Functional magnetic resonance imaging of the visual system

Functional magnetic resonance imaging of the visual system Atsushi Miki, MD, PhD, Grant T. Liu, MD, Edward J. Modestino, Mphil, Chia-Shang J. Liu, BA, Gabrielle R. Bonhomme, MD, Cristian M. Dobre, MS, and John C. Haselgrove, PhD Functional magnetic resonance imaging (fMRI), which is a technique useful for non-invasive mapping of brain function, is well suited for studying the visual system. This review highlights current clinical applications and research studies involving patients with visual deficits. Relevant reports regarding the investigation of the brain’s role in visual processing and some newer fMRI techniques are also reviewed. Functional magnetic resonance imaging has been used for presurgical mapping of visual cortex in patients with brain lesions and for studying patients with amblyopia, optic neuritis, and residual vision in homonymous hemianopia. Retinotopic borders, motion processing, and visual attention have been the topics of several fMRI studies. These reports suggest that fMRI can be useful in clinical and research studies in patients with visual deficits. Curr Opin Ophthalmol 2001, 12:423–431 © 2001 Lippincott Williams & Wilkins, Inc. Division of Neuro-Ophthalmology, University of Pennsylvania School of Medicine; and the Functional MRI Research Unit; the Children’s Hospital of Philadelphia; Philadelphia, Pennsylvania, USA. Correspondence to Dr. Grant T. Liu, MD, Division of Neuro-ophthalmology, Department of Neurology, Hospital of the University of Pennsylvania, 3400 Spruce St., Philadelphia, PA 19104, USA; e-mail: [email protected] Current Opinion in Ophthalmology 2001, 12:423–431 ISSN 1040–8738 © 2001 Lippincott Williams & Wilkins, Inc. Functional magnetic resonance imaging (fMRI) is a relatively new neuroimaging technique that can detect neural activation noninvasively in the human brain. Functional MRI is based on the assumption that local neuronal activity correlates with increases in regional cerebral blood flow [1]. The advantages of this method include: 1) high spatial and temporal resolution (ie, small areas of the brain can be sampled at relatively fast rates); 2) no ionizing radiation or exogenous contrast agent is necessary; 3) and many images can be obtained from a single subject because of 1 and 2. For these reasons, fMRI is superior to other functional imaging techniques, such as positron emission tomography (PET) or single photon emission computed tomography (SPECT). Although electroencephalography (EEG) and magnetoencephalography (MEG) provide higher temporal resolution, fMRI has much better spatial resolution. Functional magnetic resonance imaging can be implemented on many clinical MR scanners with relatively low cost. Functional magnetic resonance imaging and the blood oxygenation level dependent contrast method In most fMRI studies, neuronal activity is detected using the blood oxygenation level dependent (BOLD) contrast method (Fig. 1). The principle behind this technique is as follows: during focal neural activation, oxygen consumption increases, but there is a much greater increase in regional blood flow. As a result, the relative concentration of deoxyhemoglobin decreases in this area. As deoxyhemoglobin is paramagnetic, this phenomenon causes an increase in signal intensity with specific MRI pulse sequences sensitive to the inhomogeneity of magnetic fields [2] (see “Alternative fMRI methods”). Thus, deoxyhemoglobin acts as an endogenous contrast agent in the BOLD contrast method. Data acquisition and analysis Data acquisition Fast imaging techniques, such as echo planar imaging (EPI), are capable of acquiring one scan from the whole brain in a few seconds. In most cases, the total acquisition time should not be too long (less than one hour) to avoid habituation, fatigue, and large head movements during the session. However, longer studies may be conducted in psychological experiments. Many scanners operating at 1.5 Tesla (T) have been used for fMRI scans, but scanners at higher magnetic field strengths are de423 424 Neuro-ophthalmology Figure 1. Blood oxygenation level dependent (BOLD) contrast technique pensates for static magnetic field inhomogeneities, is essential for detecting subtle magnetic field inhomogeneities caused by the change in oxygenation level of hemoglobin. Data analysis Before statistical evaluation, motion correction of the subject’s head movements is usually performed because head movements are inevitable despite the head constraints; the head movements can produce spurious activation or mask real activation. Spatial normalization of the images to standard brain space can be performed to facilitate group data analysis and identification of activated regions in standard coordinates. Top) Relative concentrations of oxyhemoglobin and deoxyhemoglobin in an arteriole/venule interface in resting cortex. Bottom) During neuronal activaty, a presumed coupling with blood flow occurs. Oxygen extraction takes place, but a relatively larger compensatory increase in blood flow leads to a decrease in deoxyhemoglobin concentration in the venule. Since deoxyhemoglobin is paramagnetic, this leads to a local increase in signal in T2*-weighted magnetic resonance images. Adapted with permission from Cohen MS, Bookheimer SY. Localization of brain function using magnetic resonance imaging. Trends in Neurosci 1994;17:268–277. sirable for accurate localization of brain function because of the high sensitivity and high specificity for gray matter activity [3]. Recently, ultra-high field scanners operating at 7 or 8 T have become available [4]. Experimental design Normally, MR images are obtained during both baseline (eg, lights off) and activation (eg, task - lights on) conditions, and changes from the baseline condition are sought. The pair of off-on conditions is repeated several times to enhance the statistical power of detecting taskinduced neural activation. The baseline condition does not need to be a “resting condition” and can be another active task, depending on the