Brain Representation of Hemifield Stimulation in Poststroke Visual Field Defects
Background and Purpose— Plasticity in extended, parallel, or reciprocal operating networks is well recognized. Changes in neuronal activity after lesions to distinct localized structures, such as the primary visual cortex, are less well characterized. We investigated the cortical reorganization in patients with poststroke visual field defects using blood oxygen level–dependent functional MRI.
Methods— Brain activation was measured in 7 patients with a single occipital cortical lesion and partially recovered hemianopia and in 7 age-matched control subjects. Differences in activation between rest and visual hemifield stimulation were assessed with statistical parametric mapping (SPM’99).
Results— In normal subjects, significant activation was found in the contralateral primary visual cortex and bilaterally in the extrastriate cortex. During hemifield stimulation of the unaffected side of stroke patients, a similar pattern was found compared with that seen in control subjects. During stimulation of the hemianopic side, bilateral activation was seen within the extrastriate cortex, stronger in the ipsilateral hemisphere. The primary visual cortex was not significantly activated in either hemisphere during stimulation of the hemianopic side.
Conclusions— Visual field defects after stroke are associated with bilateral activation of the extrastriate visual cortex. This pattern of activation indicates altered neuronal activity in the visual system. Further investigation is necessary to determine the relationship between functional reorganization and recovery of lost visual function after poststroke hemianopia.
Hemianopia is the most common visual field abnormality, accounting for >70% of all visual field defects after infarction of the posterior cerebral artery.1 Patients with hemianopia may be severely disabled in terms of activities of daily living such as indoor and outdoor mobility, reading, or spatial orientation when overlooking unfamiliar places or rooms.2 Spontaneous recovery of hemianopia may occur, but most patients are left with some residual visual field defects.3 The unilateral representation of visual fields may explain why full restitution of visual field defects is rarely seen.
Functional brain reorganization has been proposed to explain recovery of impaired function after the resolution of peri-infarct edema and reperfusion of the ischemic penumbra. Important insights into these reorganizational processes that may mediate recovery after brain damage are derived from studies of motor4,5⇓ and language function.6 The unilateral distinct retinotopic representation in the primary visual cortex, however, may preclude reorganization similar to other neural systems that are organized in extended, overlapping, and parallel or reciprocal processing networks. In his early cytoarchitectonic studies, rodmann7 subdivided the visual cortex in the primary visual cortex (striate cortex, area 17), the surrounding peristriate cortex (area 18), and the extrastriate cortex (areas 19 and 37). More recent studies in animals propose a distinction between the primary visual (striate) cortex, V1, and the extrastriate visual cortex. The latter comprises >20 functionally and anatomically distinct areas, among which V2, V3, VP, V4, and V5 are best characterized.8 Plasticity within the visual cortex after striate cortical lesions has been documented in both animals and humans.9,10⇓ To date, the anatomic areas that may be involved in plasticity of patients with poststroke visual field defects, however, have received little attention and are poorly characterized. This topic, however, is important to gain more insight into the function of neural systems that are represented in distinct localized areas.
The aim of this study was to investigate patterns of brain activation in patients with poststroke visual field defects after a single occipital cortex stroke. We used a hemifield stimulation paradigm to segregate brain activation during stimulation of the hemianopic side and the unaffected side and compared patient data to those of age-matched control subjects using group and individual analyses.
We studied 7 stroke patients and 7 control subjects. Stroke patients were selected from the Neurological Therapy Center, an outpatient rehabilitation facility for stroke patients. Inclusion criteria were a first stroke, minimum interval from stroke onset to study of 6 months, lesion in the striate cortex causing homonymous hemianopia, and the ability to maintain fixation for at least 6 minutes. Patients with previous strokes or strokes elsewhere than in the occipital lobe were not included. Patients with neglect, those with any other visuospatial impairment, and those with contraindication for MRI were excluded. Visuospatial neglect was excluded by use of the line bisection test,11 line cancellation test,12and drawing of a clock.13 MRI of the brain and computed perimetry were performed in all patients before blood oxygen level–dependent functional MRI (fMRI; Figure 1). Seven age-matched (52.1±8.8 years) control subjects (2 women) were recruited through local advertisements. All control subjects were healthy volunteers with no history of neurological or psychiatric disease. Before each fMRI, a full neurological examination was obtained for each subject. All patients gave written, informed consent. This study was approved by the Review Board of the Ethics Committee of the University of Essen.
