Journal of Neuro-Ophthalmology, September 1999, Volume 19, Issue 3

Functional Magnetic Resonance Imaging in the Visual System

Functional magnetic resonance imaging (fMRI) is a relatively new technique for measuring brain function during resting and activated conditions with good spatial and temporal resolution. Because of a robust and reproducible activation response to visual stimuli in the occipital cortex, many studies have been directed at visual function. The methodology has been refined progressively to allow more accurate detection of the small activation signal, and using computational mapping foci of cerebral activity have been displayed in a two-dimensional format. Several factors modifying the activation signal have been identified. fMRI has been used to define the retinotopic representation of areal boundaries and the localization of higher visual functions in the occipital cortex. Motion perception in area middle temporal (MT) is well-recognized, but eye movement studies are limited. The activated signal may have significant implications for our understanding of brain metabolism, but cerebral blood flow and oxygenation sensitive recordings after prolonged visual stimulation have given conflicting results. Clinically, fMRI can follow changes in cerebral activity during a progressive neurologic illness and measure responses to treatment. Neurosurgical planning in disorders such as epilepsy may be facilitated.

Journal of Nc- itro- Ophllialmology 19( 1): 186- 200, 1999. © 1999 Lippincotl Williams & Wilkins, Inc., Philadelphia Functional Magnetic Resonance Imaging in the Visual System Robert M. McFadzean, MB, chB, DO, FRCOphth, Barrie C. Condon, PhD, and Dai B. Barr, MB, ChB, FRCOphth Functional magnetic resonance imaging ( f'MRl) is a relatively new technique for measuring brain function during resting and activated conditions with good spatial and temporal resolution. Because of a robust and reproducible activation response to visual stimuli in the occipital cortex, many studies have been directed at visual function. The methodology has been refined progressively to allow more accurate detection of the small activation signal, and using computational mapping foci of ce­rebral activity have been displayed in a two- dimensional for­mat. Several factors modifying the activation signal have been identified. fMRl has been used to define the retinotopic representation of areal boundaries and the localization of higher visual functions in the occipital cortex. Motion perception in area middle tem­poral ( MT) is well- recognized, but eye movement studies are limited. The activated signal may have significant implications for our understanding of brain metabolism, but cerebral blood flow and oxygenation sensitive recordings after prolonged visual stimu­lation have given conflicting results. Clinically, fMRI can fol­low changes in cerebral activity during a progressive neuro­logic illness and measure responses to treatment. Neurosurgical planning in disorders such as epilepsy may be facilitated. Key Words: Functional magnetic resonance imaging- Vision. There are several techniques available for measuring brain function during resting and activated conditions, including direct cortical electrical stimulation, scalp re­corded electrical evoked potentials, single photon emis­sion computed tomography ( SPECT), positron emission tomography ( PET), magnetoencephalography ( MEG), near- infrared spectroscopy ( NIRS), magnetic resonance spectroscopy ( MRS), and functional magnetic resonance imaging ( fMRI). Because these techniques depend on different principles, the resultant brain mapping may re­flect different neuronal activation processes. In addition, these techniques have different temporal and spatial resolution properties. Manuscript received February 11, 1999; accepted April 22, 1999. From the Department of Neuro- ophthalmology ( RMM), Department of Medical Physics ( BCC), and Department of Neuro- ophthalmology ( DBB), Institute of Neurological Sciences, Glasgow, Scotland. Address correspondence and reprint requests to Dr. Robert M McFadzean, Department of Neuro- ophthalmology, Institute of Neuro­logical Sciences, Southern General Hospital, Govan Road, Glasgow, G5I 4TF, Scotland. Direct electrical stimulation of the cortical surface de­pends on activation or disruption of local cortical circuits with a spatial resolution of 5 to 10 mm and a temporal resolution of a few seconds. Visually evoked potentials and MEG probably reflect underlying neuronal activity but have poor spatial resolution, although excellent tem­poral resolution of milliseconds. SPECT uses ( 99mTc) HMPAO as a cerebral blood flow ( CBF) imaging agent with a spatial resolution of 8 to 9 mm and a temporal resolution of approximately 45 seconds. PET using H2- l 5 0 tracer measures changes in local CBF after neuronal activation. Its spatial resolution is 4 to 5 mm, but its temporal resolution is governed by the time required for clearance of the tracer from the bloodstream, about 50 seconds. Optical imaging with near- infrared photons combines good temporal resolution with spatial resolu­tion of at least 5 mm. MRS provides a useful measure of regional brain chemistry but with poor spatial ( 1,000 mm3 at best) and temporal resolution. The relatively re­cent introduction of fMRI ( 1- 6) permits activation map­ping of the brain in a noninvasive manner with a poten­tial spatial resolution of < 1 mm ( 7) and temporal reso­lution of < 1 second ( 8,9) using fast MR or echo- planar imaging ( EPI) techniques ( 10- 12). fMRI is thought to measure local changes in blood oxygenation after in­creases in CBF and modification of oxygen utilization after neuronal activation ( vide infra) ( 1,5,6,13). The presence of a robust activation response to visually evoked stimuli in the occipital cortex has made the striate and extra- striate region a particularly attractive area to study. In addition, the gatekeeper role of the primary striate visual cortex ( area VI) in the determination of higher visual function is well- known, whereas the local anatomy and boundaries of the striate cortex have been studied extensively and are well- understood. The tech­nique is useful not only in the mapping of localized cerebral cortical functions in normal individuals, but also in the assessment of disease processes ( 14,15) and moni­toring of psychiatric disorders ( 16,17) and psychophar-macologic treatments ( 18). A typical activation pattern in the occipital cortex on exposure to an alternating black and white checkerboard is shown in Figure 1. The advantages of fMRI are considerable in that the technique allows noninvasive imaging with excellent 186 fMRI IN THE VISUAL SYSTEM 187 FIG. 1. Activated pixels ( yellow) in the occipital cortex after ex­posure to an alternating black and white checkerboard at 8 Hz. spatial resolution almost in real time. Thus, it is possible to carry out repetitive studies on normal subjects with reasonable reproducibility and to follow changes in ce­rebral activity during the course of a progressive disease, including assessment of recovery patterns and responses to treatment in various neurologic disorders such as stroke and head injury. Dyslexia and albinism ( 19,20) can be evaluated and neurosurgical planning facilitated ( 21- 23). This is in contrast to radionuclide techniques such as PET and SPECT, in which repeated studies are constrained by ethical considerations and radiation dos­age. Advanced technology with fast scanning times and multislice cerebral images makes it possible to scan the entire brain within a short time scale and identify regions of interest ( ROI) activated by different experimental paradigms. Furthermore, appropriate scanning facilities are widely available in neuroscience departments. Un­fortunately, as with conventional MRI, individuals with cardiac pacemakers or retained foreign metallic material in their body must be excluded and claustrophobic pa­tients find the technique difficult to tolerate. Volunteer or patient cooperation is crucial not only for participation in the activation task, but also in the maintenance of im­mobility during the scanning period to avoid misregis­tration artifacts ( 24,25). However, image processing al­gorithms are available that can overcome the effects of patient movement to some extent by realigning the im­ages ( 26). Cerebral tissue close to the air- filled sinuses ( e. g. at the base of the brain) is prone to signal void artifacts, but this is not a problem with studies of the visual cortex. While images can be acquired in close to real time, f'MRI in common with PET and SPECT scan­ning has the disadvantage of measuring alterations in CBF that change over a period of seconds compared to the millisecond time scales of the initiating neuronal event. It is unlikely that fMRI will ever be able to in­vestigate events with such a short time scale, which re­quire instead the high temporal resolution of EEC or MEG. PRINCIPLES OF METHOD- THE BOLD EFFECT Initial activation studies in the primary visual cortex were based on images of cerebral blood volume after injection of a bolus of paramagnetic contrast agent, such as gadolinium diethylene triaminc penta- acetic acid in­travenously, obtained before and during stimulation with a flashing light ( 2). Cerebral blood volume increased during visual stimulation, but this technique required the use of an external contrast agent. Based on the magnetic properties of hemoglobin and their effect on the mag­netic resonance ( 27,28), it was realized that paramag­netic deoxyhemoglobin could be detected in contrast to diamagnetic oxyhemoglobin and therefore used as an endogenous contrast agent ( 29). This difference in the physical properties of hemoglobin means that when blood flows through cerebral tissue deoxyhemoglobin creates point magnetic inhomogeneilies within blood vessels, which result in microscopic field distortions around those vessels. As the coherence of the signal from hydrogen nuclei in water in surrounding tissue is par­tially destroyed by these microscopic inhomogeneilies, the signal intensity during MRI is lower than it would be if they did not exist. During neuronal activation, a sub­stantial increase in the metabolic demand of the local cerebral tissue involved results in a corresponding in­crease in CBF and oxygenation, which more than com­pensates for that demand ( 30,31). Consequently, there is a decrease in the local concentration of deoxyhemoglo­bin resulting in an increase in the intensity of the local MR signal. This finding has enabled the use of MRI in the monitoring of local modulations in the level of blood oxygenation associated with brain activity ( 32). The activated signal in the cerebral cortex is referred to as the blood oxygenation level dependent ( BOLD) contrast effect. The size of the signal is relatively small ( range 2- 10%); therefore, optimization of the visual stimulus and recording conditions is important. Its pre­cise origin has not been identified, but it is believed to arise from around small venules in the cerebral cortex ( 33,34). The signal intensity can be correlated with local changes in CBF, tissue oxygenation, and neural activity ( 5,35). Typically, activity is averaged during a time scale of 2 to 6 seconds and the volume of each individual image element is 1 to 27 mm3 of cerebral cortex ( 36). During recordings, particular attention must be paid to avoidance of artifacts because of volunteer or patient