kind of brain activity that one is searching (eg, no motion vs motion). Therefore, a variety of task paradigms are possible. Although the task conditions are often blocked (“blocked design”), in which one block (or “epoch”) typically lasts 15 to 40 seconds, “event-related design” can make use of very short stimuli lasting for a few hundred milliseconds. Characteristics of time course of signal intensity The hemodynamic response of the brain is much slower than the neural activity. Accordingly, the signal increase and decrease is delayed by approximately 6 seconds after the onset and offset of the stimulus. The signal change (between the active and control conditions) in fMRI is relatively small (frequently 0.5–5% at 1.5 T), and statistical procedures are necessary to delineate the areas of “activation.” A procedure called shimming, which com- Functional images are analyzed by comparing different states of neural activity. Regions with statistically significant signal intensity changes (task vs baseline, for instance) are classified as “activated pixels” according to a statistical threshold (eg, p < .001). These areas can be localized by overlaying the functional images on the corresponding high-resolution anatomic (structural-often T1-weighted) images, which are often conveniently acquired in the same session but in an independent scan. Applications of functional magnetic resonance imaging for the human visual system Activation of the geniculocalcarine and extrastriate pathways can be studied with fMRI. Although various brain regions (sensorimotor cortex, auditory cortex, language regions, for example) have been mapped, visual cortex activation is easily detected because the associated signal changes are so robust. In addition, activation of the lateral geniculate nucleus can be imaged (Fig. 2, also see “fMRI of lateral geniculate nucleus”). Retinotopic organization of visual cortex (see “Retinotopic mapping”), including the areas corresponding to the peripheral temporal visual fields [5], has been confirmed using fMRI. An optimal check size of 0.5 degrees of visual angle for checkerboard stimulation has also been studied [6]. Clinical applications Presurgical mapping of visual cortex Presurgical mapping of visual cortex is a straightforward clinical application of fMRI. Hirsch et al. [7] compared visual fields and preoperative fMRI in six patients with homonymous visual field defects. In all six patients, fMRI results were consistent with visual field examinations. Presurgical mapping has also been performed in children (Fig. 3) [8,9]. However, although fMRI maps generally agree with patients’ visual fields [10–12], careful interpretation of negative fMRI findings is necessary becasue regions adjacent to lesions may have altered BOLD effects owing to a loss of autoregulation, for instance [13,14]. Functional magnetic resonance image of the visual system Miki et al. 425 Figure 2. Bilateral activation of lateral geniculate nucleus A statistical map overlaid on the SPM T1-template (standard brain) showing bilateral activation of visual cortex and LGN during flash visual stimulation. This statistical map was created with a group data analysis from 10 normal subjects’ data. The data were motion corrected and spatially normalized prior to the statistical analysis. The intersection of crosshairs shows the location of right LGN. Optic neuritis Werring et al. [15•] studied seven patients who recovered from a single episode of unilateral acute optic neuritis. All patients had normal visual acuity and color vision (Ishihara color plates), but two patients showed delayed latency in the pattern-reversal visual evoked potential (VEP) and five patients had abnormal MRI of the optic nerve. The patients had no other lesions on brain MRI. In normal control volunteers, activation by the stimulation of either eye was shown in visual cortex. Stimulation of the clinically unaffected eye in patients induced additional activation in the right insulaclaustrum. Stimulation of the recovered eye from optic neuritis induced extensive activation of bilateral insulaclaustrum, orbitofrontal cortex, lateral temporal cortex, posterior parietal cortex, thalamus, and corpus striatum. There was a positive correlation between the volume of extraoccipital activation and latency in the patternreversal VEP. Because these extraoccipital areas are known to have connections with visual areas, the authors suggested that the extraoccipital network might represent an adaptive response to the VEP delay, which could contribute to the recovery process. Comparison between the maps of patients and those of control subjects showed a greater response in visual cortex in controls compared with patients for both affected and unaffected eyes. The abnormal fMRI response in the “unaffected” eyes is consistent with a previous fMRI study [16]. Although some researchers expect better sensitivity of fMRI over electrophysiologic recordings or MRI [17], the sensitivity and interpretation of fMRI in this setting remains to be investigated in detail. Functional magnetic resonance imaging studies in amblyopia Three articles compared monocular fMRI activation of striate and extrastriate visual cortex in unilateral amblyopia. Goodyear et al. [18•] studied four patients with unilateral anisometropic/strabismic amblyopia using fMRI at 4T while the patients viewed vertical sinusoidal gratings at six spatial frequencies. A good relationship between fMRI response in early visual areas (V1 and possibly including V2) and psychophysical measurement of the perception of 22 % contrast was obtained in all control subjects and patients. Whereas three patients showed similar levels of activation by either the good eye or amblyopic eye, one patient showed smaller signal increases in early visual areas by the amblyopic eye, especially with the visual stimuli at higher spatial frequencies. However, the average number of activated voxels was significantly decreased at higher spatial frequencies for the four patients with amblyopia. They ascribed these findings