fMRI Image Acquisition
Visual stimuli were projected with a laptop computer onto a 154×115-cm2 screen placed 90 cm in front of the gantry aperture. During rest, the screen remained otherwise gray. During activation, a black-and-white checkerboard pattern reversal was used to stimulate each hemifield separately (Figure 2). The size of hemifield stimulation was 42.5×36°. The size of checks was 3.4 cm. Subjects were given clear instructions to fixate on a red point in the center of the screen during both conditions. All subjects were experienced in performing the task during hemifield stimulation and had practiced repetitively in the laboratory. Before fMRI, all subjects were acclimated to the scanner and the fixation task in a prescan session. Fixation was not monitored in the scanner but was presumed to be similar to behavioral performance measured in the laboratory, as in precedent studies of functional imaging of the visual system.14,15⇓ The light intensity of the gray screen during baseline and of the contralateral hemifield during stimulation was adjusted to be equal to the light intensity of the black-and-white checkerboard to avoid unintentional stimulation caused by changes in brightness of the environment. Luminance of the projector was 1100 ANSI-Lumen with 400:1 contrast.
With the use of a block design, 120 volumes of 48 contiguous axial fMRI slices (3.3×3.3×4.0-mm spatial resolution) were acquired with echoplanar imaging (repetition time/echo time, 4800/60 ms; flip angle, 90°; Siemens Sonata, 1.5 T; gradients, 40 mT) covering the whole brain (60 volumes for each condition; AB...AB, starting with the unaffected side in patients and the left side in control subjects). The angle (along the horizontal meridian) between the fixation point and the lateral edge of stimulation with high-contrast checkerboards pattern was 40.5° for control subjects and patients (each square, 60 arc/min). The pattern reversed at 5 Hz.
Raw data were processed and analyzed with statistical parametric mapping (SPM’99, Wellcome Department of Cognitive Neurology, London, UK).16 Functional images were realigned with sinc interpolation and then normalized to the standard stereotactic space corresponding to the template from the Montreal Neurological Institute (http://www.mrc-cbu.cam.ac.uk/Imaging/mnispace.html). Bilinear interpolation was applied for normalization. No distortion from the occipital infarction was noted on normalized images. Finally, x, y, and z spatial coordinates were transformed to the 3-dimensional anatomic space according to Talairach and Tournoux.17 Functional images were smoothed to accommodate intersubject variability of brain anatomy with an isotropic Gaussian kernel of 12 mm. Data were modeled by use of a boxcar function convolved with the hemodynamic response function. Because 6 patients had left-sided lesions, data of 1 patient with a right-sided lesion were flipped across the midsagittal plane, a procedure that has been applied previously in group studies of functional imaging.4,5⇓ Thus, for the group analyses, the hemianopic side was always right.
Assessment of significant signal changes between rest and visual stimulation was performed with SPM’99.18 Statistical inferences were based on group effects by use of a fixed-effects models.19 A voxel-by-voxel comparison according to the general linear model and t statistics was used to calculate differences in activation between hemifield stimulation and rest in patients and normal subjects (group study). The same model was used for comparisons between patients and control subjects (multigroup study). The resulting statistical parametric map was subsequently used to assign probability values (to voxels and clusters), which were corrected for multiple comparisons applied for the whole brain. Significant differences were defined at a corrected value of P<0.01. All patients were also analyzed individually. For the single-subject analysis, the same steps for data processing were used. The purpose of the individual analysis was to compare the individual activation pattern among subjects and with the group results.
