Acknowledgements We wish to thank the many collaborators, postdoctoral fellows and studentsinvolved in the Ts16pro]ect.
4 Rossor, M. and Iversen, L. (1986) Br. Med. Bull. 42, 70-74 5 Katzman, R. (1986) N. Engl. Med. J. 314, 964-973 6 Wisniewski, K. E., Wisniewski, H. M. and Wen, G. Y. (1985) Ann. Neurol. 17, 278-282 7 Glenner, G. G and Wong, C. W. (1984) Biochem. Biophys. Res. Commun. 120, 885-890 8 Kang, J. etaL (1979) Nature 325, 733-736 9 Robakis, N. K., Ramakrishna, N., Wolfe, G. and Wisniewski, H. M. (1987) Proc. Natl Acad. Sci. USA 84, 4190-4194 10 Goldgaber, D., Lerman, M. I., McBride, W. O., Saffiotti, U. and Gajdusek, D. C. (1987) Science 235, 887-880 11 St George-Hyslop, P. H. et aL (1987) Science 235, 885-890 12 Breitner, J. C., Folstein, M.F. and Murphy, E.A. (1986) J. Psychiatr. Res. 20, 31-43 13 Hook, E. B. (1981) in Trisomy 21 (Down Syndrome) (de la Cruz, F. F. and Geralk, P. S., eds), pp. 3-68, University Park Press 14 Epstein, C. J. (1986) The Consequences of Chromosomal Imbalance, Cambridge University Press 15 Coyle, J. T., Oster-Granite, M. L. and Gearhart, J. D. (1986) Brain Res. Bull. 16, 773-787 16 Jervis, G. A. (1948) Am. J. Psychiatr. 105, 102-106 17 Ball, M. J. and Nuttal, K. (1981) J. Neuropathol. Appl. Neurobiol. 7, 13-30 18 Burger, P. C. and Vogel, F. S. (1973) Am. J. Pathol. 73,457476 19 Yates, C. M. etal. (1983) Brain Res. 280, 119-126 20 Yates, C. M. etal. (1983) Brain Res. 258, 45-52 21 Mann, D. M. A., Yates, P.O. and Marcynicek, B. (1984) J. Neuropathol. Appl. Neurobiol. 10, 185-207 22 Casanova, M. F., Walker, L. C., Whitehouse, P. J. and Price, D. (1985) Ann. Neurol. 18, 310-313 23 Kirkpatrick, J. B. and Hicks, P. (1984) J. Neuropathol. Exp. Neurol. 43, 307
24 Ropper, A. H. and Williams, R. S. (1980) Neurology 30, 639644 25 Schwarz, M. et al. (1983) Science 221,781-783 26 Shapiro, M. B., Haxby, J. V., Grady, G. L. and Rapoport, S. I. (1986) in The Neurobiology of Down Syndrome (Epstein, D. J., ed.), pp. 89-108, Raven Press 27 Gropp, A., Kolbus, U. and Giers, D. (1975) Cytogenet. Cell Genet. 14, 42-62 28 Gearhart, J., Davisson, M. T. and Oster-Granite, M. L. (1986) Brain Res. Bull. 16, 789-801 29 Reeves, R. H., Gearhart, J. D. and Littlefield, J.W. (1986) Brain Res. Bull. 16, 803-814 30 Reeves, R. H., Callahan, D., O'Hara, B. F., Callahan, R. and Gearhart, J. D. (1987) Cytogenet. Cell Genet. 44, 76-81 31 Reeves, R. H. etal. (1987) Mol. Brain Res. 2, 215-221 32 Watson, D. K. et al. (1986) Proc. Natl Acad. Sci. USA 83, 1792-1796 33 Skow, L. C. and Donner, M. E. (1985) Genetics 110, 723732 34 Oster-Granite, M.L., Gearhart, J.D. and Reeves, R.H. (1986) in The Neurobiology of Down Syndrome (Epstein, D. J., ed.), pp. 137-151, Raven Press 35 Oster-Granite, M.L. et al. (1987) in Animal Model of Dementia (Coyle, J. T., ed.), pp. 279-307, Alan Liss 36 Oster-Granite, M. L. et al. (1987) Soc. Neurosci. Abstr. 13, 1121 37 Singer, H. S., Tiemeyer, M., Hedreen, J. C., Gearhart, J. and Coyle, J. T. (1984) Dev. Brain Res. 15, 155-166 38 Ozand, P. T. etal. (1984) J. Neurochem. 43,401-408 39 Saltarelli, M. D., Forloni, G-L., Oster-Granite, M. L., Gearhart, J. D. and Coyle, J. T. (1987) Dev. Genet. 8, 267-279 40 Lovett, M. et al. (1987) Biochem. Biophys. Res. Commun. 144, 1069-1075 41 Bendotti, C. etal. (1988) Proc. NatlAcad. Sci. USA 85, 36283632 42 Gearhart, J. D. et al. (1986) Brain Res. Bull. 16, 815-824
Probing visual cortical function with discrete chemical lesions William
T. N e w s o m e
a n d R o b e r t H. W u r t z
Recent anatomical and physiological experiments sugInitial glimpses into the functional roles played by gest that a neural pathway in primate visual cortex several of these visual areas emerged from single-unit selectively analyses visual motion information. By recordings in anesthetized monkeys. Zeki and cocreating small chemical lesions in identified visual workers reported that a striate-recipient zone of the areas of this pathway, a new link has been established superior temporal sulcus contained a preponderance between the physiological properties of cortical neurons of direction-selective neurons a, whereas other and the behavioral capabilities of rhesus monkeys. Such extrastriate areas appeared relatively enriched in lesions elevate psychophysical thresholds in motion- color-selective neurons 4'5. Zeki suggested that the Sensorimotor related tasks while leaving non-motion thresholds extrastriate visual areas function in parallel, with