to the hypothesis that fewer neurons are active for amblyopic eyes, but the average firing rate of the neurons is unaffected in amblyopia. Barnes et al. [19•] studied visual cortex activation in 10 patients with unilateral strabismic amblyopia. The pa- 426 Neuro-ophthalmology tients viewed radial sinusoidal grating stimuli with different spatial frequencies. LCD shutter glasses were used to achieve monocular stimulation. A lower spatial frequency well within and a higher spatial frequency above or close to the subject’s amblyopic acuity were used for each patient. In most but not all patients, decreased cortical activation was observed when the amblyopic eye was stimulated. The higher spatial frequency stimuli produced larger differences between normal and amblyopic activation. However, the decrease did not seem to relate to the subjects’ visual acuity. Also, the contrast sensitivity difference did not correlate with the difference in BOLD signal change. Stimuli with high spatial frequency that are invisible to the amblyopic eye did produce activation of V3A in two patients. Using brain flattening and retinotopic mapping techniques, they showed that there was reduction in monocular activation for the amblyopic eyes in both V1 and V2. Therefore, they concluded that V1 is the earliest anomalous site in amblyopia. Lee et al. [20] studied 5 patients with strabismic amblyopia and 6 patients with anisometropic amblyopia. Blackand-white checkerboards at a spatial frequency of 0.25, 0.5, 1, or 2 cycles per degree of visual angle were used for monocular visual stimulation. Binocular voxel indices (calculated from the number of suprathreshold voxels) were significantly lower in the strabismic amblyopes than in the anisometropic amblyopes. A stronger correlation of signal changes between the amblyopic eye stimulation and defocused normal eye stimulation was found in the anisometropic amblyopes than in the strabismic amblyopes. These findings seem to support the hypothesis that strabismic amblyopia results from the breakdown of binocular interaction. The response to higher-spatial-frequency stimulation was reduced in the anisometropic group but not in the strabismic group. The latter finding is consistent with the notion that anisometropic amblyopia results from selective undersampling of visual images at high spatial frequencies. In anisometropic amblyopia, preliminary studies using fMRI also have shown decreased activation of motion sensitive cortex [21], shifts in ocular dominance in primary visual cortex [22], and differences in visual cortex activation by magno- and parvocellular stimuli [23]. Functional magnetic resonance imaging studies in patients with retrochiasmal lesions Rausch et al. [24•] reported fMRI findings from three patients with cortical blindness (residual rudimentary vision). Activation of parietal cortex, the frontal eye field (FEF), and the supplementary eye field (SEF) as well as reduced activation in V1 was shown by a flash stimulus in these patients. One patient, who showed complete recovery of visual function, was scanned three times. Ac- tivation of FEF and SEF was found in the first two sessions when the patient had only rudimentary vision, but not in the third session when the patient’s field defects had resolved. Because these structures are related to eye movements and spatial attention, and activation of these areas other than primary visual cortex was not observed in normal volunteers, they suggested that this result might have reflected changes in responsiveness of extravisual areas related to attention in patients with severely reduced vision. Kleiser et al. [25•] examined activation of primary visual cortex in three patients with incomplete homonymous hemianopia. The patients viewed radial reversing checkerboards within their impaired, but not absolutely blind, visual fields. Data acquisition was also performed while the intact visual fields were stimulated. One patient showed comparable activation (in terms of the mean signal change and the number of significant voxels) of V1/V2 during the stimulation of either affected or intact visual field. However, activation of ipsilesional V1/V2 could not be detected during the stimulation of impaired visual fields in the other two patients. Because the patients identified the stimulus consciously, the authors speculated that conscious vision could be mediated without V1 activity. However, as the authors mentioned in the discussion, it is not possible to rule out the existence of any residual neural activity in the ipsilesional primary visual cortex from their “negative” finding because these regions may have had sub-threshold activity that was not detected by fMRI. Also, this study suggests that fMRI may not be sensitive enough to detect ipsilesional cortical activation in incomplete homonymous hemianopia. Goebel et al. [26•] tested two patients with long-standing postgeniculate lesions using fMRI to study the neural correlates of ‘blindsight’. For activation of ventral visual pathway, colored images of natural objects (fruit and vegetables) were presented in the relatively and absolutely blind visual fields of the patients. The intact visual fields also were tested with the same stimuli. For activation of dorsal pathway, a rotating spiral stimulus was used. Eccentricity and polar angle mapping experiments were performed in the same session. Although the patients indicated no awareness of the stimuli in most cases, strong responses to the blind-field stimulation in ipsilesional extrastriate areas in both dorsal (motion complex (hMT+/V5)) and ventral (LO, and V4/V8) pathways were found. There was no detectable activation of V1 on the affected side during the stimulation of the blind regions. These findings suggest that the activity of the ipsilesional extrastriate cortex alone is insufficient to generate conscious vision. Functional magnetic resonance imaging of lateral geniculate nucleus Few neuroimaging studies have been performed on the lateral geniculate nucleus (LGN) in humans because of Functional magnetic