Brain (1993), 116, 1293-1302

Conscious visual perception without VIJ. L. Barbur,1 J. D. G. Watson,2-5 R. S. J. Frackowiak3 and S. Zeki2

^Applied Vision Research Centre, Department of Optometry and Visual Science, City University,the department of Anatomy, University College and the 3MRC Cyclotron Unit,

Hammersmith Hospital, London UK

SUMMARY

We used the technique of PET to determine whether visual signals reach visual area V5,specialized for visual motion, when a human patient, blinded by a lesion in area VI,discriminates the direction of motion of visual stimuli and shows, through his verbal reports,that he is consciously aware of both the nature of the visual stimulus and its directionof motion. The results showed that area V5 was active without a parallel activation ofarea VI, implying that the visual input can reach V5 without passing first through VIand that such an input is sufficient for both the discrimination and the conscious awarenessof the visual stimulus.

INTRODUCTION

The visual cortex of the macaque monkey consists of multiple visual areas, many of whichreceive a direct anatomical input from the primary visual cortex, area VI (Zeki, 1978).In extending our work on the primate visual cortex, it seemed naturally interesting toenquire into a general principle about visual areas that has been formulated, namely thateach visual area contributes directly and explicitly to visual perception (Zeki, 1993). Suchan enquiry constitutes a daunting task, for it amounts to delving into problems of perceptionand of conscious knowledge of the visual world. It nevertheless seemed worth tryingto approach the problem in a simple form, by asking whether activity in one of thespecialized visual areas of the prestriate cortex can lead directly to the conscious perceptionof a visual stimulus without a parallel activation of area VI and whether this activitywould be sufficient to lead to conscious awareness of the stimulus attribute in question,even when the retinal signals are not subjected to the kind of processing associated withVI. An obvious way of testing this proposition would be to study visual capacities inconditions in which a given specialized visual area receives a visual input which bypassesarea VI and is therefore active while area VI is not. But such an enquiry has quite specificrequirements. First, we obviously had to restrict it to the human brain, since a correctverbal report from a human would be a quick way of establishing whether the subjectwas consciously aware, while discriminating the relevant feature of the visual stimulus;by contrast, it would take many months of tedious work to infer from the behaviour ofa monkey that it may have had conscious awareness, and such a conclusion would stillbe open to challenge. The human subject we sought was one blinded by a lesion to VI,to ensure that VI was not receiving any visual input and would not therefore be sendingCorrespondence to: Professor S. Zeki, Department of Anatomy, University College, London WC1E 6BT, UK, or ProfessorR. S. J. Frackowiak, MRC Cyclotron Unit, Hammersmith Hospital, 150 Ducane Road, London W12 OHS, UK.

© Oxford University Press 1993

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

1294 J. L. BARBUR AND OTHERS

any output to the prestriate visual areas or receiving a return input from them. Secondly,we needed a technique that would enable us to measure the activity in a given regionof the cerebral cortex; we opted for the technique of PET, which measures increasesin regional cerebral blood flow (rCBF) and thus gives an index of excess synaptic activitywithin an area. Finally, we needed to choose a well-characterized visual area outsideV1 to enable us to present to the human subject a visual stimulus tailored to the physiologyof the relevant area. We settled on area V5, which seemed admirably organized to answerour question. Specialized for visual motion in both monkey and man (Zeki, 1974; Zekiet al., 1991), the area is relatively accessible, being located laterally and ventrally inthe human brain (Watson et al., 1993). It receives a direct input from area VI (Cragg,1969; Zeki, 1969, 1971) but there are other retinal inputs to it, ones which bypass VI,at least in the monkey. One is a direct input from the lateral geniculate nucleus (Beneventoand Yoshida, 1981; Fries, 1981; Yukie and Iwai, 1981) and the other is an input via thesuperior colliculus and the pulvinar nucleus (Standage and Benevento, 1983). Moreover,it can be inferred from the studies of Beckers and Homberg (1992), using the methodof transcranial magnetic stimulation, that the non geniculo-striate visual input does, infact, reach area V5 before it reaches area VI. The latter input to V5 is no doubt thebasis for the observation that the physiological characteristics of monkey V5 are notabolished when it is deprived of an input from VI (Rodman et al., 1989; Girard et al.,1992). Instead, the directional selectivity that is so prominent a feature of V5 remains,although cells become much more broadly tuned and lose the crispness and selectivitiesthat are evident in a V5 receiving an input from VI. But such studies do not tell us whetherthe residual physiological activity in V5, after disconnecting it from VI, is sufficient tolead to the conscious perception of visual motion.

