Abstract
The aim was to to determine if the visual aura of migraine is altered by disease of the afferent visual pathways and if visual aura changes are associated with pre- or postgeniculate lesions. Functional neuroimaging during migraine demonstrates primary visual/extrastriate cortex as an anatomical substrate of visual aura. Neuro-ophthalmological records (including kinetic and static perimetry) of 25 patients with visual loss and typical aura with or without migraine headache were reviewed. Twenty-five patients (16 women, nine men) (mean age 59.8 years) with typical aura had visual loss from pregeniculate (72%) or postgeniculate lesions (28%). Eight patients (four postgeniculate cerebrovascular accidents or arteriovenous malformations, two lifelong optic neuropathy/ retinopathy, one childhood ocular trauma, one anisometropic amblyopia) reported absence or alteration of visual aura. Postgeniculate lesions were significantly associated (P = 0.017) with visual aura changes. The association of postgeniculate lesions with altered auras points to a postgeniculate effect on aura appearance (consistent with functional neuroimaging findings). Although statistically significant, this series' association of postgeniculate disease and aura changes is even more robust (P = 0.0002) when structural changes of ocular dominance columns are posited in three patients with optic neuropathy, retinopathy and keratopathy of congenital or childhood origin.
Introduction
Prior to functional neuroimaging, the modern conception of the cerebral origin of migraine visual aura (MVA) was grounded in clinical observations ranging from Lashley's meticulous mapping of the shape and drift of his own migrainous scotomas (1) to Rose's account of recurring MVA in a woman 36 years after bilateral enucleation (2). Subsequent findings of single photon emission tomography, functional magnetic resonance imaging (MRI) and magnetoencephalography (3–5) have buttressed and refined the conclusions drawn from clinical phenomenology and point to striate/extrastriate cortex as the anatomical substrate of MVA (positing a primarily neuronal, as opposed to vascular, mechanism). This interrelationship of MVA and visual cortex begs the question as to whether anatomical disruption of the central (and, in particular, postgeniculate) visual pathways can modify MVA. This study examines the characteristics of MVA in the special circumstances of 25 patients with migraine and visual loss. (The terms ‘teichopsia’, ‘scintillating scotoma’ and ‘fortification spectra’ will be used interchangeably with MVA and will underscore the frequently but not exclusively geometric appearance of MVA that is readily identified by migraineurs.) In migraineurs with visual loss the examiner encounters three kinds of usually episodic visual symptomatology: (i) purely negative (e.g. migrainous, scotoma or field cut/scotoma arising from lesion of afferent visual pathways); (ii) purely positive (e.g. release hallucinations, Charles Bonnet syndrome (6), or advancing margin of non-scotomatous teichopsia); and (iii) combined positive and negative (e.g. the scintillating scotoma preceding headache). By delineating the visual world of the visually impaired migraineur, this study documents both the interplay of positive and negative visual symptoms and the dependence of MVA on the integrity of cerebral visual pathways.
Methods
Charts of 25 patients with visual loss and typical aura with migraine headache [International Headache Society (IHS) 1.2.1] and/or typical aura without headache (IHS 1.2.3) were reviewed (7). All patients were seen in the office of a university neuro-ophthalmologist from 2002 to 2007 and underwent complete neuro-ophthalmological examination including Snellen visual acuity and Goldmann kinetic perimetry. During prolonged interviews, all patients provided detailed descriptions of their recurrent MVA (and, in 13 patients, release hallucinations). Fortification spectra (with or without scotoma) or migrating photopsias were used as inclusion criteria for this study. Criteria for visual loss included visual acuity worse than 20/25 in at least one eye and/or abnormal Goldmann or static perimetry. Lesion localization was facilitated by MRI or computed tomography of the head in all patients (with the exception of a single patient with anisometropic amblyopia).
Results
Twenty-five patients (16 women and nine men) with typical aura with migraine headache (16%), typical aura without headache (52%), or both (32%), and visual loss from pregeniculate (72%) or postgeniculate (28%) lesions (see Table 1) were evaluated. With the exception of one patient whose MVA was characterized by transient migrating photopsias associated with hemicranial headache, all patients described MVA with geometric morphology, e.g. ‘zigzags,’ ‘saws,’ ‘picket fences,’ ‘lightning’ and ‘teepees’, or expanding ‘heat waves’. Ages varied from 19 to 87 years (mean 59.8 years). Onset of visual loss varied from infancy to 67 years (Tables 2 and 3).
Clinical diagnoses in 25 migraineurs with visual loss
Migraineurs with visual loss and altered aura
OD, right eye; OS, left eye; OU, both eyes; WNL, within normal limits.
