Simple cells in striate cortex have been depicted as rectified linear operators, and complex cells have been depicted as energy mechanisms (constructed from the squared sums of linear operator outputs). This paper discusses two essential hypotheses of the linear/energy model: (1) that a cell's selectivity is due to an underlying (spatiotemporal and binocular) linear stage; and (2) that a cell's firing rate depends on the squared output of the underlying linear stage. This paper reviews physiological measurements of cat striate cell responses, and concludes that both of these hypotheses are supported by the data.

Previous biochemical and immunohistochemical studies have shown that the neurotensin-related hexapeptide LANT6 is widespread and abundant in the avian nervous system. In the present study, immunohistochemical techniques were used to show that LANT6 is present in numerous cells of the retinal ganglion cell layer in pigeons. Consistent with the possibility that these LANT6+ retinal cells might be retinal ganglion cells, it was found that (1) the distribution of LANT6+ fibers and terminals in the central retinal target areas matched the distribution of central retinal projections; (2) the LANT6+ fibers and terminals are eliminated from retinal target areas by transection of the contralateral optic nerve; and (3) LANT6+ retinal cells in the ganglion cell layer can be retrogradely labeled by injections of fluorogold in the tectum. These results suggest that LANT6 may be utilized as a neuroactive substance by the central terminals of numerous retinal ganglion cells in birds. Similar anatomical findings have been previously reported for members of several other vertebrate groups, giving rise to the possibility that LANT6 (or its homologues in nonavians) may be a phylogenetically ubiquitous neuroactive substance used by retinal ganglion cells.

The callosal connections between visual cortical areas 17 and 18 in adult normally pigmented and “Boston” Siamese cats were studied using degeneration methods, and by transport of WGA-HRP combined with electrophysiological mapping. In normal cats, over 90% of callosal neurons were located in the supragranular layers. The supragranular callosal cell zone spanned the area 17/18 border and extended, on average, some 2–3 mm into both areas to occupy a territory which was roughly co-extensive with the distribution of callosal terminations in these areas. The region of the visual field adjoining the vertical meridian that was represented by neurons in the supragranular callosal cell zone was shown to increase systematically with decreasing visual elevation. Thus, close to the area centralis, receptive-field centers recorded from within this zone extended only up to 5 deg into the contralateral hemifield but at elevations of -10 deg and -40 deg they extended as far as 8 deg and 14 deg, respectively, into this hemifield. This suggests an element of visual non-correspondence in the callosal pathway between these cortical areas, which may be an essential substrate for “coarse” stereopsis at the visual midline.

In the Siamese cats, the callosal cell and termination zones in areas 17 and 18 were expanded in width compared to the normal animals, but the major components were less robust. The area 17/18 border was often devoid of callosal axons and, in particular, the number of supragranular layer neurons participating in the pathway were drastically reduced, to only about 25% of those found in the normally pigmented adults. The callosal zones contained representations of the contralateral and ipsilateral hemifields that were roughly mirror-symmetric about the vertical meridian, and both hemifield representations increased with decreasing visual elevation. The extent and severity of the anomalies observed were similar across individual cats, regardless of whether a strabismus was also present. The callosal pathway between these visual cortical areas in the Siamese cat has been considered “silent,” since nearly all neurons within its territory are activated only by the contralateral eye. The paucity of supragranular pyramidal neurons involved in the pathway may explain this silence.

Introduction

Classical views of the optic chiasm maintain four propositions about the retinofugal pathways: (1) each optic nerve contains a retinotopic representation of its respective retinal surface; (2) this retinotopic map in the nerve is the basis for the subsequent segregation of the decussating from the non-decussating fibers; (3) this retinotopy in the nerve is also the basis for the presence of retinotopy found within the half-retinal maps in the optic tracts; and (4) the half-retinal maps from each optic nerve are brought together within the chiasm to yield a unified, binocularly congruent, map in the optic tract (Brodal, 1969; DukeElder, 1961; Polyak, 1957; Wolff, 1940). The appeal of this classical view is in its simplicity, based on the assumption that the retinofugal pathway should replicate the sensory surface along its course. We now know that each of these four propositions is incorrect, and that the error is not one simply of degree or extent (Guillery, 1982, 1991). Rather, the above description of the visual pathway is fundamentally flawed because it has failed to take into account the constraints under which the pathway develops. We shall first consider the evidence for rejecting the classical view, from recent studies on the organization of the retinofugal pathway in adult animals and on the development of that organization. We shall then describe three transformations in the fiber order which all occur in the chiasmatic region, two of which were only recently recognized, and for which we must account.