A summary of the study population is given in Table 1, and MRI scans and corresponding visual fields are given in Figure 1. In all subjects, the vascular lesion involved the striate cortex, causing dense contralateral homonymous, macula-sparing hemianopia at the time of stroke. Initially, all patients were severely impaired in their mobility because of missing obstacles and experiencing difficulties in noticing other persons in the hemianopic field. All reported paralexia with difficulty reading ensuing words fluently. At study time, 17.3±7.3 months (mean±SD) after stroke, patients reported that visual field defects had improved slowly over a period of at least 4 months, but all had persisting partial visual field defects. Patients 1 and 5 had recovered visual function in the upper quadrant (right in patient 1, left in patient 5); patients 2, 3, and 7 regained only a little visual function; and patients 4 and 6 recovered visual function in the right lower quadrant (Figure 1). All patients were able to walk outdoors independently, to care for themselves without any assistance, and to read.
Comparison of Right Coronary Blood Flow During Hemifield Stimulation and Rest in Normal Subjects
During stimulation of either hemifield, very similar patterns of activation were seen, with maximum increase in regional cerebral blood flow (rCBF) in V1 of the contralateral primary visual cortex (Z>9; see Table 2 and Figure 3). This area of maximum rCBF changes also extended to the peristriate and extrastriate cortexes covering Brodmann areas 18 and 19 (Z>9). Weaker but significant activation was also found in the peristriate and extrastriate cortexes ipsilateral to the side of stimulation (Z=7.5; Z=7.7 during stimulation of the right side).
Comparison of rCBF During Hemifield Stimulation and Rest in Patients With Hemianopia
Stimulation of the Unaffected (Left) Hemifield
During stimulation of the unaffected (left) side, an area of maximum activation was again observed in the contralateral primary visual cortex V1 (Z>9; Figure 4, top). As in control subjects, activation extended to the peristriate and extrastriate cortexes. Similar to normal subjects, only peristriate and extrastriate structures corresponding to areas 18 and 19 were activated ipsilaterally.
Stimulation of the Hemianopic (Right) Hemifield
In contrast, during stimulation of the hemianopic side, maximum activation was found in the (unaffected) hemisphere ipsilateral to the side of hemianopia (Figure 4, middle). The most significant activations were observed in Brodmann areas 18 and 19 and extended exclusively within the peristriate and extrastriate cortexes (Z >9). The primary visual cortex V1 was not activated. Less significant increases in rCBF were also found in the contralateral (left) hemisphere that was affected by the stroke, again only in peristriate and extrastriate cortical regions (Table 3). No activation was found in the primary visual cortex of either hemisphere.
Comparisons of rCBF Between Stroke Patients and Normal Subjects
Areas that were significantly more activated in stroke patients compared with normal subjects were identified by subtracting increases in rCBF during hemifield stimulation of the right side versus rest in normal subjects from those in patients (patients versus control subjects, multigroup study analysis). Significant increases in rCBF were found in Brodmann area 18 of the peristriate cortex close to the junction with the striate cortex of the ipsilateral, unaffected hemisphere (x, y, z coordinates: 16, −90, 0; Z=6.9; Figure 4, bottom).
During stimulation of the unaffected (left) hemifield, maximum increases in rCBF were observed in the contralateral primary visual cortex V1 in each individual (Table 4⇓). All except subject 3 activated the ipsilateral extrastriate cortex. During stimulation of the hemianopic (right) side, the activation pattern was more variable, but again activation of the primary visual cortex was not observed in any of the patients (Table 4⇓). The single patient with a right-sided occipital infarct (patient 5) did not show a distinct pattern of activation.
The main finding of this study is the ipsilateral activation of the peristriate and extrastriate visual cortexes during stimulation from the side of the visual field defect. This pattern of activation was consistent in all single and multisubject analyses and was confirmed in direct comparisons of stroke patients with normal control subjects.