each Research,National unaffected. Small chemical lesions also impair a area being specialized for the analysis of a particular EyeInstitute, monkey's ability to employ motion information to guide aspect of the visual image such as motion, color, Bethesda,MD 20892. eye movements, but have no effect on eye movements to disparity, or orientation 6. More recently, the complex USA. static targets. This technique creates the opportunity for serial as well as parallel nature of extrastriate a detailed functional analysis of the motion pathway organization has been emphasized by the notion of and may be employed in other visual pathways as well. functionally specialized 'pathways' or 'streams of processing' in visual cortex 7'8. In this view, parallel One of the most intriguing developments in the study extrastriate pathways indeed process distinct types of of the mammalian visual system has been the discov- visual information, but serial principles are incorporery of multiple cortical areas that perform an exten- ated within each pathway as indicated by extensive sive analysis of the visual image beyond that carried feedforward and feedback connections within a pathout by the primary visual area (striate cortex, or V1). way and by progressively more complex response In the macaque monkey, investigators from a number properties at higher levels of a pathway (reviewed in of laboratories have identified more than 20 'extrastri- Ref. 2). ate' visual areas in the occipital, temporal and The most intensively studied pathway in primate parietal lobes. Many of these areas are illustrated in extrastfiate cortex appears to be devoted to the Fig. 1, and together they comprise roughly half of the analysis of visual motion. This pathway is characterneocortex of the m a c a q u e1 2' . A major task now ized at each stage by neurons that respond selectively confronting investigators in this field is to understand to the direction of motion of a visual stimulus while how these extrastriate visual areas contribute, indi- being relatively unselective for other aspects of the vidually and collectively, to visual perception and stimulus. The motion pathway originates in layer 4B of striate cortex, or V1, and may include the 'thick' visually guided behavior. William T. Newsome is at the Department of Neurobiology, StanfordUniversity Schoolof Medicine, Stanford, CA 94305, USAand RobertH. Wurtz is at the Laboratoryfor
© 1988.ElsevierPublications.Cambridge 0378 5912188150200
TINS, Vol. 1 I, No. 9, 1988
cytochrome oxidase stripes of area V2. These anatomical subdivisions of V1 and V2 project to area MT which appears to be the first stage of the pathway where an entire cortical area is devoted to the analysis of motion information. MT in turn transmits this information to visual areas of the parietal lobe including MST and VIP (see Fig. 1). Identification of this pathway as a 'motion analysing' system is strengthened by several observations that neurons at successive levels of the pathway encode progressively more complex information concerning motion in the visual environment (see Ref. 2 for a review of the anatomy and physiology of this pathway). Interestingly, recent neurological observations suggest that a selective pathway for analysing visual motion exists in the human visual system as well9. Though the physiological and anatomical evidence for an extrastriate motion pathway is impressive, such data cannot demonstrate a firm link between neural activity in the pathway and the animal's perceptual or behavioral responses to motion. Lesion experiments hold promise for demonstrating such a linkage, but lesions of this pathway are complicated by the relatively inaccessible location (buried in deep sulci) of several of the relevant visual areas. However, in recent years, investigators in several laboratories have successfully carried out such experiments by injecting small volumes of a neurotoxin, ibotenic acid, into area MT. In these experiments, the representation of the visual field in MT is first mapped with microelectrodes, and the neurotoxin (usually 1-4 ~tl)is then injected into a selected region of MT. Since the toxin selectively kills cell bodies, lesions can be made without damage to the underlying white matter 1°. Using this technique, small cortical lesions can be placed at known topographic locations within an identified visual area. Figure 2 shows an MT lesion created in this manner.