Observations from adult organization

The difference in the fiber order in the optic nerve and tract

Some visual stimuli produce a strong percept of motion, even though they fail to excite motion detectors based on Fourier energy or cross correlation. Models which suffice to explain the motion percept in these non-Fourier motion (NFM) stimuli include linear spatiotemporal filtering, followed by rectification, followed by standard motion analysis (Chubb & Sperling 1988). We used the human “motion-onset” evoked potential, which has been assigned to area 17 on the basis of work in the macaque (van Dijk et al., 1986; van Dijk & Spekreijse, 1989), to investigate the neural substrate of the processing stages postulated in the above models. Motion-onset VEPs elicited by FM and NFM matched for spatial and temporal characteristics were indistinguishable in temporal characteristics and scalp topography at a transverse chain of electrodes. Addition of textural cues (granularity and higher-order form) did not influence the response dynamics or scalp topography of NFM responses. However, comparison of responses to NFM stimuli and related stimuli without coherent motion but similar spatial and temporal properties showed that the motion-onset responses were distinct from responses to the onset of fixed flicker-defined contours not undergoing coherent motion. We discuss the implications of these results for computational models of motion analysis.

The effects of bilateral lesions of the centrifugal visual system (CVS) on the visual-discrimination capacity were studied in pigeons. Three different behavioral experiments, each testing different aspects of visual analysis, were performed. In the first two experiments, a grain-grit discrimination task and a visual-acuity determination, stimuli were presented in the frontal binocular visual field. A third experiment investigated the early detection of slow moving objects, introduced into the monocular lateral visual field. After bilateral lesions in the nucleus isthmo-opticus (ION) and in the ectopic nucleus isthmo-opticus (EION), a multiple linear regression analysis was employed to correlate the postoperative performance in all three tasks with the amount of structure loss within ION and EION. Deficits in the grain-grit discrimination procedure were a function of the ION lesion extent and did not depend on EION damage. Thus, these two structures could be functionally differentiated for the first time. Neither the ION nor the EION seems to be involved in visual- acuity performance or the early detection of large shadows moving forward through the visual field. Our data support the hypothesis that the CVS is involved in pecking and food selection among static stimuli at a short viewing distance in ground-feeding birds such as pigeons and chickens.

The discharges of ON- and OFF-center X and Y retinal ganglion cells in the presence of stationary patterns or of a uniform field of photopic luminance were recorded from urethane-anesthetized adult cats. The interval statistics and power spectra of these discharges were determined from these discharge records. The patterned stimuli were selected and positioned with respect to a cell's receptive field so as to generate steady discharges that were different in mean discharge rate from that cell's discharge for the diffuse field. The interval statistics of discharges recorded for diffuse or patterned illumination for all cell types can be modeled, approximately, as coming from renewal processes with gamma-distributed intervals. The gamma order of the interval distributions was found to be nearly proportional to the mean discharge rate for X cells, but not for Y cells. Typical values for the gamma orders and their dependence on mean rate for different cell types are given. The same model of a renewal process with gamma-distributed intervals is used to model the measured power spectra and performs well. When the gamma order is proportional to mean rate, the power spectral density at low temporal frequencies is independent of discharge rate. Gamma order was proportional to mean rate for X cells but not for Y cells. Nonetheless, the power spectral densities of both cell types at low frequencies were approximately independent of discharge rate. Hence, noise in this band of frequencies can be considered additive. The consequences of departures from the renewal process and of the gamma order not being proportional to mean rate are considered. The significance of different rates of discharge for signaling is discussed.

Immunocytochemical staining for microtubule-associated protein 2 (MAP 2) was used to examine the morphology of neurons, the organization of neuronal groups, and the neurochemical plasticity of cells in adult monkey area 17. MAP 2-immunostained neurons are present through the depth of area 17 but are most intensely immunoreactive in layers IVB and VI. From layer IVB, separate groups of MAP 2-positive cells invade layers 1VA and IVCα. Clusters of cells protrude upward from superficial layer IVB and occupy the central core regions of the cytochrome oxidase (CO)-stained honeycomb in layer IVA, while large neurons typical of layer IVB are distributed in irregular clusters in the subjacent layer IVCa. The somata in the layer IVA honeycomb cores give off immunostained dendrites which remain largely within the core regions. Patches of MAP 2-positive neurons are also present in layers II and III, where they coincide with the CO-stained puffs.

