies, which have been more statistically and methodologically robust, have either found little evidence, or have generally related to headache or aesthenopic symptoms, rather than specifically to migraine.

Visual fields and migraine

The visual system beyond the eye can be investigated in a number of ways. Psychophysical testing of visual processing can shed light on perceptual issues in migraine as discussed by Coleston et al. (1994), McKendrick et al. (1998) and others. These studies do not involve clinical optometric approaches and will not be discussed in detail here, but are reviewed by Chronicle and Mulleners (1996). Electrophysiology can directly measure cortical activation but is also not an optometric procedure and is extensively reviewed elsewhere (Aurora et al., 1998; Áfra et al., 1998, 2000; Cao et al., 1999).

Several studies have assessed visual fields in migraine. McKendrick et al. (1998) showed deficits to tasks involving 16 Hz flicker using a Medmont 6000 perimeter auto flicker paradigm in a single migraine sufferer. Later, McKendrick et al. (2000) performed similar temporally modulated perimetry in 16 migraine sufferers and 16 controls and suggested that migraine sufferers have selective visual dysfunction for temporally modulated targets of a temporal frequency >9 Hz.

Other visual field anomalies have been found in migraine patients. McKendrick et al. (2002) performed short-wavelength automated perimetry (SWAP) and standard automated perimetry (SAP) using a Humphrey Visual Field Analyser. Although they did not find a significant difference in mean deviation and pattern standard deviation between migraine sufferers and controls using SAP, both these parameters were significantly worse in the migraine group using SWAP. The authors suggested that migraineurs should not be included in visual field normative databases.

Klein et al. (1993) reported results from the Beaver Dam Eye Study that showed no relationship between open-angle glaucoma and migraine headache. They used diagnostic criteria based on visual fields, intraocular pressure, cup/disc ratio and history. Usui et al. (1991) found no greater prevalence of migraine in a glaucoma population compared with a normal population and Pradalier et al. (1998) commented that migraine prevalence was not significantly different between normal and high tension glaucoma sufferers.

Alternatively, other authors have found that there is a relationship between normal tension glaucoma and migraine headache (Cursiefen et al., 2000). In particular, migraine has been considered a risk factor for glaucomatous visual field progression (Drance et al., 2001). Comoglu et al. (2003) found glaucomatous-like visual field defects in patients with migraine in the absence of raised intraocular pressures and suggested that there might be a relationship between the pathophysiology of normal tension glaucoma and migraine. McKendrick agreed with this viewpoint (McKendrick et al., 2000, 2002) and concluded that the similarity of SWAP defects and temporally modulated perimetry defects in migraine sufferers and glaucoma sufferers might raise the possibility of a common pre-cortical vascular involvement in these two conditions.

We would suggest an alternative explanation that migraine headache might cause a magnocellular-specific dysfunction unrelated to glaucoma. Such an interpretation would account for the fact that some studies have suggested a link to normal tension glaucoma, as intraocular pressures would remain unaffected. We are currently comparing visual fields, ocular tensions, and optic nerve head analysis in migraine and control groups to investigate this hypothesis.

Interestingly, McKendrick and Badcock (2003) have shown that migraine sufferers with visual field loss to temporally modulated targets but not to SAP exhibit dysfunction of both the parvo-cellular and magno-cellular pathways. How this might relate to the mechanism of visual field dysfunction in migraine is yet to be investigated. Coleston et al. (1994) also found evidence suggesting both magno- and parvo-cellular deficits in migraine. These authors suggested that the deficit was pre-cortical, and they noted that this could reflect either intrinsic abnormalities or a consequence of attacks. As considerably more nerve fibres run from the cortex back to the lateral geniculate nucleus than the ascending geniculostriate pathway, they hypothesised that recurrent migraine episodes might cause cortical damage which in turn causes pre-cortical deficits. Chronicle and Mulleners (1994) suggested that cerebral ischaemia occurs in migraine and that this results in long-term damage to GABA-ergic cells in the visual cortex, which are especially sensitive to hypoxia.

