The spatial region involved in vernier acuity (integration region) was behaviorally determined in cats and compared with published human data and cortical scaling parameters for the two species. Integration region was estimated from vernier thresholds obtained with: (1) gaps of various sizes between the contiguous ends of the offset contours (gap effects); and (2) irrelevant contours placed at various distances from the vernier offset (interference effects). The gap studies showed that thresholds decreased as the gap grew from 0 to 4–7 min but then increased progressively with separations greater than 7 min. In the interference studies, thresholds were increased 2–6 times when the irrelevant contours were located within 40 min of the offset. Comparing these data with available human measures showed large species differences. Converting the interference distances for both species into cortical distances using published estimates of cortical magnification factor did not resolve the discrepancy. However, a comparison of the cat data with published human data, taken at an eccentricity at which the estimated human cortical magnification factor equals that of the cat, i.e. 7 deg, showed the effects of gap size and interference distance on vernier thresholds to be remarkably similar between the two species. This similarity supports the hypothesis that positional acuity depends upon mechanisms in striate cortex, and also suggests that such acuity may be mediated by a generic neural mechanism common to mammals.

The visual pathways from the nucleus of the optic tract (NOT) to the nucleus reticularis tegmenti pontis (NRTP) and prepositus hypoglossi nucleus (ph) were studied following injections of tritiated leucine into the NOT of pigmented rats. The cell bodies of origin of the pretectal-NRTP, NRTP-ph, and pretectal-ph projections were determined using retrograde horseradish peroxidase (HRP) technique.

The pretectum projects strongly to the rostral two-thirds of the central and pericentral subdivisions of the NRTP and sends a remarkably smaller projection to the ph. Both are entirely ipsilateral. The fibers destined for the ph travel with the NOT-NRTP bundle, pass through the NRTP, traverse the medial longitudinal fasciculus, and are distributed to the rostral one-half of the ph. The retrograde HRP studies confirm these pathways. The pretectal projections to the NRTP arise from neurons in the rostromedial NOT; those to the ph are located primarily in the rostral NOT although small numbers are found within the anterior, posterior, and olivary pretectal nuclei. Of major importance is the fact that the ph injections retrogradely label neurons within the NRTP and the adjacent paramedian pontine reticular formation. This NRTP-ph projection is entirely bilateral and arises from parts of both subdivisions of the nucleus targeted by NOT afferents.

Both the direct NOT-ph and indirect NOT-NRTP-ph connections provide the anatomical basis for the relay of visual (optokinetic) information to the perihypoglossal complex and, presumably, by virtue of reciprocal ph-vestibular nuclear connections, to the vestibular nuclei itself. Such pathways confirm previous physiological studies in rat and, in particular, clarify the contrasting effects of electrolytic lesions of NRTP in rat which completely abolishes optokinetic nystagmus (OKN) (Cazin et al., 1980a) vs kainic acid lesions which produce only minor effects on OKN slow velocity (Hess et al., 1988). Given these differential effects, one concludes that the critical pathway for OKN passes in relation to, but is not significantly relayed by, the neurons of the NRTP or adjacent pontine tegmentum. The present studies suggest that one such fiber system is the NOT-ph bundle. How this relatively small projection compares to other possible fiber of passage systems remains to be determined electrophysiologically.

The effect of binocular competition on the development of ocular activation columns in areas 17 and 18 of cats was studied using the 14C-2-deoxyglucose (14C-2DG) technique to visualize the regions of cortex activated by one eye in cats reared with equal alternating monocular exposure (equal AME), unequal AME, or monocular deprivation (MD). The average size of the ocular activation columns of the eye stimulated during administration of 2DG was positively correlated with the competitive advantage during rearing. In order of increasing percentage of visual cortex activated, the eyes were (1) deprived eye of MD cats, (2) less experienced eye of unequal AME cats, (3) either eye of equal AME cats, (4) more experienced eye of unequal AME cats, and (5) experienced eye of MD cats. In area 17, the shape of the activation columns also was affected by the relative experience of the eye. The columns of the deprived eye of MD cats were widest in layer IV, where they were about the same width as those of the less experienced eye of the unequal AME cats; in other layers they were narrower, sometimes disappearing altogether. In contrast, the activation columns of the less experienced eye of the unequal AME cats were about the same width in all layers. These results suggest that when one eye is placed at a severe disadvantage and receives no patterned input, as in MD, both geniculocortical connections and intracortical connections may be disrupted, but when the disadvantage is less, as in unequal AME, only the geniculocortical connections are disrupted.