resonance image of the visual system Miki et al. 427 Figure 3. Activation in dysplastic visual cortex Demonstration of visually activated areas in visual cortex in patients with dysplastic visual cortex but normal visual fields. (A) Axial T1-weighted anatomical MR images of Subject 1, with dysplastic cortex (arrow) in the right parieto-occipital area and a smaller area of dysplasia or heterotopia in the left mesial occipital region (not marked); and Subject 2 (T1-weighted), with semilobar holoprosencephaly, monoventricle, and marked malformation of the occipital lobes. (B) Demonstration of visually activated areas in visual cortex for both subjects. The areas are superimposed upon each subject’s T1-anatomical axial MR images, each parallel to the calcarine sulcus, and four contiguous slices containing visual cortex are shown for each subject. The most dorsal image is left, and in each image, the left side of the brain is on the right, and the right side on the left. The yellow pixels represent the activated areas with a t-statistic corresponding to a Bonferroni corrected p-value of 0.05. The red areas are those with more significant t-statistics. In Subject 2, the activated areas correlated negatively with the stimulus (performed this way because of the subject’s age and study was done under sedation). Thus, the corresponding t-statistics for Subject 1 are: yellow = 4.99 and red = 5.99, for Subject 2 are: yellow = -4.82, and red = -5.82. (reprinted with permission from Liu GT, et al. Functional MRI in children with congenital structural abnormalities of occipital cortex. Neuropediatrics 2000; 31:13–15.) its relatively small size. However, activation of the LGN has been successfully identified in several fMRI studies. Functional magnetic resonance imaging of LGN is a challenge because the LGN is a deep structure, for which the use of surface coils does not improve the signal to noise ratio, unlike for studies of the visual cortex. Functional magnetic resonance imaging studies on LGN have been performed using high field MR scanners, which allow fMRI with high signal to noise ratio [27–33]. In recent reports, activation of LGN has been shown using a conventional MR scanner at 1.5 Tesla, which is widely available, both in children [34] and in adults [35]. Retinotopic mapping Retinotopic mapping is a noninvasive way of identifying borders between visual areas. V1, V2, V3, and V4 all contain complete representations of the visual field, and their borders can be identified by locating representations of vertical and horizontal meridians. For instance, the vertical meridian is represented at the V1-V2 border, and the horizontal meridian is represented at the V2-V3 border. Extensive and groundbreaking fMRI works were performed by Sereno et al. [36], De Yoe et al. [37] and Engel et al. [38]. More recent experiments [39–41] have used techniques developed by these groups to study the responses in different visual areas to various psychophysical stimuli. Visual areas involved in motion perception Area V5 (also known as MT+, hMT+, MT/MST, MT/V5+), the human homologue to the medial temporal (MT) region in the much-studied macaque brain, is involved in motion perception. Functional magnetic resonance imaging techniques can reliably locate area V5 to the intersection of the ascending limb of the inferior temporal sulcus and the lateral occipital sulcus [42]. A variety of motion stimuli, including moving rings, gratings, and dots, have been used to elicit robust activation in area V5 [42–52], as well as in other motion-responsive extrastriate areas, including area V3A [45–48] with fMRI. A recent study found that different cortical areas respond to form coherence and motion coherence [54]. Activation in area V5 and other anterior motion-sensitive areas of cortex [49,50] is reproducible in relation to a motion aftereffect (MAE). Researchers have found that the fMRI activation in response to static stimuli designed to imply motion is located in motion-sensitive areas, and theorized that higher brain centers may modulate the motionsensitive response [51]. 428 Neuro-ophthalmology Recent studies explored the possible interconnections between areas V5 and V3a and a larger motionprocessing network. A three-dimensional structure composed of moving dots stimulated activation in V5, the dorsal posterior occipital gyrus, and the superior occipital gyrus [52]. Speed discrimination stimuli used in a recent study increased BOLD activity in areas V3a, V3v, and V4v [47]. Ahlfors et al. [46] combined the temporal resolution of MEG with fMRI to examine the cortical responses to sudden changes in direction of motion. They found that V5/MT spikes twice, both before and after primary visual cortex and area V3 [46]. When Ffytche et al. noted that hemifield motion stimulation activated contralateral V5 before ipsilateral V5, they postulated two parallel pathways to V5, each selective to motion direction [53]. In a recent case report concerning a patient with an extensive callosal lesion, ipsilateral stimulation resulted in less activation than bilateral stimulation, suggesting the non-callosal transfer of motion information between cerebral hemispheres [54]. Using fMRI techniques, Rees et al. [45] combined an fMRI paradigm with primate single-neuron data to propose a direct relationship between macaque V5 neuronal firing rate and BOLD activation in human subjects in response to coherently moving dots. This study was strengthened by subsequent data showing significantly different spikes per second per neuron measurements versus percent change in BOLD signal in V1 and V5, respectively [55]. Heeger et al. [56] also combined data from electrophysiologic cell recordings with fMRI to suggest that certain cell populations in V5 are directionally selective, and thus exhibit “motion opponency.” These latter experiments represent an exciting new trend, as fMRI’s spatial resolution is bolstered by various imaging modalities with superior temporal resolution, such as MEG and electrophysiology, to further characterize areas