Pupil anomalies and migraine

Lance (1993) has suggested that migraine could be viewed as a derangement of autonomic monoaminergic function. If so, then pupil dysfunction should be a feature of the migraine headache. However, the issue is confused by Rubin et al. (1985) who found that any difference in pupil responses between migraine sufferers and controls can be attributed, at least in part, to differences in personality. They claim that the migraine personality is more neurotic and depressive, and so responds emotionally in a different way to non-migraine controls. This, they claim, can affect the pupil responses as emotional factors are related to the autonomic nervous system.

Whilst the pupil abnormalities associated with migraine headache are often subclinical, there is some good evidence that such pupil anomalies can be unmasked by experimental procedures. Often this has involved the use of pharmacological agents to elicit different responses in migraine and non-migraine sufferers and this research is reviewed below.

Sympathetic hypofunction

Fanciullacci et al. (1977) have shown greater pupil dilation from instillation of phenylephrine and a reduced pupil dilation from the instillation of fenfluramine in idiopathic headache, as compared with controls. They concluded that this showed a super-sensitivity of iris adrenergic receptors in idiopathic headache. Herman (1983) has shown that anisocoria exists in both migraine and cluster headache sufferers but by only a mean of 0.8 mm. Gotoh et al. (1984) found sympathetic hypofunction in migraine sufferers during headache-free periods with a variety of neurological tests. Rubin et al. (1985) have shown that 70% of migraine sufferers in the inter-ictal phase have deficient sympathetic innervation of the dilator pupillae as compared with controls if challenged by a cold compress. Drummond (1987) compared the pupil diameter of the headache side and non-headache side in migraine sufferers, tension headache sufferers and non-headache controls. He showed that pupil diameter was smaller on the side of the headache both during headache and during headache-free periods in patients who habitually had headache on the same side of the head. Drummond (1990) has shown that facial temperature and pupil responses show a sympathetic deficit in migraine sufferers. The facial temperature was asymmetric and associated with the side of headache during a headache attack but not between attacks. In contrast, pupil diameter was smaller on the usual side of headache both during the headache and during the headache-free interval.

De Marinis (1994) stated that the evidence was so strong that pharmacological tests of the pupils could be used to differentially diagnose different forms of idiopathic headache. De Marinis et al. (1998) used pharmacological pupillary tests to investigate the oculosympathetic system in patients diagnosed as having migraine without aura. In contrast to the findings of Drummond (1987, 1990), De Marinis et al. claimed that the oculosympathetic hypofunction was not related to headache side and was temporally related to the migraine attack, being absent after 15 days. Battistella et al. (1989) showed that this sympathetic hypofunction existed in children with migraine but to a lesser extent which suggests a progression of the sympathetic hypofunction from childhood into adulthood.

Parasympathetic deficits

Purvin (1995) described a case of a 46-year-old woman who had suffered migraine headaches for the previous 20 years. Following one attack, she developed Adie's tonic pupil in one eye. He stated this could be caused by an unusually prolonged migrainous vasospasm leading to local ischaemia of the posterior lateral ciliary artery supplying the ciliary ganglion.

Overall considerations of the pupil and migraine

The evidence for a sympathetic hypofunction in migraine is strong although authors disagree on whether it persists in the headache-free period and if it is related to the side of the habitual headache.

The evidence of Adie's tonic pupil relates to one case study which although detailed is not good evidence and may represent a unique patient event rather than a general trend for all migraine sufferers.

Evans and Jacobson (2003) recently presented a case study of transient anisocoria in a migraineur and suggested that migraine headache can exaggerate physiological anisocoria and that in their case there were no sympathetic or parasympathetic deficits.

Pattern glare/visual stress and its relief with colour

Some people will report visual perceptual distortions (illusions), eyestrain, and headaches when viewing patterned stimuli. This has been termed ‘patterned glare’ (Wilkins and Nimmo-Smith, 1984) and more recently ‘pattern glare’ (Evans and Drasdo, 1991). Table 3 summarises the feature of patterns that cause pattern glare. When the symptoms of pattern glare are present in everyday life then this is called visual discomfort or visual stress. The early literature included several references to the anomalous visual effects of such patterns (e.g. Purkinje, 1823; Brewster, 1832) and by the 1960s and 70s these effects were being used in the art world, in a movement called ‘Op Art’.