The characteristic physiology of area V5 is therefore not uniquely dependent upon aninput from VI. This made it interesting to ask whether the direct subcortical input toV5 could activate it sufficiently to mediate the conscious experience of motion withoutthe participation of VI, either because it is responsible for the preprocessing of signalsdestined for V5 or because the return input from V5 to it, which forms so prominenta feature of the anatomical connections of V5 (Shipp and Zeki, 1989), is critical for theconscious perception of visual motion. To answer our question, we investigated a patientwith a large VI lesion and with an intact V5, together with a residual visual capacityfor the detection of transient stimuli, as demonstrated in previous studies (Barbur et al.,1980; Blytheetal., 1987).

MATERIAL AND METHODS

We investigated two patients, both with extensive and unilateral lesions of VI. Subsequent studies showedthat in one, patient R.D., there was a large increase in rCBF in inferior left area VI when he undertookthe discrimination task; this suggested that, even in spite of his hemianopia, parts of his VI were activeduring the execution of the visual tasks. Indeed, subsequent psychophysical observations have demonstratedcomplete blindness in the right lower quadrant and residual vision in the right upper quadrant. This madehim an awkward subject to draw conclusions from. Another difficulty was that, in spite of the activity inVI, he had a highly abnormal conscious awareness of the visual stimulus, describing moving lights as 'ballsof fire' and was also poor in his discrimination performance. We decided therefore to eliminate him fromour present study and concentrate instead on the other patient, G.Y. For the stimulus conditions of ourexperiment G.Y. demonstrated clear, conscious awareness of motion and discriminated correctly the direction

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

CONSCIOUS VISUAL PERCEPTION WITHOUT VI 1283

of motion in virtually all presentations, as shown in Fig. 3. In this respect, our conditions of stimulationmust have been different from those employed in 'blindsight' studies, when the subjects tested could performvisual discrimination tasks without having any conscious awareness of the visual stimulus (Sanders et al.,1974; Weiskrantz, 1986). Blindsight patients can have some conscious awareness of some attributes of avisual stimulus (Weiskrantz, 1986) and we emphasize that, in the context of our PET study, the term blindsightdoes not describe G.Y.'s residual visual capacity correctly. Whether tested subjectively—through a verbalreport or objectively—through discrimination, the subject gave every sign of having seen and having beenconsciously aware of what he had seen. Thus, when stimulated with a moving stimulus, G. Y. verbally reportedseeing movement. Equally, in the experimental condition described below, he gave a verbal report of thedirection of motion. Because, even in spite of his 'blindness', he was able to discriminate correctly and faultlesslyand to have conscious awareness of having seen the particular visual stimulus and selected its directionalcharacteristics, G.Y. satisfied absolutely the requirements set out in the Introduction and the report hereis therefore restricted to him. The studies were approved by the Hammersmith Hospital Medical EthicsCommittee. Permission to administer radioactivity was obtained from the Administration of RadioactiveSubstances Advisory Committee of the Department of Health, UK. Informed consent was obtained fromthe patients.