of motion–sensitive cortex. Functional magnetic resonance imaging of visual attention During the past few years, there has been a prolific body of research in which fMRI has been used to study visual attention. Visual attention can be defined as the process whereby the visual system uses mental focus to foveate on specific parts of the visual field, as cues in perception and preattentive processes attract interest. Early visual cortical areas respond more robustly to attended versus unattended stimuli [57]. Berman et al. [58] investigated the neural correlates of pursuits and saccades using fMRI at 3 Tesla. Pursuit and saccades compared to eye fixation activated the frontal eye field, supplementary eye field, the intraparietal sulcus, the precuneus, and the anterior and posterior cingulate cortices. In conclusion, an overlap of saccades/pursuits and spatial attention was proposed. Similarly, Nobre et al. [59] examined the overlap in brain function of visual spatial attention and saccades using fMRI. These two behaviors seem to have great overlap in neural systems including frontal and parietal regions, with greater activation during the covert spatial attention task (perceiving brief peripheral targets and reaction timed responses) in relation to the oculomotor task. Hoffman and Haxby [60] attempted to dissociate eye gaze perception and facial perception. Face perception was correlated with preferential activation in the inferior occipital and fusiform gyri; whereas eye gaze seemed to involve the superior temporal sulci and the intraparietal sulcus previously associated with spatial attention. Kim et al. [61] showed spatial attention to be associated with the frontal eye fields, posterior parietal cortex, the cingulate gyrus, the putamen, the thalamus, the anterior insula, dorsolateral prefrontal cortex, temporo-occipital cortex in the middle and inferior temporal gyri, the supplementary motor area, and the cerebellum. Unique to the spatial priming task was a preferential lateralization to the right parietal lobe. Similarly, LaBar et al. [62] showed an overlap of working memory localization with visual spatial attention that included the intraparietal sulcus, ventral precentral sulcus, supplementary motor area, frontal eye fields, thalamus, cerebellum, left temporal neocortex, and right insula. Activation unique to visual spatial attention included the occipitotemporal junction and extrastriate cortex. Many fMRI studies have provided evidence that visual attention modulates visual perception via a top-down cognitive process [63–66]. Heeger [63] showed that one’s performance on visual perceptual tasks related to visual attention. Greenlee [57] showed that attentional processes modulated the amplitude of the BOLD signal in response to motion in V5. Smith et al. [66] illustrated the converse by showing suppression of activity levels in all other regions outside of the visual cortex. More specifically, Sunaert et al. [67] and Huk and Heeger [68] both used fMRI to show that visual attention modulates perception of motion as indicated by activity levels in V5. Beauchamp et al. [69] exemplified a similar modulation of attention on color perception in the anterior and medial color selective areas of the collateral sulcus and the fusiform gyrus. Their research suggested that attentive viewing (versus passive viewing) could be modulated by color when it was germane to a task, as shown via brain activation. In summary, studies using fMRI have shown that brain areas mediating eye movements may also be involved in visual attention. Furthermore, visual attention modulates the activity of visual cortical areas in response to motion, color, and other perceptive cues via a top-down process. Functional magnetic resonance image of the visual system Miki et al. Reproducibility of functional magnetic resonance imaging Reproducibility of cortical activation in a single subject, especially in quantitative data, needs to be addressed before assessing the effects of treatments and longitudinal follow up of patients with fMRI. If activation varies considerably between sessions, effects of a treatment would be difficult to access with fMRI. In fact, reproducibility of activation between sessions (separated by several days) varies across subjects, even when a simple flash visual stimulus is used [70]. The fMRI responses tested in single subjects during motor, cognitive, and visual paradigms over multiple sessions found significant session-by-condition interactions, ie, quite variable activation in each session [71]. Various factors could affect the reproducibility of fMRI, such as variations in attention, habituation, or scanner condition. Tests are not always consistent even at higher magnetic fields, and improvement in signal to noise ratio apparently does not contribute to the reproducibility [72,73]. Acquisition of multiple sessions may disclose reliability of the activation obtained (whether it is truly active or not) but this is not always practical especially in patients, because subject cooperation limits the length and number of scanning sessions. Further investigations will be necessary to see how the variability in fMRI can be controlled. Alternative functional magnetic resonance imaging methods Whereas BOLD contrast is a reflection of brain activity via blood oxygenation and increased blood flow, it is possible using MRI techniques to measure blood flow directly. This might represent a more accurate measurement of the location of brain activity. It is possible to manipulate the proton spins of water in blood such that the blood becomes “tagged”, and this tagged blood will thus travel to a region of interest and interact with the brain parenchyma. The interaction of the tagged blood and tissue affects the region’s T1. By detecting how the T1 is changed during a sequence where blood is labeled to one where blood is not, one can calculate perfusion to that region. The techniques to do this fall under the category of arterial spin labeling (ASL). Two different types of ASL exist: pulsed and