Table 3.  A summary of the features which make geometric patterns most likely to produce an epileptic response
Contrast energy concentrated within one orientationWilkins et al. (1979)
The length of line is longWilkins et al. (1979)
High luminance, high contrastWilkins (1995, p. 17)
Square wave gratingSoso et al. (1980)
Increased size of patternWilkins et al. (1979)
Spatial frequencies between two and four cycles per degreeWilkins et al. (1979)
Pattern direction is reversed 10–20 times a secondWilkins (1995, pp. 31–34)
Binocular rather than monocular viewingJeavons and Harding (1975),  Wilkins et al. (1979, 1980).
Pattern presented in the visual hemi-field that corresponds to the side of the patients cortex that is most easily excitedWilkins et al. (1980)Soso et al. (1980)Binnie et al. (1981)

Wade (1978) listed the visual phenomena exploited in op-art and included afterimages, Hermann grid effects, Gestalt grouping principles, blurring and movement due to astigmatic fluctuations in accommodation, scintillation and streaming (possibly due to eye movements) and visual persistence. Symptoms produced from such visual phenomena can range from ‘unpleasantness’ to producing epileptic fits in susceptible individuals.

Wilkins (1995) summarised the various effects that normal subjects perceive when viewing a striped pattern as follows: red, green, blue, yellow, blurring, bending of the lines, shadowy shapes amongst the lines, shimmering of the lines, flickering of the lines, nausea, dizziness and pain. Wilkins (1995) suggested that if a person suffered from two or more of these illusions when looking at a striped pattern then they were more sensitive than average, should avoid looking at such a pattern for a long time, and could be diagnosed with visual stress. Conlon et al. (2001) showed that her patients with visual stress reported most perceptual distortions with a grating of 4 cycles per degree but that patients with little or no visual stress still had perceptual distortion but at a much higher spatial frequency of 12 cycles per degree. A commercially available test is now available for pattern glare/visual stress, which takes advantage of this (IOO Sales Ltd, London, UK).

Mechanism of visual stress

Wade (1977) had earlier suggested three mechanisms that could explain some of these illusions: physiological fixation instability, accommodative changes and the chromatic aberrations of the eye. Zanker (2002) agreed from a computational viewpoint, and claimed that the illusions could have an almost trivial solution in terms of small involuntary eye movements leading to image shifts that are picked up by motion detectors in the early motion system. Wilkins (1995) suggested that these explanations were not adequate to explain the illusions and agreed with Georgeson (1976, 1980) that the illusions had a structure that could more readily be attributed to inhibitory connections in the visual cortex.

A detailed paper by Wilkins et al. (1984) was the seminal work in establishing a neurological basis for visual stress. These authors demonstrated in a number of experiments that the illusions were produced by pattern glare, showed that if the number of illusions was more than two then the patients was more likely to have visual stress, that the illusions produced were lateralized with other symptoms and that the same stimuli that produced pattern glare also produced epileptiform EEG activity in susceptible individuals. Unlike the epileptic response to patterns, the illusion response to patterns does not spread widely across a hemisphere probably because the processing is more focal. This focal (localised) response does not spread widely because the cortex is not sufficiently hyper-excitable (Wilkins, 1995).

It should be noted that this visual stress is conceptually different to the sensory visual deficits discussed earlier (e.g. Coleston et al., 1994; McKendrick and Badcock, 2003). Visual stress seems to be a manifestation of cortical hyperexcitability resulting in a visual trigger for migraine (Wray et al., 1995), eyestrain and visual perceptual distortions. It can be thought of as a visual component to the migraine brain's over-sensitivity to environmental triggers (Welch, 2003). In contrast, the sensory visual deficits seem more likely to be a consequence of neural damage caused by migraine over a number of years. In contrast with this view, Shepherd (2000) reported a correlation between pattern glare and contrast sensitivity and supra-threshold contrast scaling in migraine, but did not find any overall effects due to migraine duration or frequency of migraine attacks.