Motion perception In one class of experiments conducted using this technique, investigators have measured the effects of MT lesions on perception of visual motionn-13. The basic strategy of these experiments is to determine the effects of MT lesions on psychophysical thresholds for motion-related perceptual tasks and for non-motion-related tasks. If MT lesions elevate motion thresholds relative to non-motion thresholds, one may infer that a link exists between neural activity in MT and motion perception. In one study, Newsome and Par612 trained monkeys to discriminate the net direction of motion in a dynamic random dot display. The visual stimulus used in this study is illustrated in Fig. 3A. In the display shown in the left-hand panel, dots are flashed briefly and in rapid succession at random positions on a CRT screen. In this display, referred to as the 'uncorrelated' or '0% correlation' state, there are many localmotion events due to random associations of the briefly flashed dots, but there is no net motion in any single direction. In the display depicted in the righthand panel, an initial sequence of random dots is flashed on the CRT, but as each dot disappears, it is replaced by a partner dot with a uniform offset in space (6x) and time (6t). Thus one perceives a random dot pattern in which the motion of each dot is identical to the net motion of the entire pattern. This TINS, Vol. 11, No. 9, 1988
Fig. 1. The location of visual areas in macaque cerebral cortex. Upper diagram. The superior temporal sulcus has been unfolded so that visual areas normally hidden from view can be seen. Lower diagram. The lunate, inferior occipital and parieto-occipital sulci have been partially unfolded. The thin lines indicate the boundaries of visual areas that are well-established; the dashed fines mark borders that are less well defined. The primary visual area, V1 or striate cortex, can be seen at the occipital pole of the hemisphere, and many extrastriate areas are also visible. The extrastriate areas are identified on the basis of several criteria including visual topography, anatomical connections, architectonics and neuronal response properties. MT and MST, important extrastriate components of the motion pathway, are located within the superior temporal sulcus as seen in the upper diagram. Abbreviations: AIT, anterior inferotemporal; DP, dorsal prelunate; MT, middle temporal; MST, medial superior temporal; PIT, posterior inferotemporal; PO, parieto-occipital; STP, superior temporal polysensory; VA, ventral anterior; VP, ventral posterior. (Reproduced, with permission, from Ref. 2.)
display is referred to as 100% correlated motion. In practice, the monkey generally viewed an intermediate form of the display such as that shown in the center panel. In this display, 50% of the dots move in a correlated fashion, but this net motion signal is embedded in random motion noise provided by the 50% of the dots in uncorrelated motion. For each trial, the experimenter could specify the percentage of dots in correlated motion as well as the spafiotemporal composition of the correlated motion events (6x and 6t). 395
Each monkey also performed an orientation discrimination as a non-motion control task. In this task the monkey reported the orientation of a stationary sine wave grating presented within the stimulus aperture. The contrast of the grating was varied randomly between trials until a threshold contrast value could be determined. The right-hand panel in Fig. 3B illustrates pre- and post-lesion contrast thresholds measured in the test hemifield on the same days as the motion thresholds in the left-hand and center panels. Clearly the lesion had little, if any, effect on contrast sensitivity. In a related set of experiments, Siegel and Andersen sa have reported similar results. They trained monkeys to detect the onset of shearing motion in a static random dot pattern and to detect the presence of 3-D structure in dynamic random dot patterns. They found that small chemical lesions of MT impaired performance on both of these tasks while leaving contrast sensitivity unaffected. Together, these results demonstrate that MT lesions can selectively impair motion perception. The experiments thus establish a link between the physiological properties of motion pathway neurons and the animal's perceptual performance.
Fig, 2. A Nisslstainedsection showing an ibotenic acid lesion of extrastriate area MT. Thesection shows a parasagittal cut through the superior ternpora/ su/cus. MT is located deep in the sulcus on the posterior bank. The lesion (arrows) appearsas a region of dramatic cell loss with consequent disruption of the normal cortical pattern of lamination. Posterioris to the left in the micrograph, and dorsal is upward.