G. Y., 36 years old, had sustained his injuries in a car accident at the age of 7 years. Magnetic resonanceimages (Fig. 1) show massive damage to the medial occipital lobe of the left hemisphere, sparing the occipital

FIG. 1. Magnetic resonance images of the brain of O.Y. snowing transverse (A) and sagittal (B) sections. The imageswere relatively T,-weighted and acquired at 1.3 mm voxel size. The transverse images are oriented parallel with theintercommissural plane, seen in the leftmost section and are 8 mm apart The images show the left side of the brain onthe left The rightmost sagittal image is a section located 4 mm to the left of the midJine, with the remain ing sectionslocated progressively more leftward in the brain, at 7 mm intervals. The damage is quite extensive, with nearly completedestruction of medial occipital cortex and the underlying optic radiation.

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

1296 J. L. BARBUR AND OTHERS

pole only, together with a total destruction of the optic radiation medially and less so laterally. The extentof his blind field was established with a novel perimetric method, implemented on the P-SCAN system whichalso allows the monitoring of eye fixation (Barbur, 1991). An 8 x 8 element checkerboard forms the visualstimulus and is generated at randomly selected locations on a uniform background field of luminance37 cd/m2. The check size expands with stimulus eccentricity and varies from 5' in the central foveal regionto — 1° at an eccentricity of 20°. With respect to the uniform background, the stimulus checks representboth increments and decrements in luminance. Each stimulus consists of 10 phase reversals of the checkelements at 5 Hz. If the subject detects the complete square pattern, the stimulus reduces to smallercheckerboards of 4 x 4 and finally 2 x 2 elements that probe the sensitivity of progressively smaller subregionsat the same location in the visual field. The stimulus covers every possible location in the visual field, butrequires significantly less time and yields very similar results to those obtained using detailed mapping ofthe visual field with single, small stimuli (Barbur et al., 1980). Our subject showed fixation stability betterthan 40' over several seconds. He has homonymous hemianopia with macular sparing (Fig. 2) but his visualacuity, colour discrimination and motion perception in his foveal and in his sighted hemifield are perfectlynormal, consistent with his neurological damage. In spite of this extensive damage, his residual vision enables '

FIG. 2. Plot of visual field sensitivity in subject G.Y. showing the clear separation of the sighted and the blind hemi-fidds (blind hemifield shown in grey, to the right) and the macular sparing, which extends —2.5° into the blindhemifield. His loss of visual sensitivity is uniform in his blind hemifield. In spite of what is clinically a completelyblind hemifield, G.Y. can detect and localize targets of high contrast when presented briefly if stationary, or whenmoving at high speed.

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

CONSCIOUS VISUAL PERCEPTION WITHOUT VI 1297

him to detect and localize rapid changes of retinal illuminance and fast-moving stimuli in his blind hemifield.We wanted to select two conditions, corresponding to a stationary and moving visual stimulus and to comparethe changes in rCBF in the two conditions. In essence, the paradigm we used was similar to the one in ourearlier studies in which we determined the position of human V5 by comparing the rCBF when subjectslooked at the same pattern in the stationary and moving modes (Zeki et al., 1991; Watson et al., 1993).The stimulus was a vertical bar subtending a visual angle of 1.5 x 11 °, and was either stationary or movingat randomly selected speeds; it was displayed on a high brightness Tektronix visual display unit in such away that only the blind hemifield was stimulated in its superior quadrant, 11° to the right of the fixationpoint and 11° above the horizontal meridian, die direction of the movement being either towards or awayfrom the vertical meridian. During the stationary condition, the stimulus was presented at a randomly selectedlocation, either to the left or the right of the visual field area covered by the moving stimulus. The displayunit operated at 120 non-interlaced frames per second and was therefore suitable for generating moving stimuli.Detection of scattered light in the sighted hemifield was eliminated by flooding the corresponding screenarea with uniform illumination. The background luminance of the Tektronix display was 34 cd/m2 and equalto that of the screen area corresponding to the sighted hemifield.