Pattern glare, visual stress and headache

Interestingly, this illusion response to patterns has a relationship to headache frequency. Wilkins et al. (1984) showed that there is a direct correlation between the number of headaches reported and the number of illusions seen whilst viewing a striped pattern of about 4 cycles per degree. Unfortunately, several of the experiments cited in this paper excluded migraine sufferers. However, experiment 7 in this paper did show that migraine sufferers perceive more illusions with a pattern glare stimulus than tension headache sufferers. The correlation between migraine headache and pattern glare only held when the pattern design was within the epileptogenic range and did not hold when other symptoms such as back pain were discussed. For these reasons Wilkins and his team suggested that the finding could not be attributed to response bias.

People are more susceptible to illusions on days when they have headaches (Nulty et al., 1987). In addition, people show more aversion to striped patterns if they are headache sufferers particularly if the headaches are migraines. Marcus and Soso (1989) showed that when viewing epileptogenic striped patterns, 82% of migraine suffers demonstrated aversion whist only 18% of a control group did so. There was no difference between migraine with and without aura. If the illusions appear more pronounced on one side of a pattern then that patient is more likely than others to experience head-pain that is consistently lateralized (Wilkins et al., 1984).

Aurora et al. (1998, 1999) used transcranial magnetic stimulation to demonstrate that the visual cortex is indeed hyperexcitable in people who suffer from migraine. Huang et al. (2003) used functional MRI in patients who had migraine with aura to show that square-wave gratings that produced pattern glare did induce a hyperneuronal response in the visual cortex.

The relief of pattern glare and visual stress with colour

Colour preference can be related to psychology (red for danger and excitement or blue being a calming colour) or to ocular pathological conditions such as the brunescence of nuclear sclerotic cataract producing yellowing vision. Some individuals may wear tinted lenses due to neuroses (Howard and Valori, 1989). Other people with certain neurological disorders, such as dyslexia, migraine or epilepsy can be helped by using individually prescribed coloured filters (Lightstone, 2000), most likely through their effect on pattern glare/visual stress (Wilkins, 2003). Griffiths (2001) stated that measuring colour preference should be part of a routine optometric examination and produced a six-colour system to do this. However, the randomised controlled trials of Wilkins et al. (1994, 2002) and Robinson and Foreman (1999) suggest that a greater degree of precision is required and this is supported by recent data alluded to by Wilkins et al. (2004). The Intuitive Colorimeter (Wilkins and Sihra, 2000) is commonly used for this purpose in the UK.

The use of individually prescribed coloured filters for children with reading difficulties has been described as Meares-Irlen syndrome, which is likely to be a manifestation of visual stress. This subject has recently been reviewed by Evans (2001) and Wilkins (2003). The benefit from coloured filters is not solely attributable to placebo effects (Wilkins et al., 1994; Robinson and Foreman, 1999); conventional optometric or orthoptic anomalies (Evans et al., 1995, 1996b; Scott et al., 2002); spatio-temporal contrast sensitivity functions (Simmers et al., 2001); or a magnocellular deficit (Evans et al., 1995, 1996a; Simmers et al., 2001). Instead, the benefit from coloured filters is most likely attributable to pattern glare (Evans et al., 1995, 1996a) which can be caused by lines of text (Wilkins and Nimmo-Smith, 1984). Deficits of visual attention in some people with reading difficulties might make them particularly sensitive to pattern glare (Evans, 2001). As people with migraine are particularly sensitive to pattern glare, it is not surprising that migraine-like headaches are prevalent in children with reading difficulties who benefit from precision-tinted lenses (Evans et al., 1996b).

It is argued that coloured filters change the distribution of the firing pattern within the visual cortex and, since cortical hyperexcitability may vary locally within the visual cortex, individually prescribed coloured filters are an effective treatment (Wilkins, 1995; Wilkins et al., 2004). This hypothesis has been supported by recent work showing that the representation of colour in the visual cortex follows topographic maps (Xiao et al., 2003).

Chronicle and Wilkins (1991) have found that the visual stress of migraineurs is determined by the colour of the illuminating light, tending to avoid red illumination. In contrast, Good et al. (1991) showed that migraine frequency was reduced in children who wore rose tinted spectacles compared with a blue tint. If the tint is prescribed precisely and individually, then the reduction in symptoms with colour is not due to alterations in binocular function or refraction (Evans et al., 1996a,b, 2002).