Intravitreal injections of tetrodotoxin (TTX) into one eye of adult monkeys produce stripes of alternating light and dark MAP 2 immunostaining in layer IVC. Stripes of light immunostaining coincide with stripes of light CO staining, and correspond to reduced MAP 2 immunoreactivity within cortical neurons dominated by the TTX-injected eye. In layers II and III, the MAP 2 immunostaining of patches overlying the injected-eye columns is similarly reduced. No change occurs in the MAP 2 immunostaining of layer IVA.

These data suggest that the anatomical and physiological heterogeneity of layers IVA and IVCα arises from the periodic invasion of neurons characteristic of layer IVB, that the neurons in layer IVA have dendrites confined to thalamocortical-recipient or nonrecipient zones and that the expression of MAP 2 changes in adult cortical neurons following the loss of retinal input.

Visual cortical areas are richly but selectively connected by “patchy” projections. We characterized these connections physiologically with cross-correlograms (CCHs), calculated for neuron pairs or small groups located one each in visual areas 17 and 18 of the cat. The CCHs were then compared to the visuotopic and orientation match of the neurons' receptive fields (RFs).

For both spontaneous and visually driven activity, most non-flat correlograms were centered; i.e. the most likely temporal relationship between spikes in the two areas is a synchronous one. Although spikes are most likely to occur simultaneously, area 17 spikes may occur before area 18 or vice versa, giving the cross-correlogram peak a finite width (temporal dispersion). Cross-correlograms fell into one of three groups according to their full-width at half peak height: 1–8 ms (modal width, 3 ms), 15–65 ms (modal width 30 ms), or 100–1000 ms (modal width 400 ms). These classificatory groups are nonoverlapping; the three types of coupling appeared singly and in combination.

Neurons whose receptive fields (RFs) are nonoverlapping or cross-oriented may yet be coupled, but the coupling is more likely to be the broadest type of coupling than the medium-dispersed type. The sharpest type of coupling is found exclusively between neurons with at least partially overlapping RFs and mostly between neurons whose stimulus orientation preferences matched to within 22.5 deg. The maximum spatial dispersion observed in the RFs of coupled neurons compares well with the maximum divergence seen anatomically in the A18/A17 projection system.

We suggest three different mechanisms to produce each of the three different degrees of observed spatial and temporal coherence. All mechanisms use common input of cortical origin. For medium and broad coupling, this common input arises from cell assemblies split between both sides of the 17/18 projection system, but acting synchronously. Such distributed common-input cell assemblies are a means of overcoming sparse connectivity and achieving synaptic transmission in the pyramidal network.

Visual pigment, sensitivity, and rhabdomere size were measured throughout a 12-h light/12-h dark cycle in Drosophila. Visual pigment and sensitivity were measured during subsequent constant darkness [dark/dark (D/D)]. MSP (microspectrophotometry) and the ERG (electroretinogram) revealed a cycling of visual pigment and sensitivity, respectively. A visual pigment decrease of 40% was noted at 4 h after light onset that recovered 2–4 h later in white-eyed (otherwise wild-type, w per+) flies. The ERG sensitivity [in w per+ flies in light/dark (L/D)] decreased by 75% at 4 h after light onset, more than expected if mediated by visual pigment (MSP) changes alone. ERG sensitivity begins decreasing 8 h before light onset while decreases in visual pigment begin 2 h after light onset. These cycles continue in constant darkness (D/D), suggesting a circadian rhythm. White-eyed period (per) mutants show similar cycles of visual pigment level and sensitivity in L/D; per's alterations, if any on the D/D cycles were subtle. The cross-sectional areas of rhabdomeres in w per+ were measured using electron micrographic (EM) morphometry. Area changed little through the L/D cycle.

The relative timing of a number of events during the development of the visual system has recently been suggested to be consistent across a number of mammalian species (Dreher & Robinson, 1988). Some conflicting reports, however, had suggested that the precocial guinea pig might represent an exception to the generalized scheme. A quantitative study was thus carried out on the development of the optic nerve and retina of the guinea pig. Consistent with the prediction of a stable relative time course of mammalian visual development, axons and growth cones were found in the optic stalk from the 24th postconceptional day (40% of the period from conception to eye opening —the cecal period), the peak number of axons was observed on the 32nd postconceptional day (56% of the cecal period), and the phase of rapid axonal loss extended to the 39th and 42nd postconceptional days (68–74% of the cecal period). The number of axons in the adult optic nerve (117,000) represented about 37% of the peak number of axons. Additional observations indicated that during development of the optic nerve the mean axonal diameter increased approximately threefold from 0.31 μm to 1.06 μm. As in other mammals studied so far, myelination was first noted after the period of rapid axonal loss and continued until in the adult 97% of axons were found to be myelinated. In the retina, the presence of pyknotic profiles in the ganglion cell layers extends throughout the periods of loss of the optic nerve axons. Finally, the presence of pyknotic profiles in the amacrine sublayer suggest that in the guinea pig, as in other mammalian species, there is a loss of displaced amacrine cells as well as ganglion cells from the ganglion cell layer.