In a typical threshold measurement, the monkey indicates on each trial whether the net motion in the random dot display is upward or downward. The monkey is rewarded with a drop of water for identifying the direction of motion correctly. The minimum correlation value for which the monkey could successfully execute the discrimination was designated as threshold 12. During threshold measurements, the monkey viewed the dot pattern perifoveally while maintaining fixation on a separate point of light. In this manner, thresholds could be measured independently for dot patterns placed in either visual hemifield. Since MT is a bilateral structure that contains a topographic representation of the contralateral half of visual space, the hemifield ipsilateral to the MT lesion could be used to measure control thresholds while performance in the contralateral hemifield revealed the effects of a unilateral MT lesion. The left-hand panel of Fig. 3B depicts pre-lesion and post-lesion motion thresholds measured in the hemifield contralateral to the MT lesion ('test' hemifield). Thresholds are shown for five different spatial intervals corresponding to a range of speeds. Before the MT lesion, thresholds were as low as 2% for the optimum spatial intervals - impressively low values that compare favorably with the best performance of human observers under similar conditions. However, on the day following the lesion, thresholds were elevated by 500-800%, demonstrating a gross impairment of the animal's ability to discriminate the direction of motion. In contrast, the center panel of Fig. 3B shows that motion thresholds measured simultaneously in the ipsilateral ('control') hemifield were entirely normal. This observation indicates that the impairment revealed in the test hemitield did not result from general fluctuations in the animal's attentional or motivational state. 396
Smooth pursuit eye movements Visual motion information can be used for the guidance of movement as well as for perception. For example, smooth pursuit eye movements are employed by primates to match movement of the eyes to the motion of a target in the visual environment. Clearly, performance of this task depends upon accurate visual information about the motion of potential targets, and the chemical lesion technique described above can be used to determine whether the cortical motion pathway contributes such information to the pursuit system. With this goal in mind, Newsome et al. 14.15 trained monkeys to pursue horizontally moving targets in order to obtain a liquid reward. Each trial began with the monkey fixating a stationary visual target on a tangent screen. After a variable fixation period, the target disappeared and subsequently reappeared in motion at a randomly chosen location on the horizontal meridian. The monkey's task was to move its eyes to the target and then track it with smooth pursuit eye movements for the duration of the trial. Figure 4A depicts a monkey's eye movements in response to such a stimulus. The upper traces depict ten superimposed responses (solid lines) obtained in a normal monkey. Following the fixation period, the target (dashed line) 'stepped' 5 deg. to the monkey's right and then moved rightward at 16 deg./s toward the edge of the screen. Note that the monkey's response had two components: a rapid, or saccadic eye movement that brought the fovea near the target, and then a smooth pursuit eye movement that kept the fovea on the target for the duration of the trial. The initial 80 ms interval of pursuit following the saccade is of particular interest. Since the pursuit response has a latency of at least 80 ms, this initial interval of pursuit must be programmed using visual motion information obtained before the saccade occurred, that is, while the eyes were still pointed toward the original fixation position. In Fig. 4A, for example, this initial interval of pursuit depended upon motion TINS, Vol. 11, No. 9, 1988
information obtained on the horizontal meridian, system while leaving intact other pathways that 5-7 deg. eccentric in the right visual hemifield. By supply information concerning the static position of varying the size and direction of the initial target visual targets. 'step', the critical interval of target motion could be placed at any desired location in the visual field. The A directional pursuit deficit The visually derived pursuit deficit described above efficacy of the target motion for pursuit initiation could then be assessed by measuring the speed of the consistently occurred after lesions in peripheral porpursuit movement during the initial 80 ms interval tions of MT where neuronal receptive fields did not include the fovea. A qualitatively different pursuit following the saccade (see also Ref. 16). The lower traces in Fig. 4A show the responses of deficit resulted from chemical lesions of the foveal the same monkey to this stimulus on the day after an representation of MT or of an adjacent visual area, injection of ibotenic acid into MT of the contralateral MST, in which many receptive fields are directionally (left) hemisphere. The most striking deficit is that the selective and include the fovea17'1a. This deficit, speed of the smooth pursuit eye movement was illustrated in Fig. 5, differed from the previously reduced during the initial interval of pursuit. This can described initiation deficit in two important respects. be seen from the reduced slope of the post-lesion eye First, pursuit speed was reduced for the entire movement traces in