Binocular competition also affected the intensity of activation within columns in area 17. We used video densitometry to determine ratios of the amount of label in cortical and thalamic structures. Both the ratio of label in area 17 to that in the lateral geniculate nucleus (LGN) and the ratio of label in the binocular segment of area 17 to that in the monocular segment were significantly less for the deprived eye of MD cats than for any other group. These results suggest that even within the smaller activation columns, deprived geniculocortical afferents are relatively ineffective at driving cortical cells. This finding is consistent with earlier reports that the synapses from the deprived pathway are both morphologically abnormal and reduced in number. The cortical labeling for the less experienced eye of the unequal AME cats and the experienced eye of the MD cats were also significantly less than that in equal AME cats. The decreased labeling for the experienced eye activation columns suggests that, in order to cover an abnormally large area, afferents representing the experienced eye must make fewer synaptic contacts within that area and that there are intrinsic limits on the number of synapses that one axon can make.

This study was performed in order to establish whether selective depletion of serotonin (5-HT) and its metabolite 5-hydroxyindoleacetic acid (5-HIAA) in the pigeon optic tectum (TeO) induced by p-chloroamphetamine (p-CA) modified tectal evoked potentials (TEPs). TEPs in response to sinusoidal gratings of different contrast, spatial and temporal frequency were recorded in control pigeons and in pigeons intraperitoneally injected with p-CA (10 mg/kg; two administrations in consecutive days). TEPs of p-CA treated pigeons, as compared to those of control pigeons, were reduced in amplitude as a function of contrast, spatial and temporal frequency. In addition, TEPs of p-CA treated pigeons differed from those recorded in controls in their transfer characteristics of contrast and spatial frequency. In particular, TEPs of p-CA treated pigeons did not saturate at moderate contrast, unlike those of controls. Furthermore, the TEP spatial tuning in p-CA treated pigeons is broader than that in controls; it thus suggests a reduction of spatial-frequency selectivity. These findings indicate that a selective neurotoxin for serotonergic systems, such as p-CA, can serve as a useful denervation tool for the study of the serotonergic function in the pigeon TeO. In addition, selective changes of TEP properties suggest the possibility that serotonergic afferents play a modulatory role on the receptive-field characteristics of tectal neurons.

Our previous light-microscopic study indicates that unilateral retinal impulse blockage with tetrodotoxin (TTX) causes a reversible decrease of cytochrome oxidase (CO) in alternating rows of metabolically active zones (puffs) in the adult macaque striate cortex (Wong-Riley & Carroll, 1984b). The goal of the present study was to determine if TTX blockade adversely affects all neurons or only a subpopulation of neurons within the puffs. Three major neuronal types were identified based on mitochondrial CO activities and morphological characteristics. Type A neurons were the most prevalent, consisting of small pyramidal and nonpyramidal neurons that received only symmetrical axosomatic synapses. They had little cytoplasm and relatively low levels of CO activity, and showed the least change with TTX treatment. Type B cells were medium-to-large pyramidal neurons that received exclusively symmetrical axosomatic synapses and were moderately reactive for CO. Impulse blockage caused a decrease in mitochondrial size and packing density, but somal size remained within the control range. Type C cells were medium-sized nonpyramidal neurons contacted by both asymmetrical and symmetrical axosomatic synapses. They contained abundant darkly reactive mitochondria and presumably are metabolically the most active. This cell type suffered the greatest decrease in somal size and packing density of mitochondria, particularly the darkly reactive ones. A rare fourth cell type, type D, was a small, darkly reactive nonpyramidal variety that gave rise to somatodendritic synapses. Their low occurrence prevented statistical analysis under normal and TTX-treated conditions. These data indicate that retinal impulse blockade is most detrimental to the metabolically most active neurons in the adult primate cortical puffs. The alterations are not permanent, because the effects of TTX are fully reversible (Carroll & Wong-Riley, 1987).

The size, pattern of terminal arborizations, and laminar specificity of individual retinogeniculate axons were studied in the macaque following injections of HRP into the optic tract. Axons that terminated in the magnocellular layers had significantly larger fiber diameters and wider terminal fields than those that terminated in the parvocellular layers. Terminal fields of magnocellular fibers spanned most of the width of their target layer, whereas those of parvocellular fibers were restricted to approximately one-half the width of their target layers; almost all terminal fields were oriented along lines of projection. All of the optic tract fibers that we examined terminated in only one layer of the lateral geniculate nucleus (GL), including a population of fine caliber fibers that project to the intercalated layers, and none had collateral projections outside the GL. The results suggest that each layer – magnocellular, parvocellular, and intercalated – receives projections from a morphologically distinct population of optic tract fibers.