movement ( e. g. head movement, eye movements, and physiologic changes such as the cardiac pulse and respi­rations). Considerable concern has been expressed that subject movement during activation may correlate with the task being performed and the resulting signal changes be mistaken for evidence of brain activation, e. g., flinch- J Nmro- Ophthulmol, Vol. 19. No. 3, 1999 188 R. M. McFADZEAN ET At. ing when a flashing light comes on or goes off ( 37,38). An assessment of the effect of stimulus- correlated mo­tion using conventional visual and motor protocols and an image coregistration technique showed stimulus-correlated motion and synthetic cumulative difference images with a striking similarity to the equivalent func­tional image in each case ( 37). Even objects moving outside the field of view may have an effect on the fMRI signal if temporally correlated with the performance of a particular task ( 39). However, these artifactual signal in­tensity changes were characterized by their location, greater magnitude, and more rapid increase to a maxi­mum than those seen from typical activations ( 25). The increasing recognition of false- positive signals has stimulated efforts to develop a robust processing tech­nique with clearly defined failure modes. Improved func­tional activation maps have been obtained using naviga­tor echoes to monitor and compensate for signal fluctua­tions caused by motion, while, by simultaneously monitoring the respiration and heartbeat during the ac­quisition of imaging data and retrospectively synchroniz­ing the imaging data with physiologic activity, physi­ologic effects have been estimated and removed ( 24). Variations in venous deoxyhemoglobin levels in re­sponse to neuronal activation represent a complex inter­play between focal changes in CBF, cerebral blood vol­ume, and regional metabolism. Various mathematical models attempt to categorize the response of venous oxy­genation to changes in these variables to obtain a quan­titative understanding of changes in blood oxygenation and to relate these changes to the observed dynamics of the fMRI signal change ( 13,35,40). METHODOLOGY Conventional MRI usually employs spin- echo phe­nomena to produce Tl- and T2- weighted pulse se­quences, which are associated with high signal and good tissue contrast and resolution. However, fMRI requires a greater sensitization to the tiny differences in magnetic field caused by differing blood oxygenation. Therefore, fMRI pulse sequences are weighted toward the T2* ef­fect, which reveals these differences, but is associated with lower signal and poorer tissue contrast ( 5,6,33,34). Originally, single slice gradient ( field)- echo pulse se­quences with echo times of approximately 60 millisec­onds and repetition times of 90 milliseconds with 120 repetitions were used, but in recent years EPI has become the sequence of choice ( 4,11,12,41). While gradient- echo imaging typically took 14 seconds to acquire data from a single slice, EPI can produce images in a single repeti­tion time of 100 milliseconds or less, although image resolution is poorer than with conventional MRI ( usually 64 x 64 compared with 256 x 256). Faster acquisition times means that a stack of two- dimensional slices can be acquired to cover the entire brain within seconds. Because (" MRI signal ehanges are larger at higher field strengths, studies are usually performed at 1.5 Tesla or higher ( 10,1 1,42- 44), although studies at 1.0 Tesla have been published ( 45,46). The percentage change in signal intensity with standard 1.5 Tesla hardware can be im­proved by optimizing section thickness, echo time, and field of view during visual activation ( 43). However, during neuronal activation the fMRI signal has been shown to increase linearly and quadratically with field strength ( 34), making the use of scanners up to 4.0 Tesla desirable ( 34,41). These scanners offer better spatial resolution but are technically more difficult to use. High field strength imaging at 4.0 Tesla provides an increased contribution from the venules and capillary bed and avoids contributions from large venous vessels or in­flow effects from large arteries, which are undesirable because of their poor spatial correspondence with the actual site of neuronal activations ( 33,34). The increased importance of the susceptibility difference between deoxygenated and oxygenated blood at higher field strengths was illustrated with activated image intensity up to 28% at 4.0 Tesla but only up to 7% at 1.5 Tesla ( 41). Easily detectable signal increases of 5% to 20% were observed in area VI at 4.0 Tesla, with signal in­creases that were predominantly restricted to areas con­taining grey matter ( 5,6). Nevertheless, some workers hesitate to use higher field strengths because of potential clinical adverse effects on volunteers and increased im­age artifacts caused by susceptibility effects. Although EPI adds the further dimension of improved temporal resolution with images of brain function almost in real time, the spatial resolution is not as good as with gradient- echo fMRI ( typically 64 x 64 cf. 128 x 128). A method of multislice interleaved excitation cycles ( 47) allows the acquisition of activation maps at multiple planes within total imaging times of a few seconds. By using a further specific method, it is possible to obtain a three- dimensional data set of the visual cortex in 20 sec­onds ( 48). EPI and signal targeting with an alternating radio ( EPISTAR) frequency technique is a rapid, nonin­vasive means of creating qualitative maps of CBF with signal intensity changes that range from 13% to 193% ( 11). Using gradient- echo and spin- echo EPI at 3.0 Tesla, the greater microvascular selectivity of high- resolution spin- echo imaging enabled distinct activation patterns sensitive to stimulus motion to be detected in area VI/ V2 that were not apparent with gradient- echo imaging ( 7). Visual stimulation takes various forms, including flashing lights, red light emitting diodes ( LED) mounted in goggles, conventional full field alternating black and white checkerboard patterns ( sometimes modified into hemifields, wedges, or circles), and video and cine- film presentations. Modifications may be made to luminance, contrast, color, spatial, and temporal frequency, and so on, and the stimuli may be viewed under different con­ditions of attention, emotion, drug effects, and so on. Higher visual functions, such as appreciation of color, object and face recognition, movement and conscious perception, and others, may be assessed using appropri­ate experimental paradigms. A binocular fiberscope ( 49) can be used to relay high- resolution images of cathode ray tube displays from an adjacent room to an observer lying in a scanner, with a display of accurately controlled ./ Neiiro- Oplillialmol. Vol. 19. No. I 1999 fMRI IN THE VISUAL SYSTEM 189 high and low contrast wide- field images to the observer. To provide high- quality visual stimulation within the in­tense magnetic field of the MR scanner, a custom Max-wellian- view optic system ( 50) has been designed to project images directly on to the retina of subjects. This visual stimulator offers the particular advantages of a large field of view ( 60°) with a binocular stereoscopic display. The stimulation frequency dependence of fMRI visual activation was found to agree with previous PET studies, when the largest signal response occurred at 8 Hz ( 4). A simple visual experimental set- up used at the Institute of Neurological Sciences, Glasgow, is illus­trated in Figure 2. The projection apparatus ( a video projector or liquid crystal display projection panel on an overhead projector) projects the computer- generated stimuli through the observation window into the magnet room. The image falls onto a translucent screen, which the subjects view through a mirror inclined at 45° above their eyes. Because the distance between the subject's head and the projection apparatus is approximately 7 m, an additional optical lens is required to reduce the size of the projected image. The available field of view is re­stricted by the MRI head coil and the bore of the magnet. During a typical fMRI examination, images are col­lected at varying intervals over a period of several min­utes during activated and resting conditions. In its sim­plest form, measurement of the activation signal involves subtraction of resting images ( e. g., in total darkness) from activated images ( e. g., on exposure to a flashing alternating black/ white checkerboard). Because of the small size of the signal change, collection of many im­ages improves the statistical reliability of the data. Simple subtraction of images acquired at baseline from activated images should theoretically define only areas of the brain in which the CBF has changed. Unfortu­nately, this process will yield artifactual areas of activa­tion because of small involuntary movements of the sub­ject's head during image acquisition ( 24,25). In addition, the signal from blood in arterial vessels will vary de­pending on the part of the cardiac cycle during which the image is acquired. To reduce the contribution from these artifactual factors, it has been necessary to resort to ad­vanced statistical approaches. Essentially, the application of the stimulus ( on, off, on, off, etc.) has a time course, which may be represented as a simple " box- car" shaped IDEALISED BOX CAR FUNCTION ST1MUUS OH Plxai ntenelry FIG. 2. Experimental set- up in which the subject views through an inclined mirror above both eyes the stimulus projected from an outer room on to a translucent screen outside or inside the mag­net bore. 0 20 40 60 80 100 120 Tlm « ( l) FIG. 3. An idealized box- car function graph compared with the activated signal in the occipital cortex in response to an alternat­ing black and white checkerboard at 8 Hz. function, and should be correlated in a time- series sense with the activated signal changes on a pixel- by- pixel basis ( see Fig. 3). Image elements showing a signal in­crease every time the stimulus is applied will be highly correlated, whereas pixels overlying arterial vessels may coincidentally show signal increases on one or two oc­casions. By setting high statistical probability levels ( typically /; < 0.005), it is possible to filter out these latter artifacts to a large degree. Various statistical packages have been developed to cope with these problems ( 51,52). Alter identification of an appropriate ROI ( e. g. area VI), an idealized box- car function graph is com­pared to the signal changes, which should temporally correspond to the visual stimulus ( Fig. 3). fMRI re­sponses are slow when compared to changes in neural activity ( e. g., the onset of a visual checkerboard pattern evoked a signal response that was delayed by 1 to 2 seconds and reached 90% of peak in 5 seconds with a slightly slower return to baseline) ( 50). The signal changes are delayed ( rise- time) by the hemodynamic re­sponses of the subject with blood flow taking up to 9 seconds to reach a plateau level after application of the visual stimulus or to fall to baseline values after its re­moval. This time- lag must be taken into account during the data analysis. Signal responses in the presence of noise may be detected using cross- correlation techniques and verified by statistical parametric mapping ( 51,52). Fuzzy cluster analysis has proved to be robust and effi­cient in the separation of functional brain activation from noise