Fig. 4A, and is even more evident duration of the trial as opposed to an initial interval of in the averaged eye-speed traces in Fig. 4B. After approximately A No Correlation 50% Correlation 200 ms of impaired pursuit, the monkey made a corrective saccade and tracked the moving target accurately for the duration of the trial. The ability of the monkey to pursue accurately near the end of the trial suggests that the deficit in pursuit initiation was visual rather than motor in origin. The monkey was able to correct the initially faulty pursuit movement because B oo Lesion Hemifield oo_ C o n t r o l Hemifield _ Lesion Hernifield the large saccadic eye movement placed the target image in a portion P / of the visual field that was undam/ t aged by the small MT lesion. In / i.. / fact, pursuit initiation was impaired /, only for target motion in a small 8 8 ~° region of the visual field that corresponded topographically to the ,3 j/ location of the lesion in the contralateral MT. Pursuit initiation was normal at all other locations in the test as well as control hemifields. ml ........ , ........ , In addition to this pursuit deficit, 001 0.1 1.0 0.01 0.I 1.0 0.I 1.0 I00 Spatial Interval (deg) Spatial Interval (deg) Spatial Frequency (cyc/deg) the monkeys were also impaired in their ability to adjust the amplitude of their saccades to compensate Fig. 3. (A) A schematic representation of the random dot stimulus used to measure motion for target motion. In Fig. 4, one discrimination thresholds. Dots are plotted individually and in rapid succession at random locations can see that the amplitude of the on a CRT screen. Each dot lasts for a brief interval after which it disappears and is replaced by initial saccade to the moving target another randomly placed dot. The operator may specify that a certain percentage of the dots be was consistently reduced following replotted with a fixed spatial and temporal offset from their partner dots. This subset of dots the MT lesion 14'15. Both the pur- provides a net motion signal that is embedded within a masking motion noise. The panel on the left depicts a condition in which all dots are plotted randomly (no correlation, or 0% correlation). There suit deficit and the saccade deficit is no net motion signal although there are many local motion events resulting from fortuitous suggest that the monkey under- associations among the stream of randomly plotted dots (arrows). In the panel on the right each estimated the speed of the moving dot is replotted with a fixed offset so that the motion of each dot is identical to that of the entire target following the MT lesion. In pattern (100% correlation). The center panel depicts an intermediate display in which 50% of the contrast, the monkeys made accu- dots constitute a 'correlated' motion signal while the other 50% of the dots comprise a masking rate saccades to stationary targets motion noise (50% correlation). Psychophysical thresholds were determined as the minimum at all locations in the visual field, correlation value for which the monkey could successfully discriminate upward from downward including those for which the pur- motion. (B) The psychophysical effects of an ibotenic acid lesion of MT. In each panel, the solid line suit response to moving targets and so bars represent the mean pre-lesion threshold and SD for each condition tested. The dashed lines indicate post-lesion thresholds obtained 24 h after the M T injection. The panel on the left was impaired. As with the percep- shows motion thresholds for five different spatial intervals in the test (contralateral) hemifield (the tual data described above, the eye temporal interval was held constant at 45 ms). The injection caused threshold increases of 400movement data indicate that 800% in this hemifield. The center panel demonstrates that motion thresholds in the control lesions of MT selectively damage a (ipsilateral) hemifield were unaffected by the lesion. The panel on the right illustrates contrast visual pathway that provides thresholds for three different spatial frequencies in the test hemifield. The injection caused little, if motion inputs to the oculomotor any, elevation of contrast thresholds. (Adapted from Ref. 12.)
TINS, Vol. 11, No. 9, 1988
200 ms. This chronic reduction in pursuit soeed. dearly observable in the averaged eye-speed traces in Fig. 5B, made it necessary for the monkey to execute numerous 'catch-up' saccades during sustained intervals of smooth pursuit (Fig. 5A, lower traces). Secondly, this deficit was present for any pursuit eye
z 0- - O . I--
movement toward the hemisphere that sustained the lesion, regardless of the point of origin of target motion. Pursuit in the opposite direction was unaffected. The known physiological properties of neurons in foveal MT and MST do not provide a clear explanation for this directional deficit. Many neurons in these areas discharge in relation to smooth pursuit eye movements, but there is no striking asymmetry in the directional preferences of these neurons that could account for the directional pursuit deficit 19-~2. However, an eventual understanding of the mechanisms underlying this deficit is of particular interest, because the deficit is virtually identical to a pursuit deficit observed in humans following unilateral lesions of parieto-occipital cortex23. Comparative investigations of these deficits may permit identification of motion-related visual areas in the cortex of humans that are homologous to those being studied in monkey cortex.