Long- and medium-wavelength cones in the turtle retina participate in complex neural interactions. They are coupled via excitatory pathways to other cones and receive negative feedback inputs from luminosity-type horizontal cells. Little information has been collected on the S- (short-wavelength or blue) cones because they are scarce in the turtle retina and of smaller dimensions compared to the other cone types.

In this paper, flash sensitivity action spectra and photoresponses of seven turtle S-cones were measured in the dark-adapted state and during chromatic background illuminations. The desensitizing action of monochromatic background lights was not uniform across the visible spectrum. A red background was most effective in desensitizing the S-cones to long-wavelength stimuli while a blue background light produced its strongest action on the photoresponses elicited by short-wavelength stimuli. The effects of chromatic adaptation on the S-cone action spectrum and on the kinetics of the small-amplitude photoresponses suggested that the S-cones in the turtle retina were involved in complex neural interactions. These included excitatory inputs probably originating in neighboring L-cones and inhibitory long-wavelength inputs probably mediated by L-type horizontal cells.

We have studied the responses of striate cortical neurons to stimuli whose contrast is modulated in time by either a single sinusoid or by the sum of eight sinusoids. The sum-of-sinusoids stimulus resembles white noise and has been used to study the linear and nonlinear dynamics of retinal ganglion cells (Victor et al., 1977). In cortical neurons, we have found different linear and second-order responses to single-sinusoid and sum-of-sinusoids inputs. Specifically, while the responsivity near the optimal temporal frequency is lower for the sum-of-sinusoids stimulus, the responsivity at higher temporal frequencies is relatively greater. Along with this change in the response amplitudes, there is a systematic change in the time course of responses. For complex cells, the integration time, the effective delay due to a combination of actual delays and low-pass filter stages, changes from a median of 85 ms with single sinusoids to 57 ms with a sum of sinusoids. For simple cells, the integration times for single sinusoids range from 44–100 ms, but cluster tightly around 40 ms for the sum-of-sinusoids stimulus. The change in time constant would argue that the increased sensitivity to high frequencies cannot be explained by a static threshold, but must be caused by a fundamental alteration in the response dynamics. These effects are not seen in the retina (Shapley & Victor, 1981) and are most likely cortical in origin.

Visual pigments from the red rods of adults of eight species of Australian anuran amphibians, from a variety of habitats, were analyzed by microspectrophotometry. The λmax in all cases fell between 502 nm and 506 nm, and the absorption spectra were well fitted by an A1-based visual pigment template curve. Red rod pigments were also analyzed for a number of tadpoles. In some cases the data were best fitted with an A1based visual pigment template, in other cases with an A2-based template, and finally some tadpoles appeared to have mixtures of the two pigments.

R1–6 dominated electroretinographic (ERG) spectral sensitivities were determined as a function of days posteclosion from carotenoid deprived and replaced white-eyed Drosophila. The sensitivity of flies deprived from egg to adult waxed (about 1.5 log units by day 3), and then waned gradually from 3–11 days (over 2 log units by day 11). Carotenoid replacement (feeding nothing but carrot juice) effected recovery to near the replete control' level in about 1 day throughout (tested at 0, 4, and 11 days). The normal yellow cornmeal-agar-molasses-brewers yeast fly food (in our laboratory, supplemented with β-carotene) renders a slower recovery (requiring 7–9 days) since it is a medium designed largely for larval growth. Placing replete adults on deprivational medium did not create a deprivational syndrome in over 11 days. At 3–7 days, deprived flies reared and maintained in constant darkness had substantially enhanced sensitivity, beyond the 1.5 log unit increment already described for cyclic light rearing conditions. All spectral analyses are consistent with the ultraviolet (UV) sensitization of the blue (480 nm) rhodopsin by a replacement-dependent retinoid including two unexpected findings: (1) sensitivity recovery with carrot juice was so fast that the UV peak was already high at 6 h; and (2) the waxing of the deprived fly's sensitivity in dark rearing was so great that the UV peak was present at 4–7 days.