Layers II/III of the primary visual cortex contain a regular pattern of histochemically detectable cytochrome oxidase (CO)-rich “puffs,” which differ from the interpuff regions in their thalamo-cortical and cortico-cortical connectivity, receptive-field properties, and the density of inhibitory GABA-containing synaptic terminals. We used an immunocytochemical method, in combination with cytochrome oxidase histochemistry, to analyze the spatial relationship between neurons that contain neuropeptide Y (NPY) and the CO puffs. Of a total of 606 neurons, only 2.6% of the NPY-containing cells are located in the puffs, whereas the rest are situated in the interpuffs, or at the interface between puffs and interpuffs. The number of NPY-containing neurons in the puffs is substantially less than that expected in an equal volume of the interpuffs (X2 = 13.86; df = 1; P < 0.001).

These observations indicate that columns containing the puffs may differ also from those in the interpuff regions in that they contain a unique array of chemically and morphologically distinct local circuit neurons.

We investigate the neural computations underlying pattern processing with stimuli based on textures balanced for spatial frequency content (and second-order correlations) but not for higher-order correlations (Julesz et al. 1978). Interchange between two such isodipole textures produces a robust human visual evoked potential (VEP). The difference in population activity driven by two isodipole textures is quantified by the antisymmetric component of the VEP. Statistical properties of the textures eliminate contributions from linear mechanisms to the antisymmetric VEP.

The dependence of the antisymmetric VEP on check size and fourth-order correlation statistics is used to test nonlinear models for the underlying neural computations. Linear summation, followed by a simple nonlinearity (such as rectification, saturation, or threshold), is inconsistent with the data. More elaborate models, in which a second nonlinear stage combines the output of local nonlinear mechanisms, are consistent with the data, provided that an appropriate spatial scale is chosen for the second stage of processing. For checks 4 min or smaller, the deduced interaction length is 10–15 min. For checks larger than 4 min, the interaction length is proportional to check size.

The morphology and distribution of neurons labeled specifically by the lectin, Vicia villosa (VVA), were examined in striate cortex of adult macaque monkeys. Following incubation with VVA conjugated to histochemical markers, fine punctate reaction product appears to cover the surface of the soma and proximal dendrites of a population of cortical neurons. Although a small number of VVA-labeled cells are located in layers 2, 3A, 5, and 6, approximately 75% are located in a strip of cortex overlying layers 3B through 4Ca. Layers 1 and 4Cβ are virtually devoid of labeled cells. The morphology of labeled cells varies throughout the layers. In the supragranular layers, the labeled cells generally display a round or multipolar soma with a small number of radially disposed dendrites. In deeper layers, labeled cells are multipolar or horizontal, and their proximal dendrites are often more densely labeled. There is no clear correlation between the distribution of labeled cells and the pattern of cytochrome oxidase staining in supragranular layers.

Double labeling of single sections for VVA and for GABA (gamma-aminobutyric acid) immunoreactivity revealed that most VVA-labeled cells are also immunoreactive for GABA. The double-labeled cells comprise approximately 30% of all GABA immunoreactive cells. Soma size analysis of double-labeled cells shows that medium-to-large GABA cells in each layer are labeled by VVA. The soma size, laminar distribution, and morphology of the VVA-labeled GABA cells suggest that they include the large basket cells originally observed in Golgi preparations.

The tectum of Xenopus frogs receives input from both eyes. The contralateral eye's projection reaches the tectum directly, via the optic nerve, and the ipsilateral eye's projection reaches the tectum indirectly, via the nucleus isthmi. Under normal conditions, the topography of the ipsilateral map relayed from the nucleus isthmi is in register with the topography of the retinotectal map from the contralateral eye. During development, the process of aligning the two maps is complicated by the dramatic changes in binocular overlap of the two eyes' visual fields which take place during late tadpole and juvenile stages. The goal of this study is to determine the branching patterns of contralaterally projecting isthmotectal axons before, during, and after the period of rapid eye migration.

Isthmotectal axons were filled by anterograde transport of horseradish peroxidase (HRP) from the nucleus isthmi. The results show that crossed isthmotectal axons enter the entire extent of the tectum before binocular overlap begins to increase. Therefore, binocular overlap is not necessary for the initial isthmotectal projection to span the tectum. The density of isthmotectal branches rises dramatically at the same time that the eyes begin to shift. During the period when eye migration is most rapid, many isthmotectal axons form arbors which resemble adult arbors but which extend over greater proportions of the tectal surface.