0 0LU >LI.I ¢~
Recovery of function Recovery of function is typically observed following each of the deficits we have described. With daily practice, recovery from the effects of small, subtotal lesions of MT is frequently complete within a week O, and almost always complete within two weeks. Longer-lasting17 or even permanent 12 deficits are ? evident following larger lesions, but the permanent ~'":" '200 '400 '600 '800 -200' deficit is small relative to the acute deficit that TIME (ms) immediately follows the lesion. The mechanisms that mediate the recovery of function are a matter of active investigation 24. Plastic changes in topography or receptive field size within MT may play a role in the recovery from subtotal lesions. However, the substantial recovery observed after a complete unilateral lesion of MT indicates that other visual areas are capable of assuming the functional roles formerly : • .,~',,',- :~ ~,,;~ji.,~.~-,," played by MT neurons. w This fending is not surprising in light of the u.i i °t~ IIi U II~ i V anatomical complexity that characterizes even a single if) cortical pathway such as that for visual motion. For 1 . 1 . 1 >example, higher level areas of the motion pathway, U.I O, such as MST, receive a varied set of afferent inputs in addition to those from MT 25. Behavioral recovery from the effects of MT lesions may reflect synaptic o changes within higher level areas that permit greater I reliance on alternative sources of afferent input. From -~bo 'o '2oh '4o6 '6ob '8ob this point of view, the complex web of anatomical TIME (ms) M 1 connections between extrastriate visual areas can provide resistance to the deleterious effects of damage to any single area. Fig. 4. Effect of a chemical lesion of MT on the initiation of smooth pursuit eye An important issue in evaluating the results we movements. (A) Pursuit eye movements before (upper traces) and 24 h after have reviewed is whether the conclusions are com(lower traces) a 1 I~1injection of ibotenic acid into/tAT. The schematic drawing in the upper left-hand corner indicates that the lesion was made in the left promised by the transient nature of the deficits. We cerebral hemisphere and that motion of the pursuit target in this example was have argued that functional recovery is to be expected to the right. The dashed line represents target motion: following the fixation given the complexity of the motion pathway and the period, the target stepped 5 deg. to the right and moved smoothly to the right restricted extent of the lesions. The significance of at 16 deg./s. The solid lines depict ten superimposed eye movement responses the deficits rests not on their longevity, but on their before and after lesion. (B) Mean speed of the pursuit eye movements shown specificity: the lesions impair performance of several in A. The dotted fine shows target speed, the solid lines depict the speed of the motion-related tasks whereas performance of nonpre-lesion pursuit movements (mean and SE), and the dashed fines indicate the motion tasks is unaffected. The data indicate convinspeed of the post-lesion pursuit movements (mean and SE). The interruption of cingly that neuronal activity in MT plays a major role the eye speed traces resulted from excision of the saccade during data analysis. The primary effect of the lesion on pursuit eye movements was a reduction in in motion processing in the normally functioning eye speed during the initial 200 ms of the pursuit response. The deficit was cerebral hemisphere. It is instructive in this respect to consider another evident only when target motion originated in a portion of the visual field that corresponded topographically to the location of the lesion in/tAT. (Modified case of recovery of function, that of saccadic eye from Refs 14 and 15.) movements following a lesion of the superior collicu'
TINS, Vol. 11, No. 9, 1988
lus. The physiology of collicular neurons suggests a prominent role for this structure in the generation of saccades. On the day after an electrolytic lesion of the colliculus, however, monkeys have nearly normal Lt3 saccadic amplitudes and lengthened saccadic latencies that recover within weeks 26. Schiller and co0 workers27 have shown that simultaneous lesions of the superior colliculus and frontal eye fields permanently reduce the amplitude of visually guided saccades while lesions of either structure alone result in z ~ . only transient deficits. Clearly, these structures act in o parallel in mediating saccadic eye movements, and it ~o. would be a mistake to conclude from the transient nature of the collicular deficits that the superior colliculus is not involved in the generation of saccades. LLI Indeed, major deficits have now been observed immediately following chemical lesions of the colliculus before compensatory mechanisms can be actio vated28. Therefore, in the analysis of complex neural systems, it can be important to assess the magnitude Lt3 I, and selectivity of behavioral deficits immediately after -250" perturbations of individual structures. The chemical lesion technique appears well suited for this task since it is minimally invasive: the lesions can be created in alert animals and behavioral testing can be resumed B without delay. Concluding r e m a r k s Our knowledge of visual cortical organization has greatly increased in recent decades. The newly discovered extrastriate visual areas are sufficiently formidable in number and complexity to raise doubts of our ultimate ability to understand their functional roles in vision. However, current physiological data are encouraging in that the extrastriate areas appear to be organized into a small number of pathways that process relatively distinct aspects of the visual image. We need experimental approaches by which the physiological properties of neurons in these pathways can be linked to the perceptual and behavioral capabilities of the animal. Discrete chemical lesions placed in identified visual areas have now provided one such link for the motion pathway. The technique can be further exploited to ask more refined questions concerning the perceptual and behavioral contributions of neuronal processing in this pathway. We are optimistic that this approach can also be used for a behavioral analysis of function in other extrastriate pathways.