or other sources and when used in combination with an appropriate model calculation allows quantifica­tion of flow and BOLD contributions in areas with dif­ferent vascularization ( 53). It is possible to modify fMRI to produce recordings that are separately sensitized to cerebral blood oxygen­ation ( CBO) and CBF ( 54). The duration of the visual stimulus used is important because CBF- sensitive re­cordings appear to remain elevated during the entire stimulation period, but in some CBO- sensitive record­ings a signal decrease occurs after prolonged stimulation ranging from 1 to 6 minutes ( 13,54- 59). However, others have not confirmed this signal decrease ( 60,61). As the J Neum- Ophlhatmol. Vol. 19. No. J, 1W9 190 R. M. McFADZEAN ET Ah. acquisition of more images improves the statistical reli­ability of the data, prolonged stimulation may be desir­able, but only if there is no decrease in the relatively small activation signal. The discrepancy in these findings may represent different experimental paradigms and re­cording techniques, but the issue is unresolved. Another major concern is the reproducibility of the activated signal. The application of adaptive correlation thresholds resulted in reasonable reproducibility in re­peated single subject studies, but reproducibility across groups was poor ( 62). Test- retest analysis ( 63) of the activated area in the human visual cortex in 18 volun­teers showed significant variability, both intersubject and intrasubject. However, using fuzzy cluster analysis inter-trial reproducibility for repeated single subject studies was improved, but dependent on signal- to- noise ratio, motion artifact, and subject cooperation ( 64). A study of within- subject reproducibility of visual brain activation using EPI in 10 healthy subjects examined on three oc­casions concluded that measures of brain activity were reasonably reproducible on a routine clinical EPI system ( 65). Nevertheless, there were difficulties in separating the contributions of motion, repositioning errors, and true physiologic changes. High reproducibility of fMRI studies is important in the development of potential clini­cal applications. It is difficult to compare published data quantitatively owing to differences in sequence design and parameters, as well as statistical methods applied to enhance function- related image contrast. Technical, methodologic, and physiologic factors influence the vari­ability of signal enhancement and the apparently acti­vated area size. These should be taken into account in the interpretation of fMRI data quantitatively ( 62) and re­quire further exploration. Because of the irregularities of the surface of the hu­man cerebral cortex and the variations from one indi­vidual to the next, it has been difficult to illustrate the activated images. An early spatial topographic technique in the visual cortex involved the application of a cortical ribbon along the 1.5- to 3- mm thick striate cortex to incorporate the subject- unique enfoldings ( 66). More re­cently, to compensate for the convolutions of the cortical gyri and sulci, reconstructions of the cortical surface have been developed using computational mapping ( 67). Surface base visualization involves reconstructing corti­cal surfaces and displaying them along with associated experimental data in various complementary formats, in­cluding three- dimensional native configurations, two-dimensional slices, extensively smooth surfaces, ellipsoi­dal representations, and cortical flat maps. In particular, this approach allows unfolding of the cortical sulci and representation of foci of cortical activity in a flattened two- dimensional mapping format. Because individual cortical areas vary in size by a factor of at least two and many areas are categorized by internal compartments or modules whose total number and dimensions vary across individuals, it would be unrealistic to expect functional correspondence to be identified more precisely than within a few millimeters for human cortex. MODIFICATION OF ACTIVATION SIGNAL Because the amplitude of the activation signal is rela­tively small and difficult to detect against background noise, it is important to be aware of potential modifying factors. These include sex, attention, emotional state, ocular dominance, luminance contrast, and duration of stimulus. Among 38 healthy subjects ( 20 males and 18 females), males were more likely than females to have an unde­tectable MR signal change after photic stimulation ( 68), but in a study of 16 healthy young subjects ( 8 male and 8 female), the signal response in area VI to binocular photic stimulation was 38% lower in females than in males and much of the difference was lateralized to the right hemisphere ( 69). However, hemoglobin levels were not measured in these studies, which is regrettable be­cause the BOLD fMRI response may be particularly sen­sitive to hemoglobin concentration ( it is the iron within heme that provides the contrast in the MR signal inten­sity, and in general young women have a lower hemo­globin level than young men). During repeated presentations of identical visual mo­tion stimuli ( 70), only the attentional component of the task was varied and attention- related enhancement of cortical responsiveness was evident in extrastriate and striate areas. Attention- related activity was demonstrated in area V1 when attention was selectively directed to one side of a moving wedge ( the attention condition) com­pared with passive viewing of the wedge ( the passive condition) ( 71). Activation of area VI was found to be higher in the attention condition. In divided- attention and direction- attention tasks, early visual processing mecha­nisms in the prestriate cortex were apparently influenced by an attentional system in the temporo- parietal areas ( 72). Attention to visual motion ( 70) increased the re­sponsiveness of the motion- selective extrastriate area V5 ( 73,74) and the posterior parietal cortex ( 75). Alternating attention between heard or seen numbers modulated cor­responding activation signals in the auditory and visual cortices ( 76). Emotional arousal, induced when subjects viewed a series of pleasant, neutral, or unpleasant pictures, pro­duced significantly greater occipital activation ( 77). Af­ter determination of ocular dominance by the near- far alignment test, the dominant eye appeared to activate a larger area of area VI than the nondominant eye ( 78). Increased activity in area VI cells accompanied in­creased luminance contrast after stimulation with a single red LED covering 2° of the subject's visual field ( 79). Using a linear systems analysis of responses in area VI, the VI signal increased monotonically with stimulus contrast ( 80). As mentioned previously, the duration of the stimulus may modify the size of the activation signal, but this finding is controversial because it has been iden­tified by some authors ( 54- 59) but not confirmed by others ( 60,61). The reason for this discrepancy has not been explained but may reflect contributions from fac­tors other than the BOLD effect ( e. g., the differential sensitivity of EPI and gradient- echo fMRI to flowing ./ Neiim- Oplillmlnml, Vol. 19, No. .?, 1999 fMRl IN THE VISUAL SYSTEM 191 blood, rather than simply its oxygenation). Most studies have been carried out on young adults and, therefore, the literature does not contain information on the effect of age on the activation signal. LOCALIZATION OF FUNCTIONS Because of the robust and reproducible responses found in the occipital cortex during fMRl, considerable attention has been focused on the striate and the extra-striate cortex. Areal boundaries outlining VI, V2, V3, VP, V3a, V4, and V4v have been defined and higher visual functions such as recognition of objects, colors, faces, and so on have been localized. Because it is im­possible to image the stria of Gennari, a varying contrast technique has been used to recognize activation confined to the striate cortex ( 81). Retinotopic Representation Data from human lesional studies show that neurones within area VI are retinotopically organized after a roughly polar coordinate system ( 82,83). Movement from posterior to anterior in the striate cortex represents the center to the periphery of the visual field, a retino­topic dimension referred to as eccentricity. Movement from the inferior to the superior bank of the calcarine fissure represents shift from the superior vertical merid­ian through the horizontal meridian to the lower vertical meridian, a retinotopic dimension referred to as polar angle. To map polar angle ( i. e., angle from the center of gaze), subjects viewed a slowly rotating semicircular checkerboard stimulus, and to measure eccentricity ( i. e., distance from the center of gaze), an expanding check­ered annulus ( Fig. 4) ( 84). Neurones responding to stimulation at different locations in the visual field were activated at different times during the stimulus sequence, and corresponding differences in the temporal phase of the fMRl response identified the retinotopic location rep­resented by each active site in the brain. Cortical loca­tions of neurones responding to stimulation along the vertical or horizontal visual field meridians were charted on three- dimensional models of the occipital cortex and an unfolded map of the cortical surface produced in six subjects. The topography of visual areas VI, V2, V3, VP, and parts of V3a and V4 were consistent with the orga­nization of these areas in macaque monkeys. Using similar phase- encoded retinal stimulation, vi­sual activation responses were recorded by EPI and ana­lyzed with a Fourier- based method ( 85). The resulting volume data set was then sampled with a cortical surface FIG. 4. Checkered annulus, wedge, and rotating semicircle used to measure eccentricity and polar angle. reconstruction made from high- resolution structural MRI images collected separately from each subject. The cor­tical surface containing the data was unfolded and ana­lyzed with the visual field sign method, which automati­cally and objectively outlines areal boundaries, adjacent areas having the opposite field sign distinguished by mir­ror image and nonmirror image representations. Through a combination of multislice fMRl, stimulus phase-encoding and Fourier analysis, cortical surface recon­struction, and visual field sign, the retinotopic organiza­tion of visual areas V1, V2, VP, V3, and V4 was recon­structed in two dimensions with accurate delineation of their borders. Cortical magnification curves for striate and extra- striate areas were determined and humans ap­peared to have a greater emphasis on the center of gaze than their monkey counterparts. After stimulation with spatially alternating flickering check stimuli in the form of iso- polar angle wedges, iso- eccentricity rings, and circles of equal polar- angle diameter ( Fig. 4), the wedges produced parallel stripes of roughly equal width in V1, the rings of radically varying widths produced stripes of roughly equal width orien­tated approximately orthogonal to them, and the circles produced circular activity patches of roughly equal cor­tical width ( 81). The location of the blind spot was dem­onstrated by testing subjects binocularly and monocu-larly with a field of scaled, black and white flickering checks, when it was found that the blind spot lay just inferior to the cortical representation of the horizontal meridian at around 15° eccentricity. The size of