Wilkins et al. (2002), in a double-masked randomised controlled study, compared the effectiveness of precision-tinted ophthalmic lenses in the prevention of headache in migraine sufferers. They showed with headache diaries that headache frequency was significantly lower when a precise optimal tint was worn when compared with a suboptimal tint used as a control. The group was a selected group of migraine sufferers that found colour helpful and their optometric characteristics were described by Evans et al. (2002). Evans et al. (2002) showed that pattern glare symptoms of visual stress were reduced with a precisely selected colour of tinted spectacles. However, this reduction in visual stress was not significantly different from that produced by only a slightly different tint that was used as a control.

To conclude, certain visual stimuli produce visual stress. Migraine sufferers are particularly susceptible to visual stress and it can be reduced with precision-tinted spectacles. By reducing visual stress in migraine sufferers, migraine frequency can be reduced.


Headache is a common symptom reported by patients who consult optometrists (Barnard and Edgar, 1996). As migraine accounts for as many as 54% of all headaches (Leone et al., 1994) this suggests that optometrists are likely to encounter patients with migraine very commonly.

Some authors have argued that optometric anomalies are a trigger for migraine (Snell, 1904; Turville, 1934; Wilmut, 1956; Waters, 1970; Griffin, 1996; McKendrick et al., 1998). In contrast, the medical literature is sceptical about the role of visual factors in headaches and migraine (Lyle, 1968; Headache Classification Sub-Committee of the International Headache Society, 2004).

In the current climate of clinical governance, there is a need for evidence-based research to guide optometrists as to the role they can play, if any, in managing some cases of migraine. This review has critically examined the evidence that correlates migraine headache and optometric factors. Each optometric correlate of migraine can be segregated into either a visual sensory or visual motor factor, and Table 4 summarises the evidence. With the exception of the sensory visual factor of visual stress/pattern glare, and sympathetic hypofunction, the evidence correlating optometric factors with migraine is generally poor.

Table 4.  Summary of visual correlates of migraine. The visual correlates have been divided into sensory and motor correlates. Levels of evidence based on the Centre for Evidence Based Medicine (Oxford, UK) (1999) recommendations have been assigned (where 1 is high evidence and 5 is low evidence) by the present authors
FactorAssessment (clinical or research)Evidence (levels 1–5)Relevance (correlate, cause, treatable?)
Visual sensory factors
 Pupil (sympathetic hypofunction)Research tests
Routine clinical tests
Level 1bCorrelate
 Pupil (parasympathetic hyperfunction)Research testsLevel 4Correlate
 FlickerRoutine clinical testsLevel 2bCorrelate
 Visual stress/pattern glareRoutine clinical testsLevel 1bCorrelate
Visual motor factors and refractive error
 ExophoriaRoutine clinical testsLevel 4Correlate
 HyperphoriaRoutine clinical testsLevel 4Correlate
 Refractive errorRoutine clinical testsLevel 4Correlate

Thus, it appears that there is acceptable evidence in the literature to suggest that both cortical hyperexcitability (as demonstrated by pattern glare) and peripheral neurological defects (as demonstrated by the sympathetic hypofunction with pupil responses in migraine sufferers) are associated with migraine headache. The cortical and peripheral theories are not incompatible. It could be suggested that cortical hyperexcitability is an interictal status that leads to pattern glare and that this sensory visual factor is a trigger for migraine. This is consistent with many other authors who have found that migraine can be triggered by certain visual stimuli. It seems that precision-tinted lenses might be one method of minimising the impact of visual triggers for migraine headache sufferers. Additionally, pre-cortical changes to the visual system (such as the pupil changes and some of the visual field anomalies found) may be a long-term consequence of the neuro-vascular interactions of migraine headache.

Acknowledgment and ethical statement

The authors are members of EyeNET, the primary care eye research network supported by the London NHS Executive. The present work was funded by EyeNET. The views expressed in this publication are those of the authors and not necessarily those of the NHS Executive. The Pattern Glare Test cited in text is marketed by IOO Sales Ltd, who raise funds for the Institute of Optometry which is an independent registered charity. One of the authors (BJWE) receives a small ‘Award to Inventors’ relating to the Pattern Glare Test.