continuous. In pulsed ASL, a bolus of blood is labeled, whereas in continuous ASL, blood is continuously labeled. Because the tagging schemes of the two types of ASL are different, the mathematics required to solve for perfusion for the two techniques are different. This means that different conditions and limitations exist for the techniques. Echo-planar imaging and signal targeting with alternating radio frequency (EPISTAR) [74], proximal inversion with a control for off-resonance effects (PICORE) [75], and flowsensitive alternating inversion recovery (FAIR) [76] are 429 examples of pulsed ASL techniques, whereas the technique by Alsop et al. [77] is an example of continuous ASL. The advantage of using ASL techniques in fMRI is its spatial specificity. Because ASL labels arteries, it should theoretically show areas of brain with increased arterial flow. BOLD, on the other hand, may be susceptible to the “draining vein” effect; veins far away from the area of activation also have increased blood oxygenation and thus is detected as active [78]. ASL avoids this because the tag is lost by the time the blood water reaches veins that are far from the region of activity. Another advantage of ASL is that the acquisition of perfusion signal is such that it is possible to obtain perfusion and BOLD data simultaneously [75]; this can confirm or possibly enhance the areas of activation detected in an fMRI experiment. Lastly, ASL’s ability to measure blood flow quantitatively is a vast advantage over BOLD; BOLD contrast have arbitrary units and is hence difficult to perform intersubject comparisons of levels of activity. ASL, on the other hand, has the ability to measure blood flow in absolute units of flow. Since the techniques are relatively new in the MR field, they have not been applied to fMRI studies of the visual system nearly as much as the BOLD contrast technique. The techniques have been applied mostly to study global cerebral blood flow, and the few functional studies done have been studies of the motor cortex. In one study of the visual cortex using FAIR, robust signal changes were detected [79]. Conclusion Functional magnetic resonance imaging has been used in clinical and research studies in patients with visual disturbances and for understanding the neural correlates of the visual system. The preliminary results are encouraging, especially in research studies. Because fMRI techniques still continue to evolve, fMRI promises to be a valuable tool in the understanding of ophthalmologic disorders with neural bases. References Papers of particular interest, published within the annual period of review, have been highlighted as: • Of special interest •• Of outstanding interest 1 Tagamets M–A, Horwitz B: Interpreting PET and fMRI measures of functional neural activity: effects of synaptic inhibition on cortical activation in human imaging studies. Brain Research Bulletin 2001, 54:267–273. 2 Ogawa S, Lee TM, Kay AR, et al.: Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci USA 1990, 87:9868–9872. 3 Kruger G, Kastrup A, Glover GH: Neuroimaging at 1.5 T and 3.0 T: comparison of oxygenation–sensitive magnetic resonance imaging. Magn Reson Med 2001, 45:595–604. 4 Yacoub E, Shmuel A, Pfeuffer J, et al.: Imaging brain function in humans at 7 Tesla. Magn Reson Med 2001, 45:588–594. 430 Neuro-ophthalmology 5 Miki A, Liu GT, Raz J, et al.: Contralateral monocular dominance in anterior visual cortex confirmed by functional magnetic resonance imaging. Am J Ophthalmol 2000, 130:821–824. 6 Miki A, Liu GT, Goldsmith ZG, et al.: Effects of check size on visual cortex activation studied by functional magnetic resonance imaging. Ophthalmic Res 2001, 33:180–184. was reduced in these patients, there was additional activation of the frontal eye fields, the supplementary eye fields, and the parietal cortex. In a patient with complete recovery, such extraoccipital activation disappeared when the patient’s visual field recovered to normal. 25 • Kleiser R, Wittsack J, Niedeggen M, et al.: Is V1 necessary for conscious vision in areas of relative cortical blindness? NeuroImage 2001, 13:654– 661. Three patients with long–standing postgeniculate lesions were studied during monocular stimulation of areas of relative blindness. In the normal hemispheres, substantial activation was evoked, but significant activation was observed in the lesioned hemispheres in only one of the three patients. 7 Hirsch J, Ruge MI, Kim KHS, et al.: An integrated functional magnetic resonance imaging procedures for preoperative mapping of cortical areas associated with tactile, motor, language, and visual functions. Neurosurgery 2000, 47:711–722. 8 Born AP, Miranda MJ, Rostrup E, et al.: Functional magnetic resonance imaging of the normal and abnormal visual system in early life. Neuropediatrics 2000, 31:24–32. 9 Liu GT, Hunter J, Miki A, et al.: Functional MRI in children with congenital structural abnormalities of occipital cortex. Neuropediatrics 2000, 31:13–15. 10 Miki A, Nakajima T, Takagi M, et al.: Detection of visual dysfunction in optic atrophy by functional magnetic resonance imaging during monocular visual stimulation. Am J Ophthalmol 1996, 122:404–415. 11 Kollias SS, Landau K, Khan N, et al.: Functional evaluation using magnetic resonance imaging of the visual cortex in patients with retrochiasmatic lesions. J Neurosurg 1998, 89:780–790. 27 Kleinschmidt A, Merboldt K–D, Hanicke W, et al.: Correlational imaging of thalamocortical coupling in the primary visual pathway of the human brain. J Cereb Blood Flow Metab 1994, 14:952–957. 12 Morland AB, Baseler HA, Hoffmann MB, et al.: Abnormal retinotopic representations in human visual cortex revealed by fMRI. Acta Psychologica 2001, 107:229–247. 28 Buchel C, Turner R, Friston K. Lateral geniculate activations can be detected using intersubject averaging and fMRI. Magn Reson Med 1997, 38:691– 694. 13 Holodny AI, Schulder M, Liu W–C, et al.: The effect of brain tumors on BOLD functional MR imaging activation in the adjacent motor cortex: implications for image–guide neurosurgery. AJNR Am J Neuroradiol 2001, 21:1415–1422. 