the blind spot was generally consistent with the human cortical magnification factor in area V1 and with the width of the blind spot representation in human histologic material ( 86). This study demonstrated that area VI could be ac­tivated preferentially to extrastriate areas by manipula­tion of luminance contrast and presentation of radial gratings alternating between 6% and 100% contrast. Evi­dence for orientation selectivity in VI was provided by measuring transient f'MRI increases produced at the change in response to gratings of different orientations. The band width of the orientation " transients" was ap­proximately 45°. Using an alternating black and white checkerboard to create simple visual stimuli that generated continuous traveling waves of neural activity in the visual cortex, activity could be localized to within 1.1 mm ( 87). From measurements of the motion of the traveling wave, the borders between retinotopically organized visual areas were identified and striate cortical positions related to visual field eccentricity. The foveal response was appar­ent in the posterior striate cortex and increasingly ante­rior locations responded to increasingly eccentric stimuli. Retinotopically organized responses extended along a 3- to 4- cm strip of striate cortex, although the stimulus only extended 12°. Retinotopic representation agreed with previous human lesional studies in the revised represen­tation of the visual field hypothesis and electrophysi­ologic data from nonhuman primates. The representation of the ipsilaleral visual field in 12 J Neum- Oplilhcilinol, Vol. 19, Nti. .1, IW9 192 R. M. McFADZEAN ET Ah. subjects appeared to reveal activity along the vertical meridian in retinotopic areas and in two large branches anterior to that, in presumptive higher- tier areas ( 88). Human area V3a was found to have a retinotopy similar to that found in macaque monkeys providing a continu­ous map of the contralateral hemifield immediately an­terior to area V3, with a unique retinotopic representation of the upper visual field in the superior occipital cortex ( 89). A view of the retinotopic representation in the occipi­tal cortex is illustrated in Figure 5, constructed from activation changes in healthy volunteers ( 90), and con­trasted with homologous areas in the macaque. The precise representation of specific isopters of the FIG. 5. Location and topography of presumptive visual areas in hu­man and macaque cortex. ( A & B) One representative cortical hemi­sphere of human brain in the normal ( folded) state, based on high-resolution " anatomic" magnetic resonance images. Visual areas have been rendered in pseudocolor onto the surface, based on data from functional magnetic resonance imaging ( fMRI) as described later. ( C) The same anatomic and functional data depicted in " flattened" corti­cal format. Additional areas ( bounded by red borders) are based on data from other subjects and were added for completion. Based on quantitative differences in cortical curvature, the locations of gyri and sulci in the folded brain depicted in A and B are represented in light and dark gray, respectively, in the flattened representation in C. For comparison, ( D) shows the corresponding flat map from macaque monkey, based on previously published data ( 136). Both flattened maps are taken from the right hemisphere, which was artificially split along the length of the calcarine fissure ( approximately the horizontal meridian representation in V1). Borders of human visual areas are presumptive. However, each cortical visual area in A, B, and C has been reliably produced in approximately the same cortical location ( with similar topographic relationships to surrounding areas, likewise defined) in several scan sessions in at least four subjects ( usually many more), in response to the same visual stimulus or set of stimuli. Names for human visual areas have been adopted from apparently corresponding areas in macaque when there is topographic and func­tional evidence of homology ( for example, V1, V2, V3, ventral pos­terior ( VP), V3A, and middle temporal area ( MT or V5); this is quali­fied when evidence for homology is encouraging but undefinite ( for example, in the case of the posterior division of dorsal medial supe­rior temporal area [ pMSTd]). Otherwise, new names have been in­vented ( i. e., lateral occipital [ LO], SPO, and LSPO). Presumptive corresponding areas in the two primates are assigned the same color on the map. The foveal representation is a shared strip connecting most of the retinotopic areas and is centered roughly at the star. Based on fMRI data and inferences from the connectional hierarchy in macaques ( 136), the human areas are grouped into three broad categories: retinotopic ( blues and purples); parietal ( greens); and temporal ( yellow to red)- such categories are heuristic and tentative. Subjects were scanned in a 1.5- Tesla magnetic resonance imager, retrofitted with echo- planar imaging. In each 8- minute 32 second scan, 2,048 images were collected ( repetition time ( TR) = 4 seconds) in multislice mode ( 16 slices; 4- mm thick) at 3 x 3 mm resolution, using a bilateral surface coil ( covering occipital and posterior tempo­ral and parietal lobes) and an asymmetric spin- echo sequence ( echo time [ TE] = 70 milliseconds; offset = 25 milliseconds). Visual stimuli were presented to subjects within the magnet, with an extensive field of view. Phase- encoded, retinotopically varying stimuli ( thinner versions of those described in previous studies) ( 84,137) were used to distinguish polar angle and eccentricity axes in retinotopically organized areas. This information was combined to derive the visual field sign polarity ( 84), which reverses at area boundaries. The boundaries of areas V1, V2, V3, VP, V3A, and ventral V4 ( V4v), mapped in the same individual, were derived in this way. The motion- responsive areas MT ( V5), MSTd, and V3A were selectively activated ( in the same subject) by a pattern of moving concentric rings. A set of black- and- white images of objects or faces, compared with scrambled versions of the same objects or faces, were used to activate area LO selectively, again in the same subject. Some variability was observed in the location of object- selective activity across individuals. Area LSPO responds selectively during saccade tracking, but not fixation, in near- total darkness. Area SPO is activated by the coherent motion of random, low- density dots, compared with random motion of the same set of dots, with otherwise identical motion parameters. In some subjects, area MT ( V5) can be reliably and selectively activated by the motion coherence test and by tests for greater interhemispheric activation. The topography of human cortical areas is generally similar to that in macaques, except for an overall expansion; however, there are some noteworthy differences. 1) Human maps contain a posterior anterior retinotopic area ( V4v) and a nonretinotopic, form- related area ( LO) between MT ( V5) and VP, but there is no clear area border corresponding to the border between V4v and LO in macaque maps. 2) There is proportionately more area between MT ( V5) and foveal V3 and VP in human compared with macaque maps. 3) Human V3 and VP are proportionately several times wider than macaque V3 and VP. The assignment of separate names (' V3' compared with ' VP') to mirror- symmetrical, quarter- field representations ( otherwise known as superior and inferior arms of V3) is based entirely on macaque data; functional differences have not been reported between these two areas. In C, linear and angular distortion caused by flattening is minimized and averages approximately 15% overall ( 84); the corresponding distortion in D is presumably similar ( 138). Scale bars, 1 cm. Reprinted from Tootell RBH, Dale AM, Sereno Ml, Malach R. New images from human visual cortex. In: Trends in Neurosciences, Vol. 19. 1996; 481- 9. With permission from Elsevier Science. J Neum- Oplulmlmol, Vol. 19, No. 3. 1999 fMRI IN THE VISUAL SYSTEM 193 visual field in the striate cortex has not been clarified, but study of the central 11°, 19°, and 56° confirms the in­creased emphasis placed on the representation of the cen­tral 10° in at least 50% of the striate cortex posteriorly according to the revised representation hypothesis ( 91). However, in the 1 1 ° and 19° conditions, some activation spread anteriorly to the parieto- occipital/ calcarine fissure junction, indicative of unknown behavioral patterns in the striate cortex. Clinically, further knowledge of the precise retinotopic organization of area VI could even­tually lead to objective perimetry using fMRI. Color Recognition Activation by an array of six colored circles and their after- images suggested that the fusiform gyri play a criti­cal role in human color perception ( 92). In a repetitive stimulation protocol, isoluminant chromatic or isochro-matic luminance modulation was alternated with steady light of the same mean chromaticity and luminance as a reference condition. Color- sensitive activation was ob­served in collateral sulci and area VI ( 93). Color tuning in response to a large number of colors suggests that color signals relevant for perception are included in a large population of areas VI and V2 neurones. The strongest response is to red- green stimuli receiving op­posing inputs from L and M retinal cones ( 94). Color vision may be considered in terms of perception and imagining of color. Using a colored and grey scale Mon-drian display and contrasting a relative color judgement with a spatial task requiring the generation of mental images, it was shown that color perception activates the posterior fusiform gyrus bilaterally ( V4), plus the right anterior fusiform and lingual gyri, striate cortex ( VI), and left and right insula. Color imagery activated the right anterior fusiform gyrus, left insula, right hippocam­pus, and parahippocampal gyri, but not areas V4 nor VI. The findings suggest that the anterior fusiform and para­hippocampal gyri and the hippocampus are the location for stored representation of colored objects ( 95). Object Recognition Convergence of visual cues in the form of motion, texture, and luminance contrast were demonstrated on the lateral aspect of the occipital lobe ( LO complex) providing strong evidence for its role in object process­ing ( 96). There was preferential activation of LO by im­ages of objects, when compared to a wide range of tex­ture patterns. The " Lincoln" illusion, in which blurring of objects digitized into large blocks paradoxically in­creases their recognizability, significantly enhanced LO activation. However, LO did not seem to be involved in the final stages of the recognition process ( 97). Face Recognition During a face- matching task, there was a significantly increased MR signal in the ventral occipitotemporal cor­tex, extending from the inferior occipital sulcus to the lateral occipito- temporal sulcus and fusiform gyrus ( 98). Selective involvement of the fusiform face area in face perception has also been demonstrated ( 99). Comparing cerebral hemispheres, faces primarily activated the fusi­form gyrus bilaterally, but there was greater activation in the right than in the left hemisphere ( 100). Ocular Dominance Ocular dominance was determined by means of the near- far alignment test, when stimulation of the domi­nant eye activated a larger area of V1 than the nondomi-nant eye ( 78). Ocular dominance