Migraineurs with visual loss and unchanged aura
Eight patients (Table 2) reported alteration of MVA or absence of MVA from one eye or part of the visual fields. Aetiologies of visual loss included postgeniculate cerebrovascular accidents (CVA) or arteriovenous malformations (four patients), lifelong optic neuropathy or retinopathy (two patients), childhood penetrating ocular trauma (one patient) and anisometropic amblyopia (one patient). Postgeniculate lesions were significantly associated (P = 0.0169, Fisher's exact test) with MVA changes. If patients (three with and one without altered auras) with lifelong or childhood-acquired pregeniculate disease are reassigned to the group with postgeniculate lesions (by virtue of secondary ocular dominance column abnormalities) (8), the strength of the association is increased (P = 0.0002, Fisher's exact test).
Binocular visual loss or Snellen visual acuity equal to or worse than 20/50 in at least one eye was not significantly associated with modified MVA. In addition to MVA, 13 patients described release hallucinations ranging from elementary (nine patients), e.g. photopsias, to complex (four patients), e.g. ‘clothed rabbits’. (6) Patients distinguished MVA from release hallucinations, and release hallucinations were not significantly associated with altered MVA.
Discussion
For nearly 50 years MVA has been attributed to cortical spreading depression (CSD) occurring in visual cortex (9). Subsequently, Hadjikhani's functional neuroimaging has directly linked MVA to the retinotopic organization of the visual cortex (4). Although CSD has never been recorded in uninjured human neocortex, MRI mapping of blood oxygenation level-dependent (BOLD) events during MVA persuasively argues for MVA as a primarily neuronal and CSD-driven phenomenon.
This study has demonstrated that pathology of the visual cortex and postgeniculate pathways in migraineurs is significantly associated with complete cessation of MVA (patients 1 and 5 initially had aura without headache and patient 1 also had persistent common migraine), altered MVA appearance (patient 2), and absence of MVA from monocular or binocular field defects (patients 3, 4, 6, 7 and 8). The precise anatomical origin of teichopsia is not known with certainty. Although Richards has speculated that migrainous fortifications arise from the columnar organization of visual cortex neurons specializing in ‘detection of lines of a particular length and orientation’ (10), Penfield's bipolar electrode stimulation of cortex was unable to reproduce ‘zigzag outlines of migraine images’ (11).
The present lesion-based study reaffirms that ‘to destroy a delicate instrument is not the best way of studying its function’, (12) but this study does establish that integrity of the postgeniculate pathways is a critical factor in the appearance of MVA and, by extension, propagation of CSD. Further research is requisite to determine how CSD can be altered, inhibited or diverted to clinically ‘silent’ brain regions (13) by lesions of the geniculocalcarine white matter, occipital and parietal cortices, and ocular dominance columns found in this study.
Migraineurs with visual loss witness multiple positive and negative visual phenomena and they can readily distinguish teichopsia from release hallucinations. For example, patient 6 (with a right occipital lobe CVA) experienced baseline left visual field loss, and at times painless MVA (‘zig-zag’) outside the scotoma would occur simultaneously with release hallucinations (‘3 dimensional yellow moving “Xs”’) within the region of field loss. Other patients report the ‘paradox’ of teichopsia (usually associated with the cortical ‘signature’ of binocular symptomatology) affecting only one eye. Patients 4 and 7 with a history of monocular visual loss of childhood onset and patient 3 with congenital hyaline bodies of the optic disc experienced monocular teichopsia. The complexity of their positive visual symptomatology disallows the diagnosis of retinal migraine (7), and monocular perception of a visual symptom of cortical origin in an amblyope suggests that CSD can be inhibited by amblyopia-induced changes of ocular dominance columns subserving one eye (8).
Currently, more is known about the pathophysiology than the anatomical origin of MVA. Despite lack of precise electrophysiological delineation of CSD in humans (13), CSD is the prevailing hypothesis for the cause of migraine (2, 4, 13). Visual cortex has been posited as the site of MVA since Lashley's self-report (1), but the evolving concept of MVA anatomy is illustrated by Hadjikhani's evidence for visual aura arising in the extrastriate area, V3A (4). The present study does not precisely localize MVA, but demonstrates that postgeniculate lesions alter MVA. Moreover, it affirms that postgeniculate disease changes not only the patient's view of the veridical and external world, e.g. hemianopsia, but also the perception of internally generated visual phenomena, e.g. MVA. Similarly, other kinds of internal visuospatial perception such as mental representation of the environment have been shown to be altered by unilateral hemispheric and postgeniculate lesions (14). Lacking a comprehensive ‘atlas’ or model (15) of migraine auras, careful description of the visual world of migraineurs both with and without visual loss is necessary to refine the diagnosis of migraine and elucidate the basis of MVA (16).
Footnotes
Competing interests
None to declare.
Acknowledgements
Presented at the 132nd annual meeting of the American Neurological Association, Washington, DC, USA.