The axons appear to be directed toward appropriate mediolateral positions as they enter the tectum. Their trajectories are roughly rostocaudal, with relatively little change along the mediolateral dimension. These data, when combined with available physiological data, suggest that mediolateral order is initially established by vision-independent mechanisms but can be altered by vision-dependent mechanisms. Rostrocaudal order becomes discernable only at the time when binocular visual cues become available and appears to be established primarily on the basis of the activity of the retinotectal and isthmotectal axons.

The responses of neurones in the cat's visual cortex are very variable in amplitude. Thus, although the average response amplitude of a single neurone depends closely upon the contrast of a sinusoidal grating, the instantaneous amplitude of the response can convey little information about the grating's contrast. This paper shows that a typical cortical neurone can convey less than one bit of information about contrast in 0.5 s. The amount of information that a neurone can convey is closely correlated with the neurone's responsivity.

Unilateral retinal impulse blockage with tetrodotoxin (TTX) induces reversible shrinkage and decreased cytochrome oxidase (CO) activity in alternate rows of supragranular, CO-rich puffs in the adult macaque striate cortex (Wong-Riley & Carroll, 1984b: Carroll & Wong-Riley, 1987). The present study extended the findings to the electron-microscopic (EM) level to determine if various neuropil profiles in control puffs exhibit heterogeneous levels of CO activity, and whether specific processes were more susceptible to intravitreal TTX than others.

Within the neuropil of control puffs, 60% of the total mitochondrial population resided in dendrites, and the majority of dendritic mitochondria were highly reactive for CO. Axon terminals forming symmetrical synapses also contained darkly reactive mitochondria, whereas those forming asymmetrical synapses possessed very few and mainly lightly reactive mitochondria. Unmyelinated axon trunks, myelinated axons, and glia all exhibited low levels of CO activity. Synaptic count revealed a 3:1 ratio of asymmetrical to symmetrical synapses.

Intravitreal TTX for 2–4 weeks adversely affects dendrites and symmetrical terminals much more so than other neuropil processes. There was a general decrease in darkly and moderately reactive mitochondria and an increase in lightly reactive mitochondria throughout the puffs, especially in dendrites. This indicates that afferent blockade is more detrimental to processes of higher metabolic activity. Changes also differed between central and peripheral regions of puffs, and indications of axonal and synaptic reorganization were more evident in the latter. Thus, stabilization of neuronal structure and synapses appears to be activity-dependent even in the adult. A working model of these metabolic and morphological responses to chronic TTX is proposed.

The two main receptor subtypes for 5-hydroxytryptamine (5HT) were measured and localized in visual cortical areas of macaque monkey. [3H]5HT was used to label all 5HT-1 receptor subtypes and [3H]ketanserin was used to label 5HT-2 receptors. Both receptor types could be demonstrated in membranes prepared from macaque primary visual cortex. The specificity of these ligands for 5HT-1 or 5HT-2 receptors was demonstrated by the pharmacological profile of inhibitors of the specific binding. 5HT-1A receptor sites were detected by displacement experiments and by direct labeling with [3H]8-hydroxy-2(di-n-propylamino) tetralin 8OH-DPAT. Receptor autoradiography showed that the distribution of these receptor subtypes varied from one part of visual cortex to another. 5HT-1 receptors, labeled with [3H]5HT were present in several bands through layer IV of primary visual cortex with the densest band seen in and above layer IVA: another band was in lower layer VI. The band in layer VI was predominantly 5HT-1A sites. There were two main bands of 5HT-2 receptor sites, the most prominent around the IV/V boundary, and the other extending from layer IVA upwards. Adjacent areas showed 5HT receptors in a broad band corresponding to layer IV. 5HT-1A sites were found in superficial layers of adjacent areas, except V2. These layering patterns did not correspond precisely with cytoarchitectonic layering, nor with the pattern of 5HT-containing presynaptic fibres in published reports. It is important, therefore, in considering the role of the 5HT-containing neurons in cortical function to take account not only of the anatomy of the presynaptic terminals, but also of the postsynaptic receptors upon which the released transmitter will act, and their location within the cortex.