in VI reminiscent of the single cell recordings of Hubel and Wiesel was dem­onstrated when seven categories of response were iden­tified, varying from left only to binocular only to right only responses. Using differential techniques neuronal activity in cortical columns raised the possibility of fur­ther mapping of specialized neurones in human visual cortex ( 101). Motion Perception Area MT cortex ( the human homologue of the motion sensitive middle temporal area, MT or V5 of monkeys) responded selectively to moving compared to stationary stimuli, consistent with previous PET and anatomic stud­ies. In addition, area MT has a much higher contrast sensitivity, particularly when compared with area VI. Using color varying stimuli and changes in luminance, activity in MT decreased at and near individually mea­sured equiluminance, compatible with the psychophysi­cal phenomenon that visual motion appears to diminish when moving color varying stimuli are equated in lumi­nance. Activity in area MT appeared much less retino­topic than in areas VI, V2, V3, and VP ( 102). Viewing a stationary stimulus after adaptation to stimuli moving in a single local direction creates a visual motion after­effect of illusionary motion, also known as the waterfall illusion, to which human cortical area MT ( V5) is re­sponsive in a direction- specific manner. The time course of the motion after- effect measured psychophysically was essentially identical to the time course of the fMRI motion after- effect ( 103). Motion perception studied when subjects viewed a stationary black and while grat­ing, a moving grating, and a moving spot generated ac­tivation in the lateral occipitotemporal cortex ( area MT). During pursuit, extraretinal signals may be received be­cause the signal intensity during pursuit of the moving dot was greater than during viewing of the moving grat­ing. This occurred despite the fact that the moving grat­ing generated more retinal image motion, while signal intensity in the striate cortex was least during pursuit of the moving dot ( 104). Each of three different motion displays activated specific parts of the V5 complex, but also activated neighboring, although nonoverlapping, re­gions of the auditory cortex that arc normally activated by the perception of speech ( 105). Motion boundaries produce a boundary specific signal that is retinotopically organized within area VI, but appears to be largely ab­sent from the motion selective area MT- V5 ( 106). A region more activated by kinetic gratings than by lumi­nance- defined gratings, uniform motion, or transparent motion, anatomically and functionally distinct from areas MT- V5, V3, and V3a but minimally overlapping the LO region, has been identified and appears to be genuinely ./ Neiiro- Ophlhulmol, Vol. 19. No. .1. 1999 194 R. M. McFADZEAN ET AL. specialized for processing kinetic boundaries, created by discontinuities in motion direction ( 107). Various motion stimuli have been used to study the effects of first and second order motion, ft appears that first order motion sensitivity is localized to area VI, whereas second order motion is represented in areas V3 and VP, and area V5 and possibly areas V3a and V3b are involved in further processing of motion information, including the integra­tion of motion signals of both types ( 108). Human area V3a appears to be different from its macaque counter­parts in being relatively motion- selective when com­pared with area V3 ( 89). Eye Movements Pursuit eye movements during visual motion percep­tion, while viewing sequences of random dot motion and moving dots under conditions of fixation or pursuit, highlighted a motion- specific area in the lateral occipito­temporal cortex ( V5). There was also activation in a region approximately 12 mm dorsal to area V5 ( 74). Voluntary saccades in light or dark and imagined sac-cades with electro- oculogram monitoring produced sig­nificant activation of the frontal eye field ( precentral and posterior medial frontal gyrus) in all conditions and the supplementary eye field ( superior frontal gyrus) during the voluntary condition. In addition, voluntary saccades activated the primary visual cortex, but imagined sac­cades did not ( 109). During visually guided saccades, bilateral activity was detected in the occipital cortex, the precentral sulcus, and the deep region of the intraparietal sulcus. This intraparietal area borders areas 39 and 40 of Brodmann and apparently corresponds to the human pa­rietal eye field ( 110). Other Functions Visual perception and visual imagery may share a common neural anatomic substrate. During recall of a visual stimulus, focal signal changes were detected in areas VI and V2. It appears that processes involved in visual perception may also be applicable to visual imag­ery ( 111). When viewing the corners of a Kanizsa square, specific extrastriate regions primarily in the right hemisphere responded to illusory contour perception, an important aspect of perceptual grouping ( 112). On pre­sentation of dissimilar images to the two eyes, perception alternates spontaneously between each monocular view ( i. e., binocular rivalry occurs). Fronto- parietal areas ap­pear to play a central role in conscious perception biasing the content of visual awareness toward abstract internal representations of visual scenes, rather than simply to­ward space ( 113). Mental image generation asymmetri­cally activated the visual association cortex on the left side ( 114). During a test of differential sensitivity to faces, letter strings, and textures, faces primarily acti­vated the fusiform gyrus bilaterally, letter strings the left occipitotemporal and inferior occipital sulci, and tex­tures portions of the collateral sulcus. Thus, different regions of ventral extrastriate cortex are specialized for processing facial features and letter strings and are in­termediate between earlier processing in striate and peristriate cortex and later lexical, semantic, and asso­ciative processing in downstream cortical regions ( 100). String length contrast alone was sufficient to account for the activation pattern in the medial visual cortex by word- like stimuli, when contrasted with single characters ( 115). METABOLIC EFFECTS The behavior of the fMRI signal in the visual cortex has significant implications for our understanding of brain metabolism. PET studies have demonstrated that visual activation results in an increased CBF and hyper-oxygenation, without a corresponding increase in the oxygen extraction fraction, but accompanied by an in­creased cerebral glucose metabolic rate ( 31). These fea­tures are compatible with anaerobic glycolysis. Further­more, MRS studies on visual stimulation demonstrate increased glucose utilization with the accumulation of lactate in the visual cortex, which on continued stimula­tion subsequently declines, again indicative of anaerobic glycolysis initially ( 116,117). Therefore, it has been pro­posed that the initial MR activation signal represents an uncoupling between regional CBF and oxygenation and oxidative metabolism ( i. e., anaerobic glycolysis), but the subsequent return of the activation signal toward base­line at an interval that varies from 1 to 5 minutes is indicative of a return to oxidative metabolism through cerebral autoregulation ( 54- 59). During prolonged activation of the visual cortex by a video presentation, CBO- and CBF- sensitive fMRI re­cordings were monitored separately ( 54). Oxygen-sensitive recordings displayed an initial signal increase, followed by a subsequent signal decrease extending over 4 to 5 minutes, and a signal drop at the end of stimula­tion. However, flow sensitive recordings demonstrated that the in- flow effect remained elevated during the en­tire stimulation period. Thus, gradually decreasing CBO despite persisting elevation of CBF reflected an accumu­lation of deoxyhemoglobin caused by progressive up-regulation of oxidative metabolism after prolonged stimulation. Dynamic uncoupling and recoupling of per­fusion and oxidative metabolism were observed by mea­suring changes in glucose consumption, lactate produc­tion, and CBO during prolonged neuronal activation using dynamic MRS and fMRI ( 57). After visual stimu­lation, a decrease of steady- state glucose by 40% because of enhanced use by 21 % was accompanied by a transient accumulation of lactate 2.5 minutes after stimulation on­set. Again, nonoxidative glucose metabolism during functional activation was gradually complemented by a slow return to oxidative metabolism, with recoupling of perfusion and oxygen consumption at a new equilibrium. After stimulation of the visual cortex with photic stimuli of varying duration, a postactivation undershoot may oc­cur during fMRI ( 54). However, the absence of an un­dershoot after longer stimulation periods provides further evidence of a gradual shift from an uncoupling between regional CBF and oxygen consumption toward a steady-state ( 118). ./ Neitro- Ophtlwlnml, Vol. 19, No. J, 1999 fMRI IN THE VISUAL SYSTEM 195 By contrast, other studies have suggested that the CBF and oxygen consumption remain constant during the en­tire time that primary visual cortical neurones are acti­vated ( 60,61). In an fMRI study sensitized to flow and oxygenation changes as well as PET- sensitized to flow ( 60), several types of visual stimulation were used and flow and oxygenation were evaluated in separate time-course series as well as simultaneously using two differ­ent MRI methods. In most cases, the activation- induced increase in flow and oxygenation remained elevated for the entire stimulus duration, suggesting that flow rate and oxygen consumption remained constant during the acti­vation period. In addition, a prolonged visual stimulation study using different MRI techniques and two different visual stimuli showed that the signal- time course from areas of significant activation remained largely elevated throughout the duration of stimulation, unaffected by the imaging method used ( 61). These data suggested that recoupling between blood flow and oxygen extraction was not a general phenomenon during extended visual stimulation. The interpretation of the discrepancy in these findings is difficult, especially because different studies use dif­ferent experimental paradigms and imaging sequences. However, this is an important issue that may, in particu­lar, offer an explanation for the nature of the fMRI sig­nal. Hopefully, further metabolic studies will provide a solution. CLINICAL USAGE fMRI may be used in preoperative neurosurgical plan­ning. Two patients with complex partial seizures refer­able to the temporal and occipital regions showed sig­nificant signal intensity changes up to 15% between the activated and resting conditions during repetitive photic stimulation near the surgical targets ( 21). Systematic fMRI studies in patients with vascular malformations adjacent to the primary visual cortex demonstrated dis­placement of the activated region and hemispheric asym­metry in the number of activated voxels in the functional region, which appeared to reflect the anatomic and physi­ologic impact of the vascular malformation. Changes in fMRI findings after intervention monitored the conse­quences of therapy and paralleled clinical recovery ( 23). After fMRI in two patients with a homonymous hemi-anopia, one after an established cerebral infarction and the other during the recovery phase from an episode of multiple sclerosis to a normal visual