The subject reported 'near' or 'far' for the stationary stimulus and 'towards' or 'away' for the movingone. Residual vision could only be demonstrated when the visual stimulus changed rapidly in contrast ormoved at high speed. A two-alternative, forced-choice method was used to assess G. Y.'s ability to discriminatethe correct interval of stimulus presentation. Fixation devices were used to restrict head movement and highmagnification infrared images of the pupil allowed monitoring of the subject's fixation. No significant eyemovement was observed in response to stimulus presentation in any trial.

Figure 3 shows that the optimal condition for detecting the presence of a stationary visual stimulus wasa high contrast target which changed rapidly in luminance, while the optimal condition for detecting thepresence of motion was a high contrast, fast moving target {see shaded areas in Fig. 3). The shaded areasof Fig. 3 thus correspond to the two conditions determined as being optimal for comparison from ourpsychophysical studies. In one condition, corresponding to a moving stimulus, G.Y. detected the movingtarget and discriminated its direction of motion verbally in all presentations; in the other condition, correspondingto a stationary target, his performance was at chance. These were the very two conditions used in the PETstudy to reveal which areas of the brain were active when the subject detected the moving target.

Once the stimuli to be used in scanning die brain for changes in rCBF were determined, G.Y. was putin the PET machine where he viewed the identical Tektronix screen and was asked to discriminate the stimuli.The bore of the PET scanner, the ancillary equipment and the subject's torso were covered with black velvetto absorb unwanted light. The two stimuli were presented six times in alternation over 12 consecutive PETscans. Each PET scan lasted 3.25 min, with a 6.75 min interval to the next scan. During the scan an audiblesignal was used to indicate the end of each stimulus presentation and prompted the subject to report verballyeither the direction of motion (i.e. away or towards the point of regard), or position of the stationary stimulus(i.e. near or far). G.Y.'s correct responses, pooled over all the PET scans, were 99 and 49% for the movingand stationary stimuli, respectively (chance = 50%).

We measured increases in rCBF with a Siemens CTI 953B PET camera in its three-dimensional dataacquisition mode, three to five times more sensitive than conventional two-dimensional scanning (Bailey et al.,1991), thus allowing us to collect scans with sufficient statistical power for analysis of each subject separately.Relative rCBF was measured by recording the distribution of cerebral radioactivity caused by the freelydiffusible, positron-emitting, 13O-labelled tracer, water (H2

13O) infused intravenously, using techniquesdescribed before. Approximately 555 MBq or 15 mCi of H2

I5O was infused in each of 12 scans. Thedifference between the mean values of rCBF obtained for each of the conditions was evaluated for eachpixel by use of the t statistic. We used the method known as statistical parametric mapping (SPM; MRCCylcotron Unit, London, UK) for this purpose (Friston and Frackowiak, 1991). This generated a statisticalparametric map (SPM[;]) of the areas of significant rCBF change associated with die difference in the tasks.The most significant sites of change were determined and correlations with anatomical areas made by referenceto G.Y.'s MRI. We co-registered the PET image of G.Y.'s brain (Woods et al., 1992) with high resolution,volumetrically acquired T,-weighted MRIs of his brain because our previous studies have shown that thereis a consistent relationship between the position of area V5 and the sulcal pattern of the occipital lobe (Watsonetal., 1993).

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

1298 J. L. BARBUR AND OTHERS

FIG. 3. A, surface representation of the G. Y.'s correct discrimination scores for detection of the stationary bar stimulusplotted as a function of its luminance contrast and the temporal standard deviation of the Gaussian weighting function.The two variables investigated, namely contrast and temporal standard deviation of the stimulus, were chosen randomlyfrom preselected values and each combination was presented to the subject 36 times. A two-alternative, forced-choiceprocedure was used throughout: 50% correct response represents chance level, B, percentage scores for correct discriminationby G.Y. of the moving stimulus, again in a two-altemative, forced-choice procedure. The parameters investigated wereluminance contrast and the speed of movement, c, similar representation showing G.Y.'s ability to discriminate correctlythe direction of stimulus movement as a function of luminance contrast and speed. The hatched areas shown in A andB denote the parameters used for the stationary and the moving stimuli used in the PET experiment.