7 - 200
'400 TIME (ms)
Fig. 5. A directional pursuit deficit that followed injection of ibotenic acid at the
border of foveal M T and MST. The lesion was in the right hemisphere, as indicated in the schematic drawing in the upper left-hand corner. The target stepped 5deg. to the leftand then moved to the right at 16 deg./s. (A) Pursuit eye movements made before (upper trace) and 24 h after (lower trace) the /V1T-/VlSTlesion. The dashed line represents target motion, and the solid lines depict ten superimposed pursuit responses for each case. (B) Mean speed of the pursuit eye movements shown in A. The dotted line shows target speed, Selected references the solid lines depict the speed of the pre-lesion movement (mean and SE), and 1 Van Essen, D. C. (1985) in CerebralCortex (Vol. 5) (Peters, A. the dashed lines indicate the speed of the post-lesion pursuit movements and Jones, E. G., eds), pp. 259-329, Plenum Press (speed and SE). Unlike the deficit illustrated in Fig. 4, pursuit speed was reduced 2 Maunsell, J. H. R. and Newsome, W. T. (1987) Annu. Rev. throughout the trial when the lesion involved MST or the foveal region of M T. Neurosci. 10, 363-401 This deficit was evident for all pursuit movements toward the hemisphere 3 Dubner, R. and Zeki, S. M. (1971) Brain Res. 35, 528-532 containing the lesion (to the right in this example) regardless of the point of 4 Zeki, S. M. (1973) Brain Res. 53,422-427 origin of the target motion. (Modified from Ref. 17.) 5 Zeki, S. M. (1977) Proc. R. Soc. Lond. Ser. B 197, 195-223 6 Zeki, S. M. (1978) J. Physiol. (London) 277, 273-290 7 Ungedeider, L. G. and Mishkin, M. (1982) in Analysis of 12 Newsome, W. T. and ParE, E. B. (1988) J. Neurosci. 8, 22012211 Visual Behavior (Ingle, D. J., Goodale, M. A. and Mansfield, 13 Siegel, R. M. and Andersen, R.A. (1986) Soc. Neurosci. J. W., eds), pp. 549-580, MIT Press Abstr. 12, 1183 8 Van Essen, D. C. and Maunsell, J. H. R. (1983) Trends 14 Newsome, W. T., Wurtz, R. H., Diirsteler, M. R. and Mikami, Neurosci. 6, 370-375 A. (1985) J. Neurosci. 5, 825-840 9 Zihl, J., von Cramon, D. and Mai, N. (1983) Brain 106, 313340 10 Olney, J. W. (1983) in Excitotoxins (F0xe, K., Roberts, P. and Schwartz, R., eds), pp. 82-95, Macmillan Press 11 Newsome, W. T. and ParE, E. B. (1986) Soc. Neurosci.Abstr. 12, 1183
TINS, Vol. 11, No. 9, 1988
15 Newsome, W. T., Di~rsteler, M. R. and Wurtz, R. H. (1986) in Adaptive Processesin Visualand Oculomotor Systems(Keller, E. L. and Zee, D. S., eds), pp. 223-230, Pergamon Press 16 Lisberger, S. G. and Westbrook, L. E. (1985) J. Neurosci. 5, 1662-1673
17 DOrsteler, M. R., Wurtz, R. H. and Newsome, W. T. (1987) J. NeurophysioL 57, 1262-1287 18 Dirsteler, M. R. and Wurtz, R. H. J. NeurophysioL (in press) 19 Sakata, H., Shibutani, H. and Kawano, K. (1983) J. Neurophysiol. 49, 1364-1380 20 Kawano, K., Sasaki, M. and Yamashita, M. (1984) J. NeurophysioL 51,340-351 21 Komatsu, H. and Wurtz, R. H. J. Neurophysiol. (in press) 22 Newsome, W. T., Wurtz, R. H. and Komatsu, H. J. NeurophysioL (in press) 23 Troost, B. T. and Abel, L.A. (1982) in Functional Basis of Ocular Motility Disorders (Lennerstrand, G., Zee, D. S. and Keller, E. L., eds), pp. 511-515, Pergamon Press 24 Yamasaki, D. S. G. and Wurtz, R. H. (1987) 5oc. Neurosci. Abstr. 13, 625 25 Boussaoud, D., Ungerleider, L. G. and Desimone, R. (1987) Soc. Neurosci. Abstr. 13, 1625
26 Wurtz, R. H. and Goldberg, M. E. (1972)J. Neurophysiol. 35, 587-596 27 Schiller, P. H., True, S.D. and Conway, J.L. (1980) J. NeurophysioL 44, 1175-1189 28 Hikosaka, O. and Wurtz, R. H. (1985) J. NeurophysioL 53, 266-291
Acknowledgements We gratefully acknowledge the contributions of our colleagues, Drs J. A. Movshon, M. R. D~rsteler and A. Mikami to the work reviewed in this paper. Dr Movshon provided the software for the random dot display used in the perceptual experiments, and Drs D~rsteler and MiEami collaborated with us on the smooth pursuit experiments. Preparation of this review and much of the work it describes were supported by a National Eye Institute grant (EY05603) and a Sloan Research Fellowship to WTN.