field, it was con­cluded that fMRI activation may prove to be a useful way of objectively measuring visual function. In the pa­tient with infarction, the signal change was decreased on the affected side, whereas in the patient with multiple sclerosis the activations were within normal limits ( 119). In five patients with homonymous hemianopic visual field loss caused by retrochiasmal lesions, activation ab­normalities were compatible with the visual field defect in three patients, but two patients with macular sparing showed symmetric responses ( 120). Dynamic suscepti­bility contrast imaging, also known as relative cerebral blood volume mapping, may provide further information about the hemodynamics after strokes. In a study of five patients with occipital infarction with visual field analy­sis, cortical activation mapping and dynamic susceptibil­ity contrast imaging during full field visual stimulation demonstrated that fMRI techniques can accurately map functional and perfusion deficits and provide useful clinical information ( 14). Abnormal brain tissue function may extend beyond the limit of an infarct seen on con­ventional imaging, but by determining areas of spared cortex fMRI techniques may help to assess the efficacy of therapeutic interventions and improve rehabilitation strategies. After severe craniocerebral trauma affecting the optic radiations, a patient developed an incomplete macular splitting homonymous hemianopia with blind sight. The visual responsiveness of deafferrentcd VI was examined, when ipsilesional VI displayed no stimulus-related signal change, but activation was observed in ipsilesional extrastriate cortex ( 121). Thus, blind sight does not appear to depend on deafferrented striate cortex. In seven patients with visual field loss caused by le­sions of the optic nerve and the optic chiasm, there ap­peared to be a correlation with failure of activation in that part of the striate cortex corresponding to the central visual field defect ( 122). In a patient with optic atrophy but central visual field sparing, it was claimed that the cortical activation pattern corresponded to the visual field loss ( 123), but this finding has been disputed largely on the grounds of anatomic localization ( 124). To determine the pathophysiology of dyslexia, fMRI was used to study visual motion processing in normal and dyslexic men ( 19). In all of the dyslexic patients, presentation of moving stimuli did not produce the same task- related functional activation in area V5- MT ob­served in controls. Stationary patterns produced equiva­lent activation in both groups. The relationship between brain activity and reading performance was explored in dyslexic patients, who showed reduced activity com­pared with controls in the primary visual cortex and in a secondary cortical visual area ( MT+) believed to receive a strong magno- cellular ( M)- pathway input ( 125). Sig­nificant correlations were found between individual dif­ferences in reading rate and brain activations. The results support the hypothesis for an M- pathway abnormality in dyslexia and imply a strong relationship between the integrity of the M- pathway and reading ability. Consistent with a chiasmal crossing anomaly in albi­nism is the finding in albino patients that monocular stimulation caused predominantly contralateral activa­tion, whereas control subjects had symmetrical patterns of activation ( 20). In patients with schizophrenia, the mean signal inten­sity change in the primary visual cortex was significantly greater on photic stimulation than in normal controls, possibly reflecting structural brain changes or impair­ment of mitochondrial function or energy metabolism ( 16). In one patient with cortical Lewy body dementia who experienced persistent and vivid complex visual hallucinations, photic stimulation produced a normal bi- .1 Neuro- Ophllmlimil, Vol. 19. No. .(, 1999 196 R. M. McFADZEAN ET AL. lateral activation response in area VI when he was not hallucinating, but limited activation during hallucina­tions ( 17). Depressed subjects were studied while using positively and negatively balanced visual stimuli before and alter venlafaxine ( 18). Preliminary results suggested that I'MRI would be useful to study emotional processes in normal and depressed subjects and to examine the mechanisms of action of antidepressant drugs. These studies are indicative of I'MRI's potential for assessing brain abnormalities in psychiatric disorders and monitor­ing treatment. Pharmacologic induction of vasodilatation by acet-azolamide attenuated signal changes under photic stimu­lation, but there appeared to be a persisting autoregula-tory responsiveness to functional challenge ( 126). After cocaine infusion, it was possible to map dynamic pat­terns of brain activation and provide evidence of dynami­cally changing brain networks associated with cocaine-induced euphoria and craving, which affected a number of brain regions including the occipital cortex ( 127). Al­though cocaine infusion produced a small but definite decrease in global cortical CBF, visual stimulation re­sulted in comparable signal increases in visual cortex before and after cocaine and saline infusion ( 128). The elucidation of the effects of drugs on the central nervous system could be an important future use of fMRI. OTHER FUNCTIONAL IMAGING TECHNIQUES While detailed comparison with other functional im­aging techniques in the visual cortex is beyond the scope of this review, it is possible to profitably allude to a few studies. Comparison with PET studies are interesting because it is known that the PET signal measures intraparenchy-mal changes in CBF. Primary visual cortical activation was evaluated using fMRI and H2 - l 50 PET in six male subjects visually stimulated by means of red LED flash goggles. The PET technique demonstrated substantially greater relative signal change on visual stimulation than the fMRI technique. The fMRI signal changes were con­centrated in loci around the periphery of brain paren­chyma exhibiting increased radiotracer activity. It was concluded that signal changes using fMRI based on gra­dient- echo techniques reflected primarily phenomena oc­curring within small veins and under- represented activity intrinsic to brain parenchyma, while PET directly mea­sured changes in metabolically related activity within the parenchyma. This finding suggested potential inaccura­cies in localization of activated brain tissue using fMRI ( 129). However, functional activation maps for the visual cortex have been obtained at a spatial resolution almost two orders of magnitude greater than achievable by PET and within measuring times of a few seconds using gra­dient- echo imaging at 2.0 Tesla ( 130). Cortical areas associated with face perception identified by PET and fMRI were in similar anatomic locations, but in addition fMRI revealed interindividual variation with greater ana­tomic precision ( 98). Comparison of activation findings with MEG and fMRI indicates that these techniques offer different, largely complementary capabilities. While fMRI offers excellent spatial resolution of < 1 mm, MEG has a tem­poral resolution measured in milliseconds. It is possible to treat MEG and fMRI data within a unified computa­tional framework, fMRI precisely defining the locations of activation to regions and MEG the temporal dynamic ( 131). MEG studies of retinotopic representation in con­junction with anatomic MRI correlation have been car­ried out by presenting small pattern stimuli near the ver­tical and horizontal meridians. The results suggested that the representation of the horizontal meridian did not nec­essarily correspond in a one- to- one manner with the base of the calcarine fissure. Significant individual variability in the details of how area VI maps around the calcarine fissure was identified ( 132). FURTHER DEVELOPMENTS Using flow- sensitive fMRI with steady- state arterial water labeling, it is possible to measure the relative CBF increase during brain activation. On visual stimulation, CBF in the visual cortex increased by 17% to 35%. Such quantification facilitates comparative fMRI studies under different conditions ( 133). The use of magnetically vari­able tissue water proton spins as a freely diffusible tracer provides maps of absolute CBF changes ( delta CBF). The individual mean values of delta CBF measured in five healthy volunteers ranged from 69 ± 18 to 99 ± 26 ml/ minute per 100 g on visual stimulation with an alter­nating black and white checkerboard at 8 Hz ( 134). Since the advent of fMRI, there has been a continual improvement in spatial and temporal resolution. Using the latest scanners, it is possible to record activated sig­nals with a spatial resolution of < 1 mm and a temporal resolution of < 1 second. Continued technologic develop­ment may enhance this performance. However, further investigations are required to localize in terms of mi-crovessels or macrovessels the magnetic effect respon­sible for the fMRI image and to develop understanding of the coupling between neuronal activity and CBF and oxygenation. Comparison with PET, MRS, and NIRS may elucidate some of the cerebral metabolic changes in response to activation experiments and in disease states. While functional imaging studies can localize cortical activations in response to different experimental para­digms, it is impossible to image the white matter inter­connecting pathways ( e. g., the fronto- parieto- occipital connections during viewing of a moving scene). Never­theless, with improved temporal resolution, it may be possible to infer the nature of higher cortical connections during activation processes. In general, fMRI is con­strained to investigation of sustainable and reproducible activation, but attempts are being made to examine single events rather than repeated stimuli by more resolvable temporal analysis. The technique of magnetic source im­aging combines the high spatial resolution of fMRI and MRI with the high temporal resolution of MEG in this respect ( 135). J Neiim- Ophlhalniol, Vol. 19, No. 3, 1999 fMRI IN THE VISUAL SYSTEM 197 REFERENCES 1. Bellivcau JW, Kennedy DN Jr, McKinstry RC, cl al. Functional mapping of the human visual cortex by magnetic resonance im­aging. Science 1991; 254: 716- 9. 2. Rosen BR, Bellivcau JW, Aroncn HJ, et al. Susceptibility contrast imaging of cerebral blood volume: human experience. Magn Re-son Med 1991; 22: 293- 9. 3. Bclliveau JW, Kwong KK, Kennedy DN, et al. Magnetic reso­nance imaging mapping of brain function. Human visual cortex. Invest Radiol l992; 27( Suppl 2): S59- 65. 4. Kwong KK, Bclliveau JW, Chesler DA, et al. Dynamic magnetic resonance imaging of human brain activity during primary sen­sory stimulation. Proc Nail Acad Sci USA 1992; 89: 5675- 9. 5. Menon RS, Owaga S, Kim SG, et al. Functional brain mapping using magnelic resonance imaging. Signal changes accompany­ing visual stimulation. Invest Radiol 1992; 27( Suppl 2): S47- 53. 6. Ogawa S, Tank DW, Menon R, et al. Intrinsic signal changes accompanying sensory stimulation: functional brain mapping with magnetic resonance imaging. Proc Natl Acad Sci USA 1992; 89: 5951- 5. 7. Thulborn KR, Chang SY, Schen GX, Voyvodic JT. High-resolution echo- planar fMRI of human visual cortex at 3.0 tcsla. NMR Biomed 1997; 10: 183- 90. 8. Schad LR, Weiner E, Baudendistel KT, Mullcr E, Lorenz WJ. Event- related functional MR imaging of visual cortex stimulation at high temporal resolution using a standard 1.5 T imager. Magn Reson Imaging 1995; 13: 899- 901. 