RESULTS

We directed our search of the SPM[f] principally to area V5 which is situated ventrallyin the occipital lobe, just posterior to the ascending limb of the inferior temporal sulcus.Our earlier studies had shown that, in addition to area V5, two other areas, putative V3and Brodmann's area 7 are also active when subjects view a moving pattern. By constrainingour search for activity to this part of the occipital lobe, we were able to threshold the

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

CONSCIOUS VISUAL PERCEPTION WITHOUT VI 1299

SPM at P < 0.05, corresponding to a Z score of 1.61. In order to report any areaselsewhere as significantly activated, a much harsher statistical threshold of P < 0.05,with a correction for multiple (voxel) comparisons was used (Z score of —3.7, comparedwith Z = 3.1 for P < 0.001 omnibus). In fact, the results showed that there was activitywithin V5, left putative V3 and area 7 (Fig. 4). In addition, there was significant activity

FIG. 4. Positron emission tomography data in which the MRIs and SPM[r] images have been coregistered and superimposed.On the left is a solid image of the brain at 70° rotation from the occipital pole. The arrows indicate the sulcal landmarksthat are associated with human area V5. The horizontal lines indicate the corresponding levels of the transverse sectionsshown on the right. Starting with the leftmost transverse section, the arrows indicate the activation of V5 (slice a), putativeV3 (slice b) and Brodmann's area 7 (slice c). The SPM[f] images share a common colour scale for their pixels' Z values,indicated on the right of the figure. The remaining areas that show activation, about which we had no prior hypothesis,do not have Z values necessary to achieve significance. In particular we could find no significant increase in the pulvinar;we believe that the rCBF changes induced in such nuclei are likely to be too small to be detected consistently by currentmethods.

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

1300 J. L. BARBUR AND OTHERS

in the superior vermis and an area in the right middle temporal lobe. We note in particularthat there was no activity in any region that could be said to be VI or its adjoining areaV2 and that, in this respect, the PET results confirmed the results obtained from MRIand also the clinical observations on the extent of his blind field.

DISCUSSION

Our results show directly that, during visual motion stimulation, signals reach the verycortical area which has been implicated in visual motion perception, without necessarilypassing through area VI and that, under these conditions, a patient may be consciouslyaware that the stimulus in his 'blind' field is a moving stimulus and report so verballyand also report verbally and correctly the direction of motion of the visual stimulus. Giventhe results from monkey experiments, it seems very likely that the hallmark of area V5,directional selectivity, is not conferred upon it by a cortical input from VI alone, butalso by a subcortical input since cells in V5 remain directionally selective, thoughabnormally so, even in the absence of a VI input (Rodman et al., 1989; Girard et al.,1992). Our results show that such subcortical input is sufficient not only for conferringupon the cells of V5 their characteristic physiology but also that V5, unaided by an inputfrom VI and without a reference back to VI in the well-documented projections linkingthe two areas (Shipp and Zeki, 1989), can contribute explicitly to conscious visualperception. The result thus gives credence to the general principle enunciated above, thateach area is capable of contributing directly and explicitly to visual perception in relationto its capacities (Zeki, 1993). It stands to reason that V5, deprived of its connectionswith VI, is not as efficient as normal. Hence the conscious awareness of visual motionin G.Y., though surprisingly good in the context of our experimental conditions, cannotbe expected to be as faultless as it would be in one with an intact VI in addition to anintact V5. It is nevertheless surprising how capable V5 is, on its own, in mediatingconscious visual perception.