29 Chen W, Kato T, Zhu X–H, et al.: Mapping of lateral geniculate nucleus activation during visual stimulation in human brain using fMRI. Magn Reson Med 1998, 39:89–96. 14 Miki A, Nakajima T, Takagi M, et al.: Functional magnetic resonance imaging of visual cortex in a patient with cerebrovascular insufficiency. Neuro–ophthalmology 2000, 23:83–88. 30 Chen W, Kato T, Zhu X–H, et al.: Human primary visual cortex and lateral geniculate nuclues activation during visual imagery. NeuroReport 1998, 9:3669–3674. Werring DJ, Bullmore ET, Toosy AT, et al.: Recovery from optic neuritis is associated with a change in the distribution of cerebral response to visual stimulation: a functional magnetic resonance imaging study. J Neurol Neurosurg Psychiatry 2000, 68:441–449. Seven patients with a single episode of unilateral optic neuritis were studied. Stimulation of the affected eye induced extensive extraoccipital activation that was not observed in normal control subjects. 31 Chen W, Zhu X–H, Thulborn KR, et al.: Retinotopic mapping of lateral geniculate nuclues in humans using functional magnetic resonance imaging. Proc Natl Acad Sci USA 1999, 96:2430–2434. 32 Chen W, Zhu XH. Correlation of activation sizes between lateral geniculate nucleus and primary visual cortex in humans. Magn Reson Med 2001, 45:202–205. 16 Rombouts SARB, Lazeron RHC, Scheltens P, et al.: Visual activation patterns in patients with optic neuritis: an fMRI pilot study. Neurology 1998, 50:1896– 1899. 33 Miki A, Liu GT, Raz J, et al.: Visual activation in functional magnetic resonance imaging at very high field (4 Tesla). J Neuro–ophthalmol 2001, 21:8–11. 17 Gareau PJ, Gati JS, Menon RS, et al.: Reduced visual evoked responses in multiple sclerosis patients with optic neuritis: comparison of functional magnetic resonance imaging and visual evoked potentials. Multiple Sclerosis 1999, 5:161–164. 34 Morita T, Kochiyama T, Yamada H, et al.: Difference in the metabolic response to photic stimulation of the lateral geniculate nucleus and the primary visual cortex of infants: a fMRI study. Neurosci Res 2000, 38:63–70. 35 Miki A, Raz J, Haselgrove JC, et al.: Functional magnetic resonance imaging of lateral geniculate nucleus at 1.5 Tesla. J Neuro–ophthalmol 2000, 20:285– 287. 36 Sereno MI, Dale AM, Reppas JP, et al.: Borders of multiple visual areas in humans revealed by functional magnetic resonance imaging. Science 1995, 268:889–893. 37 DeYoe EA, Carman GJ, Bandettini P, et al.: Mapping striate and extrastriate visual areas in human cerebral cortex. Proc. Natl. Acad. Sci, 1996, 93:2382– 2386. 38 Engel SA. Rumelhart DE, Wandell BA, et al.: FMRI of human visual cortex. Nature 1994, 369:525. 39 Engel SA, Glover GH, Wandell BA. Retinotopic organization in human visual cortex and the spatial precision of functional MRI. Cerebral Cortex 1997, 7:181–192. 40 Hadjikhani N, Tootell RB. Projection of rods and cones within human visual cortex. Hum Brain Mapp 2000, 9:55–63. 41 Singh KD, Smith AT, Greenlee MW. Spatiotemporal frequency and direction sensitivities of human visual areas measured using fMRI. Neuroimage 2000, 12:550–564. 15 • 26 •• Goebel R, Muckli L, Zanella FE, et al.: Sustained extrastriate cortical activation without visual awareness revealed by fMRI studies of hemianopic patients. Vision Res 2001, 41:1459–1474. Responsiveness of dorsal and ventral stream areas was investigated with fMRI during the stimulation of cortically blind visual fields in two patients with hemianopia. The stimulation to the blind fields produced strong responses in ipsilesional extrastriate cortex (but not in the early visual areas) without the patients’ awareness of the stimuli. 18 • Goodyear BG, Nicole DA, Humphrey GK, et al.: MRI response of early visual areas to perceived contrast in human amblyopia. J Neurophysiol 2000, 84:1907–1913. Four patients with strabismic/anisometropic amblyopia were studied. A good relationship between fMRI response and psychophysical measurements was found. Stimulation of the amblyopic eye was associated with a decreased number of activated voxels compared with stimulation of the good eye. 19 Barnes GR, Hess RF, Dumoulin SO, et al.: The cortical deficit in humans with strabismic amblyopia. J Physiol 2001, 533:281–297. •• Ten patients with strabismic amblyopia were studied using fMRI during visual stimulation with LCD shutter glasses. Decreased cortical activation in both V1 and V2 was found when the amblyopic eyes were stimulated regardless of the spatial frequency of the visual stimulus. No close correlation was found between fMRI responses and psychophysical deficits. 20 Lee K–M, Lee S–H, Kim N–Y, et al.: Binocularity and spatial frequency dependence of calcarine activation in two types of amblyopia. Neurosci Res 2001, 40:147–153. 21 Bonhomme GR, Liu GT, Miki A, et al.: Decreased cortical activation in response to a motion stimulus in anisometropic amblyopic eyes. Invest Ophthalmol Vis Sci 2000, 41 (suppl): 703. 22 Liu GT, Miki A, Goldsmith ZG, et al.: Ocular dominance histograms of visual cortex using functional MRI (fMRI). An alternative method. Invest Ophthalmol Vis Sci 2000, 41 (suppl): 535. 42 Dumoulin SO, Bittar RG, Kabani NJ, et al.: A new anatomical landmark for reliable identification of human area V5/MT: a quantitative analysis of sulcal patterning. Cerebral Cortex 2000, 10:454–463. 23 Miki A, Siegfried JB, Liu GT, et al.: Magno– and parvocellular visual cortex activation in anisometropic amblyopia studied with functional MRI. Invest Ophthalmol Vis Sci 2001, 42 (suppl): 52. 43 Watson JD, Myers R, Frackowiak RS, et al.: Area V5 of the human brain: evidence from a combined study using positron emission tomography and magnetic resonance imaging. Cerebral Cortex 1993, 3:79–94. 44 Tootell RBH, Reppas JB, Kwong KK, et al.: Functional analysis of human MT and related visual cortical areas using magnetic resonance imaging. J Neuroscience 1995, 15:3215–3230. 