Cellular interactions during early neurogenesis of Drosophila
Jose A . C a r n p o s - O r t e g a
How are diverse cell types generated from initially homogeneous precursor cells? This question is at the Institutfor heart of developmental biology, and has been studied Entwicklungs- extensively in insects because of their relatively small physiologie, number of embryonic cells and because, in the case of Universit~tzu K61n, Drosophila, they are amenable to genetic analysis. Gyrhofstrasse 17, 5000 K61n41, FRG. During neurogenesis in Drosophila, the progenitor cells of the neurogenic ectoderm give rise to two cell types: neural and epidermal. This developmental switch is not genetically hard-wired, but depends instead on cell-cell communication. These interactions are determined by the products of the neurogenic genes, which seem to provide a molecular signal chain that leads to this neural~epidermal developmental decision. JoseA. CamposOrtega isat the
In insects, neurogenesis is initiated by the separation of individual neural progenitor cells, or neuroblasts (NBs), from a sheet of epidermal progenitor cells called epidermoblasts (EBs) 1-3. In Drosophila melanogaster, the area from which the NBs and EBs derive is termed the neurogenic region (NR), or neuroectoderm3'4, which by the blastoderm stage contains approximately 2000 cellss. About 500 of these cells will move to deeper levels of the embryo to develop as NBs that then form the neural primordium, whereas the remaining 1500 cells remain at the embryonic surface to give rise to an important fraction of the epidermal sheath3,s. Most NBs segregate, i.e. become internalized without having divided since the formation of the blastoderm3, although some arise from cells that have divided once prior to segregation, sometimes producing both neural and epidermal siblings. No morphological differences can be detected among the cells of the NR immediately before the separation of the two lineages occurs3: all neuroectodermal cells of Drosophila become conspicuously large prior to NB segregation; after segregation the EBs shrink considerably. Since the initial large size is indicative of NBs, it was assumed that all neuroectodermal cells are able to adopt the neural fate. Thus, it was thought that all cells of the NR acquire a primary neural fate, which is later substituted in most cells by a secondary, 400
epidermal fate3. Although the overall pattern of NBs and EBs is fairly constant from animal to animal, experimental evidence has shown that the developmental fate of individual cells is not determined a priori 6. When a single labelled cell of the NR of one embryo (donor) is homotopically transplanted into the NR of another (host) embryo, the transplanted cell may develop as either a NB or an EB 6. In some cases, the transplanted cell may undergo a mitosis before the separation of lineages occurs, and daughter cells of both histotypes are found among its progeny. One striking feature of these transplant experiments is that although the proportion of EBs to NBs is normally 3:1, most transplanted cells adopt the neural fate. This observation supports the hypothesis that in normal development the proportion of EBs to NBs is not fixed prior to lineage separation, but that most (or all) neuroectodermal cells have neurogenic capabilities. Since only 500 of 2000 potentially neurogenic cells eventually develop as NBs, a substantial fraction of the 2000 cells must be prevented from adopting the neural fate. Cell interactions and cell determination: a regulatory neuralizing signal? The results of these transplantation experiments indicate that the developmental fate of a cel depends largely on interactions with the neighboring cells of the neuroectoderm (Fig. 1). The homotopic transplantation of ectodermal cells from outside the NR, e.g. from the dorsal ectoderm of the donor into the same region of the host, never yields neural progeny. However, following heterotopic transplantation into the NR, dorsal ectodermal cells occasionally adopt a neural fate6. This could be interpreted as a release from an inhibitory signal that normally acts on cells in the (non-neurogenic) dorsal regions to suppress the neural fate. However, another interpretation is that the surrounding neuroectodermal cells give the transplanted dorsal cells the capability of developing neural progeny. In fact, no intercellular influences that would actively prevent neurogenesis can be experi-
© 1988,ElsevierPublications,Cambridge 0378-5912/88/$02.00
TINS, VOI. 1 I, NO. 9, 1988