It is of course possible to argue that some spared, but undetected, islands of VI allowedthe latter to make a direct contribution to the discrimination of the visual stimuli andto the conscious awareness. We discount this for two reasons: first both the MRI andthe PET scans did not show any vestige of VI, apart from the pole (which accounts forthe macular sparing) and, secondly, GY was completely hemianopic in the classical sense.We have also concentrated our explanation here on V5 and have done so principally becausewe tailored our visual stimulus to the physiology of V5, i.e. to directional motion selectivity.The stimulus we used and asked G. Y. to discriminate was a bar in motion, his task beingto discriminate the direction of motion. However, our PET results show that V5 wasnot the only visual cortical area to be active. There was also activity in putative areaV3 as well as in Brodmann's area 7. While we discount, at present, the latter area asbeing of pivotal importance in the discrimination we report here, it is very likely thatV3 may have been of some importance. Area V3, as defined in the monkey (Cragg,1969; Zeki, 1969), shares a common, M-dominated input with area V5 and receives itsVI input from the same layer as area V5, layer 4B (Lund et al., 1975; Zeki and Shipp,1988). More significantly, most cells in it are orientation selective and many of theseare, in addition, directionally selective (Zeki, 1978; Felleman and Van Essen, 1987).Thus the human equivalent of area V3 would stand a good chance of contributing to the

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

CONSCIOUS VISUAL PERCEPTION WITHOUT VI 1301

kind of visual discrimination task used here and, in the case of G.Y., is presumably alsoactivated by a direct route which bypasses area VI. If V3 did contribute to the consciousexperience of vision in our apparently blind subject, then this merely reinforces ourconclusion that activity in visual areas outside the prestriate cortex, without a parallelactivity in VI or indeed any activity in VI, is sufficient to lead to a correct discriminationof a visual stimulus and one, moreover, of which the subject is consciously aware evenif his awareness is not the same as that of normal individuals.

The role of VI in vision is better understood than perhaps any other visual area andyet many aspects of it remain mysterious and none more so than its role in consciousvision. Here we show for the first time that VI does not have a monopoly in mediatingconscious vision and that other areas, even when isolated from it, can have direct accessand contribute to that conscious experience.

ACKNOWLEDGEMENTS

We wish thank Dr E. Paulesu, Mr G. Lewington, Ms A Williams and Mr A. Blyth for their help in carryingout the PET scans, and Mr A. Harlow for his help with the psychophysical investigations. We are gratefulto the members of the Methods and Radiochemistry Sections at the MRC Cyclotron Unit for all their assistance.We thank Dr Stewart Shipp and Dr Georg Beckers for reading, and commenting on, the manuscript. Thiswork was supported by the Wellcome Trust.

REFERENCES

BAILEY DL, JONES T, SPINKS TJ (1991) A method for measuring the absolute sensitivity of positron emissiontomographic scanners. European Journal of Nuclear Medicine, 18, 374—379.

BARBUR JL (1991) The P-SCAN 100 system for simultaneous measurements of pupil size and eye movements.Journal of Psychophysiology, 5, 231-239.

BARBUR JL, RUDDOCK KH, WATERFIELD VA (1980) Human visual responses in the absence of the geniculo-calcarine projection. Brain, 103, 905-928.

BECKERS G, HOMBERG V (1992) Cerebral visual motion blindness: transitory akinetopsia induced by transcranialmagnetic stimulation of human area V5. Proceedings of the Royal Society of London, B, 249, 173 -178.

BENEVENTO LA, YOSHIDA K (1981) The afferent and efferent organization of the lateral geniculo-prestriatepathways in the macaque monkey. Journal of Comparative Neurology, 203, 455—474.

BLYTHE EM, KENNARD C, RUDDOCK KH (1987) Residual vision in patients with retrogeniculate lesionsof the visual pathways. Brain, 110, 887-905.

CRAGG BG (1969) The topography of the afferent projections in the circumstriate visual cortex of the monkeystudied by the Nauta method. Vision Research, 9, 733-747.

FELLEMANDJ.VAN ESSEN DC (1987) Receptive field properties of neurons in area V3 of macaque monkeyextrastriate cortex. Journal of Neurophysiology, 57, 889—920.