9. Konishi S, Yoneyama R, Itagaki H, et al. Transient brain activity used in magnetic resonance imaging to detect functional areas. Neuroreporl 1996; 8: 19- 23. 10. Constable RT, Kennan RP, Puce A, McCarthy G, Gore JC. Func­tional NMR imaging using fast spin- echo at 1.5 T. Magn Reson Med 1994; 31: 686- 90. 11. 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- 20. 12. Kim SG, Hu X, Adriany G, Ugurbil K. Fast interleaved echo-planar imaging with navigator: high resolution anatomic and functional images at 5 Tcsla. Magn Reson Med 1996; 35: 895- 902. 13. Hathout GM, Gambhir SS, Gopi RK, et al. A quantitative physi­ologic model of blood oxygenation for functional magnetic reso­nance imaging. Invest Radiol 1995; 30: 669- 82. 14. Sorenscn AG, Wray SH, Weisskoff RM, et al. Functional MR of brain activity and perfusion in patients with chronic cortical stroke. AJNR 1995; 16: 1753- 62. 15. Buchel C, Price C, Frackowiak RS, Friston K. Different activa­tion patterns in the visual cortex of late and congenitally blind subjects. Brain 1998; 121: 409- 19. 16. Rcnshaw PF, Yurgelun- Todd DA, Cohen BM. Greater hemody­namic response to photic stimulation in schizophrenic subjects: an echo- planar MRI study. Am J Psychiatry 1994; 151: 1493- 5. 17. Howard R, David A, Woodruff P, et al. Seeing visual hallucina­tions with functional magnetic resonance imaging [ Review]. De­ment Gerialr Cogn Disord 1997; 8: 73- 7. 18. Kalin NH, Davidson RJ, Irwin W, ct al. Functional magnetic resonance imaging studies of emotional processing in normal and depressed patients: effects of venlafaxine. J Clin Psychiatry l997; 58( Suppl l6): 32- 9. 19. Eden GF, Van Meter JW, Rumsey JM, Maisog JM, Woods RP, Zcffiro TA. Abnormal processing in visual motion in dyslexia revealed by functional brain imaging. Nature 1996; 382: 66- 9. 20. Hcdera P, Lai S, Haacke EM, et al. Abnormal connectivity of the visual pathways in human albinos demonstrated by susceptibility-sensitized MRI. Neurology 1994; 44: 1921- 6. 21. Fried I, Nenov VI, Ogemann SG, Woods RP. Functional MR in PET imaging of rolandic and visual cortices for neuro- surgical planning. J Neurosurg 1995; 83: 854- 61. 22. Latchaw RE, Hu X. Functional MR imaging in the evaluation of the patient with epilepsy. Functional localization [ Review]. Neu-roimag Clin N Am 1995; 5: 683- 93. 23. Schlosser MJ, McCarthy G, Fulbright RK, Gore JC, Awad IA. Cerebrovascular malformations adjacent to sensorimotor and vi­sual cortex. Functional magnetic resonance imaging studies be­fore and after therapeutic intervention. Stroke I997; 28: l 130- 7. 24. Hu X, Kim SG. Reduction of signal fluctuation in functional MRI using navigator echoes. Magn Reson Med 1994; 31: 495- 503. 25. Weisskoff RM. Functional MRI: are we all moving towards ar-tifactual conclusions? Or fMRI fact or fancy? [ Review| NMR Biomed 1995; 8: 101- 3. 26. Hu X, Le TH, Parrish T, Erhard P. Retrospective estimation and correction of physiological fluctuation in functional MRI. Magn Reson Med 1995; 34: 201- 12. 27. Pauling L, Coryell C. The magnetic properties of and structure of hemoglobin, oxyhemoglobin and carbonmonoxyhemoglobin. Proc Natl Acad Sci USA 1936; 22: 210- 6. 28. Thulborn KR, Waterton JC, Mathews PM, cl al. Oxygenation dependence of the transverse relaxation time of water protons in whole blood at high field. Biochem Biophys Acta 1992; 7I4: 265- 70. 29. Ogawa S, Lee T- M. Magnetic resonance imaging of blood vessels at high fields: in vivo and in vitro measurements and image simulation. Magn Reson Med 1990; 16: 9- 18. 30. Fox PT, Raichle ME. Focal physiological uncoupling of cerebral blood flow and oxidative metabolism during somatosensory stimulation in human subjects. Proc Natl Acad Sci USA 1986; 83: 1140- 4. 31. Fox PT, Raichle ME, Minttin MA, Deuce C. Non- oxidative glu­cose consumption during focal physiologic neural activity. Sci­ence 1988; 241: 462- 4. 32. Jezzard P. Functional magnetic resonance imaging of the brain: principles, pp. 296- 297. In: Lc Bihan D, Moderator. Functional magnetic resonance imaging of the brain. Ann Intern Med 1995; 122: 296- 303. 33. Menon RS, Ogawa S, Tank DW, Ugurbil K. Tcsla gradient re­called echo characteristics of photic stimulation- induced signal changes in the human primary visual cortex. Magn Reson Med 1993; 30: 380- 6. 34. Ugurbil K, Garwood M, Ellcrmann J, et al. Imaging at high magnetic fields: initial experiences at 4 T. Magn Reson Q 1993; 9: 259- 77. 35. Kim SG, Ugurbil K. Comparison of blood oxygenation and ce­rebral blood flow effects in fMRI: estimation of relative oxygen consumption change. Magn Reson Med 1997; 38: 59- 65. 36. Courtney SM, Ungerleider LG. What fMRI has taught us about human vision [ Review]. Curr Opin Neurobiol 1997; 7: 554- 61. 37. Hajnal JV, Myers R, Oatridge A, et al. Artefacts due to stimulus correlated motion in functional imaging of the brain. Magn Reson Med 1994; 31: 283- 91. 38. Hajnal JV, Bydder GM, Young 1R. Stimulus- correlated signals in functional MR of the brain. AJNR I996; 17: 101 1- 2. 39. Yetkin FZ, Haughton VM, Cox RW, el al. Effect of motion outside the field of view on functional MR. AJNR 1996; 17: 1105- 9. 40. Ogawa S, Menon RS, Tank DW, ct al. Functional brain mapping by blood oxygenation level- dependent contrast magnetic reso­nance imaging. A comparison of signal characteristics with a biophysical model. Biophys J 1993; 64: 803- 12. 41. Turner R, Jezzard P, Wen H, cl al. Functional mapping of the human visual cortex at 4 and 1.5 tcsla using deoxygenation con­trast EPI. Magn Reson Med 1993; 29: 277- 9. 42. Constable RT, McCarthy G, Allison T, Anderson AW, Gore JC. Functional brain imaging at 1.5 T using conventional gradient echo MR imaging techniques. Magn Reson Imaging I993; ll: 451- 9. 43. Thompson RM, Jack CR Jr, Butts K, et al. Imaging of cerebral activation at 1.5 T: optimizing a technique for conventional hard­ware. Radiology 1994; 190: 873- 7. 44. Wiener E, Schad LR, Baudendistel KT, Essig M, Mullcr E, Lorenz WJ. Functional MR imaging of visual and motor cortex stimulation at high temporal resolution using a FLASH technique on a standard 1.5 Tesla scanner. Magn Reson Imaging 1996; 14: 477- 83. 45. Lundervold A, Er. sland L, Gjesdal KI, el al. Functional magnelic J Neiiro- Ophtlmlmol, Vol. 19, No. .1, 1999 198 R. M. McFADZEAN ET AL. resonance imaging of primary visual processing using a 1.0 Tesla scanner [ Review], IntJ Neurosci 1995; 81: 151- 68. 46. Santosh CG, Rimmington JE, Best J. I. Functional magnetic reso­nance imaging at 1 T: motor cortex, supplementary motor area and visual cortex activation. Br J Radiol 1995; 68: 369- 74. 47. Loenneker T, Hennel F, Hcnnig J. Multislice interleaved excita­tion cycles ( MUSIC): an efficient gradient- echo technique for functional MRI. Magn Reson Med 1996; 35: 870- 4. 48. Duyn JH, Mattay VF, Sexton RH, et al. 3- dimensional functional imaging of human brain using echo- shifted FLASH MRI. Magn Reson Med 1994; 32: 150- 5. 49. Cornelisscn FW, Pelli DG, Farell B, Huckins SC, Szevcrcnyi NM. A binocular fiberscope for presenting visual stimuli during l'MRI IRcvicw]. Spat Vis 1997; l 1: 75- 81. 50. Dc Yoc EA, Bandcttini P, Neitz .1, et al. Functional magnetic resonance imaging ( FMR1) of the human brain../ Neurosci Meth­ods 1994; 54: 171- 87. 51. Friston K, Jczzard P, Turner R. The analysis of functional MRI time scries. Hum Brain Map 1994; 1: 153- 71. 52. Chin MJ, Yang JR, Huang KM, Chen JH. Functional MRI of the activated human visual cortex. Taiwan I Hseuh Hut Tsa Chili 1996; 95: 396- 8. 53. Moser E, Diemling M, Baumgartner R. Fuzzy clustering of gra­dient- echo functional MRI in the human visual cortex. Part II: quantification, JMRI 1997; 7: 1102- 8. 54. Krugcr G, KJeinschmidt A, Frahm J. Dynamic MRI sensitized to cerebral blood oxygenation and flow during sustained activation of human visual cortex. Magn Reson Med 1996; 35: 797- 800. 55. Frahm J, Bruhn H, Mcrboldt K, Hanicke W. Dynamic MR im­aging of human brain oxygenation during rest and photic stimu­lation. JMRI 1992; 2: 501- 5. 56. Hathout GM, Kirlew KAT, So GJK, ct al. MR imaging signal response to sustained stimulation in human visual cortex. JMRI 1994; 4: 537- 43. 57. Frahm J, Kruger G, Mcrboldt K, Kleinschmidt A. Dynamic un­coupling and re- coupling of perfusion and oxidative metabolism during focal brain activation in man. Magn Reson Med 1996; 35: 143- 8. 58. Condon B, McFadzcan R, Hadley DM, Bradnam MS, Shahani U. Habituation- like effects cause a significant decrease in response in MRI neuroactivation during visual stimulation. Vision Res 1997; 37: 1243- 7. 59. Fransson P, Krugcr G, Mcrboldt KD, Frahm J. A comparative FLASH and EPI study of repetitive and sustained visual activa­tion. NMR Biomed 1997; 10: 204- 7. 60. Bandcttini PA, Kwong KK, Davis TL, et al. Characterisation of cerebral blood oxygenation and flow changes during prolonged brain activation. Hum Brain Map 1997; 5: 93- 109. 61. Howseman AM, Porter DA, Hutton C, Josephs O, Turner R. Blood oxygenation level dependent signal time courses during prolonged visual stimulation. Magn Reson Imaging 1998; 16: 1 - II. 62. Moser E, Teichtmcistcr C, Diemling M. Reproducibility and post­processing of gradient- echo functional MRI to improve localisa­tion of brain activity in the human visual cortex. Magn Reson Imaging 1996; 14: 567- 79. 63. Rombouts SA, Barkhof F, Hoogcnraad FG, Sprenger M, Valk J, Scheltens P. Tcst- retest analysis with functional MR of the acti­vated area in the human visual cortex. AJNR 1997; 18: 1317- 22. 64. Baumgartner R, Scarth G, Teichtmcistcr C, Somorjai R, Moser E. Fuzzy clustering of gradient- echo functional MRI in the human visual cortex. Part I: reproducibility. JMRI 1997; 7: 1094- 101. 65. Rombouts SA, Barkhof E, Hoogenraad FG, Sprcnger M, Schel­tens P. Within- subject reproducibility of visual activation patterns with functional magnetic resonance imaging using multislice echo planar imaging. Magn Reson Imaging 1998; 16: 105- 13. 66. Schneider W, Noll DC, Cohen . ID. Functional topographic map­ping of the cortical ribbon in human vision with conventional MRI scanners. Nature 1993; 365: 150- 3. 67. Van Essen DC, Drury HA. Structural and functional analyses of human cerebral cortex using a surface- based atlas. ./ Neurosci 1997; 17: 7079- 102. 68. Hcdera P, Wu D, Collins S, et al. Sex and electroenccphalo-graphic synchronisation after photic stimulation predict signal changes in the visual cortex on functional MR images. AJNR 1998; 19: 853- 7. 69. Levin JM, Ross MH, Mendelson JH, Mcllo NK, Cohen BM, Renshaw PF. Sex differences in blood- oxygenation- level-dependent functional MRI with primary visual stimulation. Am J Psychiatry 1998; 155: 434- 6. 70. Buchel C, Josephs O, Rees G, Turner R, Frith CD, Friston KJ. The functional anatomy of attention to visual motion. A func­tional MRI study. Brain 1998; 121: 1281- 94. 71. Watanabe T, Sasaki Y, Miyauchi S, et al. Attention- regulated activity in human primary visual cortex. J Neuro- physiol 1998; 79: 2218- 21. 72. Fink GR, Halligan PW, Marshall JC, Frith CD, Frackowiak RS, Dolan RJ. Neural mechanisms involved in the processing of glob­al and local aspects of hierarchically organised visual stimuli. Brain 1997; 120: 1779- 91. 73. O'Craven KM, Rosen BR, Kwong KK, Trcisman A, Savoy RL. Voluntary attention modulates f'MRI activity in human MT- MST. Neuron 1997; 18: 591- 8. 