A 30-deg-high horizontally rotating random-dot display was presented to the central field, and with its more central edge at vertical eccentricities of 0, 2.5, 5, and 10 deg above or below the horizon. Stimulus velocities of 25–100 deg/s and two directions of motion were presented. The mean gain of the slow phases of optokinetic nystagmus (OKN) for five subjects was significantly higher when the stimulus was presented to the lower visual field than when the stimulus was presented to the upper field. This difference was most pronounced when the display was displaced 5 deg from the fovea and moving below 100 deg/s. Our results are consistent with existing psychophysical and physiological evidence for the superiority of the upper retina. In addition, four of the five observors showed significant directional asymmetries.

Studies of the physiology of the central visual system and visual behavior frequently require large tangent screens, and it has been difficult to create a smooth surface with uniform illumination. We have developed a technique for holding drafting paper on a sheet of acrylic so that it can be used as a tangent screen for both front and rear projections. The process utilizes a vacuum system that holds the paper firmly and produces a uniformly smooth surface with good light diffusion; it is simple, flexible, and inexpensive and can be used for any size screens.

Immunocytochemical methods were used to examine neuropeptide Y (NPY) immunoreactive neurons and fibers in area 17 of rhesus monkeys during the first year of life. NPY-immunoreactive (+) neurons are nonpyramidal cells which are either multipolar, bipolar, or bitufted in shape. They occur most frequently in layer 6 and the subjacent white matter, are sparser in the supragranular layers, and absent from layer 4C. Labeled somata in the supragranular layers are smaller compared to those in layer 6 and the white matter. A typical axon originates from the NPY+ soma or from a primary dendrite and frequently is varicose. Distribution and morphologies of NPY+ neurons in area 17 of infants are similar to those of adult monkeys. Thus, it seems that NPY+ neurons in rhesus monkeys are mature from birth. NPY+ fibers occur in area 17 from birth; however, they differ in density and distribution from those of older infant and adult monkeys. At birth, a prominent fiber plexus is found in the deepest part of layer 1, and another in the white matter. Immunoreactive processes are sparse in the remaining cortical gray, except for some vertical fibers extending from pia to white matter. By 4 months of age, labeled fibers form a coarse network in layers 2, 3, 5, and 6. In addition, a distinct plexus extends through layers 4B, 4A, and the lowest aspect of layer 3. Also, a thin immunoreactive fiber band is found at the bottom of layer 4C. In the remainder of layer 4C, NPY+ fibers are scant. The supragranular layers also exhibit a unique immunoreactive “snarl” of fibers. Increases in density of NPY+ processes in the older infants are gradual so that between 7 and 13 months of age, NPY+ fibers appear to have achieved adultlike densities. These observations indicate that NPY+ fibers in area 17 of newborn rhesus monkeys undergo postnatal maturation which reaches a plateau around 4 months of age. After monocular visual deprivation from birth to 4 months of age, either by eyelid suture or by occlusion with an opaque contact lens, density and distribution of NPY+ neurons and fibers, including snarls, appear similar to those of age-matched undeprived infants. Thus, disruption of the normal binocular input does not seem to arrest the maturation of the NPY system in area 17 of rhesus monkeys during a sensitive period of early postnatal development.

The sensitivity of neurons in area 17 of the cat's visual cortex to vernier offset was expressed as the percentage reduction in response caused by the introduction of a given offset into a bar stimulus moving across the receptive field. There was a wide variation in sensitivity: in some cells response could be halved by an offset equal to a fifth of receptive-field width (defined as twice the standard deviation of a Gaussian curve fitted to the response profile), while other cells showed no sensitivity. The highest absolute sensitivities of complex and simple cells were similar, although most cells with poor sensitivity were complex.

Sensitivity was largely unaffected by changes in stimulus velocity and stimulus length, although there was a tendency for sensitivity to increase with decreasing bar length.

Comparisons of orientation tuning curves with vernier tuning curves showed that the response to a vernier stimulus approximated the response to a single bar of the same overall length and an orientation equal to that of a line joining the midpoints of each bar. This was true for a wide range of sensitivity values.

Vernier sensitivity was correlated with a measure of length summation H, which is positive when there is net facilitation between the bars, and negative when there is net inhibition. Vernier sensitivity was highest in cells with large values of H, and least in cells where H was negative.

We examined a linear model of the simple cell receptive field which, together with a variable response threshold, was able to explain the correlation between vernier acuity and length summation. Although this model accounted qualitatively for many of our findings, the majority of simple cells had tuning curves that were sharper than the predicted ones. This suggests that there are nonlinearities in the behavior of many simple cells whose effect is to increase the sharpness of orientation tuning and consequently vernier sensitivity.