24 • Rausch M, Widdig W, Eysel UT, et al.: Enhanced responsiveness of human extravisual areas to photic stimulation in patients with severely reduced vision. Exp Brain Res 2000, 135:34–40. Three patients with severely reduced vision were tested. Although activation of V1 Functional magnetic resonance image of the visual system Miki et al. 431 45 Rees G, Friston K, Koch C. A direct quantitative relationship between the functional properties of human and macaque V5. Nature Neuroscience 2000, 3:716–723. 62 LaBar KS, Gitelman DR, Parrish TB, et al.: Neuroanatomic overlap of working memory and spatial attention networks: a functional MRI comparison within subjects. Neuroimage 1999, 10:695–704. 46 Ahlfors SP, Simpson GV, Dale AM, et al.: Spatiotemporal activity of a cortical network for processing visual motion revealed by MEG and fMRI. J Neurophysiology 1999, 82:2545–2555. 63 Heeger DJ: Linking visual perception with human brain activity. Current Opinion in Neurobiology 1999, 9:474–479. 64 47 Sunaert S, Van Hecke P, Marchal G, et al.: Attention to speed of motion, speed discrimination, and task difficulty: an fMRI study. Neuroimage 2000, 11:612–623. Coull JT, Frith CD, Buchel C, et al.: Orienting attention in time: behavioural and neuroanatomical distinction between exogenous and endogenous shifts. Neuropsychologia 2000, 38: 808–819. 65 48 Braddick OJ, O’Brien JM, Wattam–Bell J, et al.: Form and motion coherence activate independent, but not dorsal/ventral segregated, networks in the human brain. Current Biology 2000, 10:731–734. Hopfinger JB, Buonocore MH, Mangun GR: The neural mechanisms of top–down attentional control. Nat Neuroscience 2000, 3:284–291. 66 Smith AT, Singh KD, Greenlee MW: Attentional suppression of activity in the human visual cortex. Neuroreport 2000, 11:271–277. 67 Sunaert S, Van Hecke P, Marchal G, et al.: Attention to speed of motion, speed discrimination, and task difficulty: an fMRI Study. Neuroimage 2000, 11:612–623. 68 Huk AC, Heeger DJ: Task–related modulation of visual cortex. J Neurophysiol 2000, 83:3525–3536. 69 Beauchamp MS, Haxby JV, Jennings JE, et al.: An fMRI version of the Farnsworth–Munsell 100–Hue test reveals multiple color–selective areas in human ventral occipitotemporal cortex. Cerebral Cortex 1999, 9:257–263. 49 Taylor JG, Schmitz N, Ziemons K, et al.: The network of brain areas involved in the motion aftereffect. Neuroimage 2000, 11:257–270. 50 He S, Cohen ER, Hu X. Close correlation between activity in brain area MT/V5 and the perception of a visual motion aftereffect. Current Biology 1998, 8:1215–1218. 51 Kourtzi Z, Kanwisher N. Activation in human MT/MST by static images with implied motion. J Cognitive Neuroscience 2000, 12:48–55. 52 Paradis AL, Cornilleau–Peres V, Droulez J, et al.: Visual perception of motion and 3–D structure from motion: an fMRI study. Cerebral Cortex 2000, 10:772–783. 70 Miki A, Raz J, van Erp TGM, et al.: Reproducibility of visual activation in functional MR imaging and effects of postprocessing. AJNR Am J Neuroradiol 2000, 21:910–915. 53 Ffytche DH, Howseman A, Edwards R, et al.: Human area V5 and motion in the ipsilateral visual field. European Journal of Neuroscience 2000, 12:3015– 3025. 71 McGonigle DJ, Howseman AM, Athwal BS, et al.: Variability in fMRI: an examination in intersession differences. Neuroimage 2000, 11:708–734. 72 54 Clarke S, Maeder P, Meuli R, et al.: Interhemispheric transfer of visual motion information after a posterior callosal lesion: a neuropsychological and fMRI study. Experimental Brain Research 2000, 132:127–133. Tegeler C, Strother SC, Anderson JR, Kim SG: Reproducibility of BOLD– based functional MRI obtained at 4 T. Human Brain Mapping 1999, 7:267– 283. 73 55 Heeger DJ, Huk AC, Wilson SG, et al.: Spikes versus BOLD: what does neuroimaging tell us about neuronal activity? Nature Neuroscience 2000, 3:631–633. Miki A, Raz J, Englander SA, et al.: Reproducibility of visual activation in functional magnetic resonance imaging at very high field strength (4 Tesla). Jpn J Ophthalmol 2001, 45:1–4. 74 56 Heeger DJ, Boynton GM, Demb JB, et al.: Motion opponency in visual cortex. J Neuroscience 1999, 19:7162–7174. Edelman RR, Siewert B, Darby, DG, et al.: Qualitative mapping of cerebral blood flow and functional localization with echo–planar MR imaging and signal targeting with alternating radio frequency. Radiology 1994, 192:513– 520. 57 Greenlee MW. Human cortical areas underlying the perception of optic flow: brain imaging studies. International Review of Neurobiology 2000, 44:269– 292 75 Wong EC, Buxton RB, Frank LR. Implementation of quantitative perfusion imaging techniques for functional brain mapping using pulsed arterial spin labeling. NMR in Biomedicine 1997, 10:237–249. 58 Berman RA, Colby CL, Genovese CR, et al. Cortical networks subserving pursuit and saccadic eye movements in humans: an fMRI Study. Human Brain Mapping 1999, 8:209–225. 76 Kim S-G: Quantification of relative cerebral blood flow change by flow–sensitive alternating inversion recovery (FAIR) technique: application to functional mapping. Mag Reson Med 1995, 34:293–30. 59 Nobre AC, Gitelman DR, Dias EC, et al.: Covert visual spatial orienting and saccades: overlapping neural systems. Neuroimage 2000, 11:210–216. 77 Alsop DC, Detre JA: Reduced transit–time sensitivity in noninvasive magnetic resonance imaging of human cerebral blood flow. J Cerebral Blood Flow Metabol 1996, 16:1236–1249. 60 Hoffman EA, Haxby JV: Distinct representations of eye gaze and identity in the distributed human neural system for face perception. Nat Neurosci 2000, 3:80–84. 78 Lai S, Glover GH, Haacke EM: Spatial selectivity of BOLD contrast: effects in and around draining veins. In Functional MRI. Edited by Moonen CTW, Bandettini PA. New York: Springer; 1999:221–231. 61 Kim YH, Gitelman DR, Nobre AC, et al.: The large–scale neural network for spatial attention displays multifunctional overlap but differential asymmetry. Neuroimage 1999, 9:269–277. 79 Li T–Q, Haefelin TN, Chan B, et al.: Assessment of hemodynamic response during focal neural activity in human using bolus tracking, arterial spin labeling and BOLD techniques. NeuroImage 2000, 12:442–451.