74. Freitag P, Greenlee MW, Lacina T, Scheffler K, Radu EW. Effect of eye movements on the magnitude of functional magnetic reso­nance imaging responses in extra- striate cortex during visual mo­tion perception. Exp Brain Res 1998; 119: 409- 14. 75. Buchel C, Friston KJ. Modulation of connectivity in visual path­ways by attention: cortical interactions evaluated with structural equation modelling and ( MRI. Cereb Cortex 1997; 7: 768- 78. 76. Woodruff PW, Benson RR, Bandettini PA, et al. Modulation of auditory and visual cortex by selective attention is modality-dependant. Neuroreport 1996; 7: 1909- 13. 77. Lang PJ, Bradley MM, Fitzsimmons JR, et al. Emotional arousal and activation of the visual cortex: an l'MRI analysis. Psycho-physiology 1998; 35: 199- 210. 78. Rombouts SA, Barkhof F, Sprcnger M, Valk J, Scheltens P. The functional basis of ocular dominance: functional MRI ( l'MRI) findings. Neurosci Lett 1996; 221: 1- 4. 79. Goodyear BG, Menon RS. Effect of luminance contrast on BOLD fMRl response in human primary visual areas. ./ Neurophysiol 1998; 79: 2204- 7. 80. Boynton GM, Engel SA, Glover GH, Hccger DJ. Linear systems analysis of functional magnetic resonance imaging in human V I. ./ Neurosci 1996; 16: 4207- 21. 81. Tootell RBH, Hadjikhani NK, Vanduffel W, et al. Functional analysis of primary visual cortex ( VI) in humans. Proc Natl Acad Sci USA 1998; 95: 811- 7. 82. Horton JC, Hoyt WF. The representation of the visual field in human striate cortex. A revision of the classic Holmes map. Arch Ophthalmol 1991; 109: 816- 24. 83. McFadzean R, Brosnahan D, Hadley D, Mutlukan E. Represen­tation of the visual field in the occipital striate cortex. Br J Oph­thalmol 1994; 75: 185- 90. 84. De Yoe EA, Carman GJ, Bandcttini P, et al. Mapping striate and extrastriate visual areas in human cerebral cortex. Proc Natl Acad Sci USA 1996; 93: 2382- 6. 85. Sereno MI, Dale AM, Reppas JB, ct al. Borders of multiple visual areas in humans revealed by functional magnetic resonance im­aging. Science 1995; 268: 889- 93. 86. Horton JC, Dagi LR, McCrane EP, De Monastcrio FM. Arrange­ment of ocular dominance columns in human visual cortex. Arch Ophthalmol 1990; 108: 1025- 31. 87. Engel SA, Glover GH, Wandell BA. Retinotopic organisation in human visual cortex and the spatial precision of functional MRI. Cereb Cortex 1997 ; 7: 181 - 92. 88. Tootell RBH, Mendola JD, Hadjikhani NK, Liu A, Dale AM. The representation of the ipsilateral visual field in human cerebral cortex. Proc Natl Acad Sci USA 1998; 95: 818- 24. 89. Tootell RB, Mendola JD, Hadjakhani NK, et al. Functional analy­sis of V3a and related areas in human visual cortex. J Neurosci 1997; 17: 7060- 78. 90. Tootell RBH, Dale AM, Sereno MI, Malach R. New images from human visual cortex. Trends Neurosci 1996; 19: 481- 9. 7 Neiiro- Ophllialiliol, Vol. 19, No. .1 1999 fMRI IN THE VISUAL SYSTEM 199 91. Condon BC, McFad/. ean RM, Barr DB, Dai D, Hadley DM. Further observations on the representation of the visual field in the occipital cortex using functional MRI ( abstract). Neuro-ophthalmology 1998; 20: ll. 92. Sakai K, Watanabe E, Onodera Y, ct al. Functional mapping of the human colour centre with echo- planar magnetic resonance imaging. Proc Roy Soc Loud ( Biol) 1995; 261: 89- 98. 93. Kleinschmidt A, Lee BB, Requardl M, Frahm J. Functional map­ping of color processing by magnetic resonance imaging of re­sponses to selective P- and M- pathway stimulation. Exp Brain Res 1996; 110: 279- 88. 94. Engel S, Zhang X, Wandell B. Colour tuning in human visual cortex measured with functional magnetic resonance imaging. Nature 1997; 388: 68- 71. 95. Howard RJ, Ffylche DH, Barnes J, et al. The functional anatomy of imagining and perceiving colour. Neuroreporl I998; 9: 10I9- 23. 96. Grill- Spector K, Kushnir T, Edelman S, Itzchak Y, Malach R. Cue- invariant activation in object- related areas of the human oc­cipital lobe. Neuron 1998; 21:' l91- 202. 97. Malach R, Reppas . IB, Benson RR, et al. Object- related activity revealed by functional magnetic resonance imaging in human occipital cortex. Proc Null Acad Sci USA 1995; 92: 8135- 9. 98. Clark VP, Keil K, Maisog JM, Courtney S, Ungerleidcr LG, Haxby . IV. Functional magnetic resonance imaging of human visual cortex during face matching: a comparison with positron emission tomography. Neuroimage 1996; 4: 1- 15. 99. Kanwisher N, McDcrmolt J, Chun MM. The fusiform face area: a module in human cxtrastriatc cortex specialised for face per­ception. ./ Neurosci 1997; 17: 4302- 11. 100. Puce A, Allison T, Asgari M, Gore JC, McCarthy G. Differential sensitivity of human visual cortex lo faces, letter strings, and textures: a functional magnetic resonance imaging study. J Neu­rosci 1996; 16: 5205- 15. 101. Menon RS, Ogawa S, Strupp . IP, Ugurbil K. Ocular dominance in human VI demonstrated by functional magnetic resonance im­aging. .1 Neurophysiol 1997; 77: 2780- 7. 102. Tootell RB, Reppas . IB, Kwong KK, et al. Functional analysis of human MT and related visual cortical areas using magnetic reso­nance imaging. J Neurosci 1995; 15: 3215- 30. 103. Tootell RB, Reppas JB, Dale AM, et al. Visual motion after- effect in human cortical area MT revealed by functional magnetic reso­nance imaging. Nalure 1995; 375: 139- 41. 104. Barton JJ, Simpson T, Kiriakopoulos E, ct al. Functional MRI of lateral occipitotemporal cortex during pursuit and motion percep­tion. Ann Neurol 1996; 40: 387- 98. 105. Howard RJ, Brammer M, Wright 1, Woodruff PW, Bulmore ET, Zeki S. A direct demonstration of functional specialisation within motion- related visual and auditory cortex of the human brain. Curr Biol 1996; 6: 1015- 9. 106. Reppas JB, Niyogi S, Dale AM, Sereno MI, Tootell RB. Repre­sentation of motion boundaries in retinotopic human visual cor­tical areas. Nature 1997; 388: 175- 9. 107. Van Oostcnde S, Sunaert S, Van Hecke P, Marchal G, Orban GA. The kinetic occipital ( KO) region in man: an fMRI study. Cereb Cortex 1997; 7: 690- 701. 108. Smith AT, Greenlee MW. Singh KD, Kraemer FM, Hennig J. The processing of first- and second- order motion in human visual cor­tex assessed by functional magnetic resonance imaging ( fMRI). J Neurosci 1998; 18: 3816- 30. 109. Bodis- Wollner I, Bucher SF, Seelos KC, Paulus W, Reiser M, Oertel WH. Functional MRI mapping of occipital and frontal cortical activity during voluntary and imagined saccades. Neu­rology 1997; 49: 416- 20. 110. Muri RM, Iba- Zizen MT, Derosier C, Cabanis EA, Pierrol- Deseilligny C. Location of the human parietal eye field with functional magnetic resonance imaging. .1 Neurol Neurosurg Psy­chiatry 1996; 60: 445- 8. 111. Lc Bihan D, Turner R, Zeffiro TA, Cuenod CA, Jezzard P, Bon-nerol V. Activation of human primary visual cortex during visual recall: a magnetic resonance imaging study. Proc Natl Acad Sci USA 1993; 90: 11802- 5. 112. Hirsch J, De La Paz, RL, Relkin NR, ct al. Illusory contours activate specific regions in human visual cortex: evidence from functional magnetic resonance imaging. Proc Natl Acad Sci USA 1995; 92: 6469- 73. 113. Lumer ED, Friston KJ, Rees G. Neural correlates of perceptual rivalry in the human brain. Science 1998; 280: 1930- 4. 1 14. D'Esposito M, Detrc JA, Aguirre GK, el al. A functional MRI study of mental image generation. Neuro- psychologia I997; 35: 725- 30. 115. Inclcfrcy P, Kleinschmidt A, Mcrboldl KD, ct al. Equivalent re­sponses to lexical and nonlexical visual stimuli in occipital cor­tex: a functional magnetic resonance imaging study. Neuroimage 1997; 5: 78- 81. 116. Prichard J, Rothman D, Novolny E, el al. Lactate rise delected by ' H NMR in human visual cortex during physiologic stimulation. Proc Natl Acad Sci USA 1991 ; 88: 5829- 31. 117. Sappcy- Marinicr D, Calabrcsc G, Fein G, Hugg . IW, Biggins C, Wciner MW. Effect of photic stimulation on human visual cortex lactate and phosphates using ' H and ! , P magnetic resonance spectroscopy. ./ Cereb Blood Flow Melab 1992; 12: 584- 92. 118. Merboldt KD, Kruger G, Hanicke W, Kleinschmidt A, Frahm J. Functional MRI of human brain activation combining high spatial and temporal resolution by a CINE FLASH technique. Magn Reson Med 1995; 34: 639- 44. I 19. Miki A, Nakajima T, Fujila M, ct al. Functional magnetic reso­nance imaging of the primary visual cortex: evaluation of human afferent visual system. Jpn ./ Ophthalmol 1995; 39: 302- 8. 120. Miki A, Nakajima T, Fujita M, Takagi M, Abe H. Functional magnetic resonance imaging in homonymous hemianopsia. Am .1 Ophthalmol 1996; 121: 258- 66. 121. Stoerig P, Kleinschmidt A, Frahm .1. No visual responses in de-nervaled VI: high- resolution functional magnetic resonance im­aging of a blindsighl patient. Neuroreporl 1998; 9: 21- 5. 122. Miki A, Nakajima T, Takagi M, Shirakashi M, Abe I- I. Detection of visual dysfunction in optic atrophy by functional magnetic resonance imaging during monocular visual stimulation. Am .1 Ophthalmol 1996; 122: 404- 15. 123. Miki A, Nakajima T, Hascbe 11, Abe H, Functional magnetic resonance imaging of visual function in post papilledema optic atrophy. J Neuro Ophthalmol 1997; 17: 223- 5. 124. McFadzean RM, Condon BC. Letter. ./ Neitro- ophlhalmol 1998; 18: 294- 5. 125. Demb JB, Boynton GM, Hccger DJ. Brain activity in visual cor­tex predicts individual differences in reading performance. Proc Natl Acad Sci USA 1997; 94: 13363- 6. 126. Bruhn H, Kleinschmidt A, Boecker H, Mcrboldl KD, Hanicke W, Frahm J. The effect of aeelazolamide on regional cerebral blood oxygenation at rest and under stimulation as assessed by MRI. ./ Cereb Blood Flow Metahol 1994; 14: 742- 8. 127. Breiter HC, Gollub RL, Weisskoff RM, el al. Acute effects of cocaine on human brain activity and emotion. Neuron 199749: 591- 611. 128. Gollub RL, Breiter FIC, Kantor H, el al. Cocaine decreases cor­tical cerebral blood flow but docs not obscure regional activation in functional magnetic resonance imaging in human subjects. ./ Cereb Blood Flow Melab 1998; I 8: 724- 34. 129. Kraut MA, Marcnco S, Softer BJ, Wong DF, Bryan RN. Com­parison of functional MR and H,-' sO positron emission tomog­raphy in stimulation of the primary visual cortex. AJNR 1995; 16: 2101- 7. 130. Frahm J, Merboldt KD, Hanicke W. Functional MRI of human brain activation at high spatial resolution. Magn Reson Med 1993; 29: 139- 44. 131. George JS, Sanders JA, Lcwine . ID, Caprihan A, Aine CJ. Com­parative studies of brain activation with MEG and functional MRI. In: Baumgartner C, et al, eds. Biomagnctism: Fundamental Research and Clinical Applications. Elsevier Science, IOS Press, 1995; 60- 5. 132. Aine CI, Supek S, George . IS, et al. Retinotopic organisation of ./ Neuro- OplUhalmol, Vol. 19. No. .!. I9W 200 R. M. McFADZEAN ET AL. human visual coitex: departures from the classical model. Cereb Cortex I996: 6; 354- 6I. 133. Talagala SL, Noll DC. Functional MRI using steady- state arterial water labeling. Magn Reson Med 1998; 39: 179- 83. 134. Schwarzbauer C, Heinke W. BASE imaging: a new spin labeling technique for measuring absolute perfusion changes. Magn Reson Med 1998; 39: 717- 22. 135. Gallcn C, Hirschoff E, Buchanan D. Magnetoenccphalography and magnetic source imaging: capabilities and Limitations. Neuro- imaging Clin North Am 1995; 5: 227^ 19. 136. Felleman DJ, Van Essen DC. Distributed hierarchical processing in the primate cerebral cortex. Cereb Cortex 1991; 1: 1- 47. 137. Engel SA, Rumelliart DE, Wandcll BA, et al. I'MRI of human visual cortex. Nature I994; 369: 525. 138. Van Essen DC, Maunsell JHR. Two- dimensional maps of the cerebral cortex. J Comp Neurol 1980; 191: 255- 81. ./ Neiim- Oplillldlinal, Vol. 19, No. .1 1999