FRIES W (1981) The projection from the lateral geniculate nucleus to the prestriate cortex of the macaquemonkey. Proceedings of the Royal Society of London, B, 213, 73-80.

FRISTON KJ, FRACKOWIAK RSJ (1991) Imaging functional anatomy. In: Brain Work and Mental Activity.Edited by N. A. Lassen, D. H. Ingvar, M. E. Raichle and L. Friberg. Copenhagen: Munksgaard,pp. 267-279.

GIRARDP, S A U N P A , BULLIERJ (1992) Response selectivity of neurons in area MT of the macaque monkeyduring reversible inactivation of area VI. Journal of Neurophysiology, 67, 1437—1446.

LUND JS, LUND RD, HENDRICKSON AE, BUNT AH, FUCHS AF (1975) The origin of efferent pathwaysfrom the primary visual cortex, area 17, of the macaque monkey as shown by retrograde transport ofhorseradish peroxidase. Journal of Comparative Neurology, 164, 287—303.

RODMAN HR, GROSS CG, ALBRIGHT TD (1989) Afferent basis of visual response properties in area MTof the macaque. I. Effects of striate cortex removal. Journal of Neuroscience, 9, 2033-2050.

at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from

1302 J. L. BARBUR AND OTHERS

SANDERS MD, WARRINGTON EK, MARSHALL J, WEISKKANTZ L (1974) 'Blindsight': vision in a field defect.Lancet, i, 707-708.

SHUT S, ZEKI S (1989) The organization of connections between areas V5 and VI in macaque monkeyvisual cortex. European Journal of Neuroscience, 1, 309—332.

STANDAGE GP, BENEVENTO LA (1983) The organization of connections between the pulvinar and visualarea MT in the macaque monkey. Brain Research, Amsterdam, 262, 288—294.

WATSON JDG, MYERS R, FRACKOWIAX RSJ, HAJNAL JV, WOODS RP, MAZZIOTTA JC et al. (1993) AreaV5 of the human brain: evidence from a combined study using positron emission tomography and magneticresonance imaging. Cerebral Cortex, 3, 79-94.

WEISKRANTZ L (1986) Blindsight: A Case Study and Implications. Oxford: Clarendon Press.WOODS RP, CHERRY SR, MAZZIOTTA JC (1992) A rapid automated algorithm for aligning and reslicing

PET images. Journal of Computer Assisted Tomography, 16, 620-633.Y U H E M, IWAI E (1981) Direct projection from the dorsal lateral geniculate nucleus to the prestriate cortex

in macaque monkeys. Journal of Comparative Neurology, 201, 81—97.Z E H SM (1969) Representation of central visual fields in prestriate cortex of monkey. Brain Research,

Amsterdam, 14, 271-291.Z E H SM (1971) Convergent input from the striate cortex (area 17) to the cortex of the superior temporal

sulcus in the rhesus monkey. Brain Research, Amsterdam, 28, 338—340.ZEKI SM (1974) Functional organization of a visual area in the posterior bank of the superior temporal sulcus

of the rhesus monkey. Journal of Physiology, London, 236, 549—573.ZEKI SM (1978) Functional specialisation in the visual cortex of the rhesus monkey. Nature, London, 274,

423-428.ZEKI S (1993) A Vision of the Brain. Oxford: Blackwell Scientific.ZEKI S, SHIPP S (1988) The functional logic of cortical connections. Nature, London, 335, 311—317.ZEKI S, WATSON JDG, LUECK CJ, FRISTON KJ, KENNARD C, FRACKOWIAK RSJ (1991) A direct

demonstration of functional specialization in human visual cortex. Journal of Neuroscience, 11, 641 —649.

Received September 14, 1993. Revised October 14, 1993. Accepted October 14, 1993 at University C

ollege London on D

ecember 23, 2011

http://brain.oxfordjournals.org/D

ownloaded from