In control subjects, the most significant increases in activation were found in the primary visual cortex V1 contralateral to the side of hemifield stimulation. Peristriate and extrastriate cortical activation occurred bilaterally. The weaker ipsilateral activation of the extrastriate cortex in the control group corroborates earlier studies that have described a very similar activation in response to hemifield stimulation in normal subjects with fMRI.20,21⇓ Transcallosal transfer may explain the ipsilateral extrastriate activation in response to unilateral presentation of visual information. Evidence for interhemispheric pathways is given by a study that combined visually evoked potentials with fMRI in an experiment of hemifield motion stimuli in normal subjects.22 During hemifield stimulation of the unaffected side in patients, maximum activation was again seen in the contralateral striate cortex V1. The activation of the extrastriate cortex was also similar to that in normal control subjects with stronger activation of the contralateral hemisphere and weaker activation of the ipsilateral hemisphere.
In contrast, during stimulation of the hemianopic side, maximum activation was found in the ipsilateral peristriate and extrastriate cortexes. The primary visual cortex contralateral to the visual field defect was not activated in any of our analyses. Lack of activation in contralateral V1 has been described previously and is best explained by the relatively large occipital lesions in our patients.23 Other studies have reported cortical activation maps in the primary visual cortex that corresponded to the clinical visual field defect.24 A study using 18-fluoro-2-deoxyglucose PET has demonstrated a reduction in the size of metabolic lesion and improvement in striate cortex metabolism only in patients with recovery of vision, not in those patients with persistent visual field defects.25 These data confirm that the damaged primary visual cortex has little or no evidence of neuronal activity, at least in those patients with persisting visual field defects. Because of the variability of anatomic boundaries across individual subjects,26 we cannot rule out weak activation of the striate cortex, although no V1 activation was found in the individual analyses.
The increased activation of the ipsilateral peristriate and extrastriate cortexes has not been described previously with fMRI. Several reasons may explain this observation. As in control subjects, this activation could result from preexisting transcallosal pathways. Transcallosal shift of visual information in response to focal injury may be enhanced by axonal sprouting and establishment of new connections in the visual cortex.27 Because restitution of visual field defects is usually limited and incomplete, an ipsilateral activation of the extrastriate cortex by transcallosal pathways may not be directly related to a restorative mechanism.
Another possible explanation for the observed bilateral activation is small deviations from central fixation that may stimulate the contralateral macula and directly activate V5. Because we did not control fixation during fMRI, we cannot rule out minor eye deviations from the fixation point. However, it is unlikely that such eye movements account exclusively for V5 activation, because no activation was found in ipsilateral V1 and activation of V5 was more significant in patients. An alternative hypothesis is a bilateral representation of extrastriate cortical areas. In contrast to V1, extrastriate cortical area V5 may extend into the opposite hemisphere.20,28⇓ Representation of 1 hemifield in both hemispheres implies a more complex anatomic input than traditionally assumed by entirely crossed visual pathways. It has further been shown that visual information related to fast motion can bypass the primary visual cortex to reach V5 via direct projections from the lateral geniculate nucleus.29 This hypothesis could explain why some recovery occurs after visual field defects but full restoration of visual fields is rarely seen. A third explanation would suggest that distinct visual areas may be capable of performing different functions at different times. Evidence for the existence of such multitasking neurons in the parietal cortex with potential for simultaneous visual, saccadic, and attentional function has been provided by animal studies.30 Furthermore, congenitally blind individuals show differential activation of the extrastriate cortex but not of the primary visual cortex during Braille reading relative to auditory reading.31 In accord with these findings are reports about blind raised monkeys that show reorganization of extrastriate and parietal areas.32,33⇓ After the previously sutured eyelids were reopened, 20% of cells studied in Brodmann area 19 responded exclusively to tactile stimuli.34 Area 17, in contrast, did not respond to tactile stimuli in this animal model. These data from human and animal studies suggest cross-modal reorganization in extrastriate cortex without primary visual cortical involvement. Blind sight, the ability to respond to stimuli placed within the scotoma, may have been present in some subjects, but this phenomenon cannot readily explain activation of the ipsilateral extrastriate cortex.
In conclusion, brain activation during unilateral stimulation from the side of visual field defect was confined to the extrastriate cortex and was more pronounced in the ipsilateral (unaffected) hemisphere. Because recovery from visual field defects is usually limited and incomplete, the association between this activation pattern and clinical recovery requires further investigation. Recently, longitudinal studies with PET have described dynamic changes in brain activation patterns over time in patients with recovery from hemiparesis.35,36⇓ Whether such ongoing plasticity is only a property of distributed overlapping networks, as in the motor system, or is also found in the visual cortex needs further study with correlation of longitudinal functional brain imaging data and behavioral changes.
We thank Isabelle Janssen for expert technical assistance during fMRI scanning.
Presented in part at the 26th AHA International Conference on Stroke and Cerebral Circulation, Ft Lauderdale, Fla, February 14–16, 2001.
- Received September 19, 2001.
- Revision received January 11, 2002.
- Accepted January 16, 2002.
- ↵Zihl J. Eye movement patterns in hemianopic dyslexia. Brain. 1995; 118(pt 4): 891–912.
- ↵Zihl J, von Cramon D. Visual field recovery from scotoma in patients with postgeniculate damage: a review of 55 cases. Brain. 1985; 108(pt 2): 335–365.
- ↵Nelles G, Spiekermann G, Jueptner M, Leonhardt G, Muller S, Gerhard H, Diener HC. Reorganization of sensory and motor systems in hemiplegic stroke patients: a positron emission tomography study. Stroke. 1999; 30: 1510–1516.
- ↵Brodmann K. Vergleichende Lokalisationslehre der Grosshirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues. Leipzig, Germany: Barth; 1909.
- ↵Eysel UT, Schweigart G. Increased receptive field size in the surround of chronic lesions in the adult cat visual cortex. Cereb Cortex. 1999; 9: 101–109.
- ↵Albert ML. A simple test of visual neglect. Neurology. 1973; 23: 658–664.
- ↵Heilmann KM, Watson RT, Valenstein E. Neglect and related disorders.In: Heilmann KM, Valenstein E, eds. Clinical Neuropsychology. New York, NY: Oxford University Press; 1992.
- ↵Kastner S, De Weerd P, Desimone R, Ungerleider LG. Mechanisms of directed attention in the human extrastriate cortex as revealed by functional MRI. Science. 1998; 282: 108–111.
- ↵Frackowiak R, Friston K, Frith C, Dolan R, Mazziotta J. Human Brain Function. San Diego, Calif: Academic Press; 1997.
- ↵Talairach J, Tournoux P. Coplanar Stereotaxic Atlas of the Human Brain. New York, NY: Thieme Verlag; 1988.
- ↵Tootell RB, Mendola JD, Hadjikhani NK, Liu AK, Dale AM. The representation of the ipsilateral visual field in human cerebral cortex. Proc Natl Acad Sci U S A. 1998; 95: 818–824.
- ↵Sorensen AG, Wray SH, Weisskoff RM, Boxerman JL, Davis TL, Caramia F, et al. Functional MR of brain activity and perfusion in patients with chronic cortical stroke. AJNR Am J Neuroradiol. 1995; 16: 1753–1762.
- ↵DeYoe EA, Carman GJ, Bandettini P, Glickman S, Wieser J, Cox R, Miller D, Nietz J. Mapping striate and extrastriate visual areas in human cerebral cortex. Proc Natl Acad Sci U S A. 1996; 93: 2382–2386.
- ↵Colby CL, Duhamel JR, Goldberg ME. Visual, presaccadic, and cognitive activation of single neurons in monkey lateral intraparietal area. J Neurophysiol. 1996; 76: 2841–2852.
- ↵Buchel C, Price C, Frackowiak RS, Friston K. Different activation patterns in the visual cortex of late and congenitally blind subjects. Brain. 1998; 121(pt 3): 409–419.
- ↵Rauschecker JP, Tian B, Korte M, Egert U. Crossmodal changes in the somatosensory vibrissa/barrel system of visually deprived animals. Proc Natl Acad Sci U S A. 1992; 89: 5063–5067.
- ↵Calautti C, Leroy F, Guincestre JY, Baron JC. Dynamics of motor network overactivation after striatocapsular stroke: a longitudinal PET study using a fixed-performance paradigm. Stroke. 2001; 32: 2534–2542.