The calcium-binding proteins calbindin (CaBP) and parvalbumin (PV) are important in regulating intracellular calcium in brain cells. PV immunoreactivity is reduced by enucleation in the lateral geniculate nucleus (LGN) and by enucleation and visual deprivation in the striate cortex of adult monkeys. The effects of enucleation and visual deprivation on CaBP immunoreactivity in the LGN are not known. We therefore have studied cells and neuropil in the LGN that are labeled by antibodies to CaBP in normal and visually deprived Rhesus monkeys to determine if there is an effect on this calcium-binding protein. One group of monkeys had one eye removed 2 weeks to 4.3 years before sacrifice. A second group had one eye occluded with opaque lenses from infancy without enucleation. A final group had one eye occluded long-term followed by short-term enucleation 2 weeks before sacrifice.

In normal monkeys, CaBP-immunoreactive neurons were found throughout the LGN. They were sparsely distributed within the six main laminae, and more densely distributed within layer S and the interlaminar zones (ILZ). The labeled ILZ neurons had a distinct morphology, with fusiform somata and elaborate dendritic trees that were confined primarily to the ILZ. Most CaBP-labeled neurons in the main layers had dendrites that radiated in all directions from the soma. ILZ and main layer cells labeled by CaBP thus probably represent two different cell types.

Monocular enucleation with or without occlusion produced a significant reduction in antibody labeling in the deafferented laminae. Field measures revealed an average 11.5% reduction in optical density in each deafferented lamina compared to its adjacent, nondeprived layer. The differences in field optical density between deprived and nondeprived layers were statistically significant. CaBP neurons were still visible, but the optical density of antibody labeling in these cells also was reduced. Occlusion without enucleation had no effect. Thus, deafferentation, but not light deprivation, reduces concentrations of CaBP in monkey LGN. This effect is different than that seen in striate cortex of adult monkeys, where visual deprivation as well as enucleation alters CaBP immunoreactivity.

The developmental time courses for vernier acuity and grating acuity were measured longitudinally in infant Macaca nemestrina monkeys. Behavioral measurements of vernier and grating acuity were made at regular intervals during development. Near birth, grating acuity is relatively more mature than vernier acuity. The proportional rate of vernier acuity development is faster than that for grating acuity. During the course of development, grating acuity improves approximately 15-fold whereas vernier acuity improves about 60-fold. Both visual functions approach adult levels at about the same age, around 40 weeks postnatally. Although grating acuity develops about four times faster in monkeys than in humans, vernier acuity development in monkeys and humans does not appear to reflect the same relationship. Adult levels of vernier acuity for the monkeys are about a factor of 2 poorer than are typically reported for humans. The differential development of vernier acuity and grating acuity does not necessarily reflect development at different levels of the visual system.

The motion after-effect (MAE) lasts longer when the test period does not immediately follow adaptation, a phenomenon called storage. Does storage of the MAE occur if the test target is present but rendered phenomenally invisible owing to the presence of a rival target presented to the other eye during the storage period? Our experiment addressed this question. Following adaptation to a drifting grating, an intervening period preceded testing with a stationary grating. During this period, the adapted eye either viewed the test target immediately or was occluded, and the unadapted eye either viewed a high-contrast rival target or was occluded. Thus four conditions were employed. The duration of the residual MAE was found to be longer for the rivalry condition (grating and rival target viewed) than for the normal MAE condition (grating viewed), and comparable to that in the stored MAE condition (both eyes occluded). Thus, the MAE is stored when the test target is rendered invisible due to binocular rivalry, indicating that a suppressed target is ineffective at promoting decay of the MAE. So while suppression does not prevent information about the adapting grating from reaching the site of generation of the MAE (Lehmkuhle & Fox, 1975), it can prevent information about the test target from reaching the site of the stored MAE. Current models attribute the MAE to reduced responsiveness of direction-selective cortical neurons (Sutherland, 1961; Barlow & Hill, 1963). Thus, storage should reflect a differential return of these adapted cells to preadapted response levels, dependent on postadaptation stimulation. From our results we deduce that storage does not occur at all sites at which motion adaptation occurs. Rather, decay of the MAE is dependent on postadaptation stimulation at higher levels of adaptation, and independent at earlier levels.

In an aseptic operation under surgical anesthesia, one optic nerve of a cat was exposed and subjected to pressure by means of a special cuff. The conduction of impulses through the pressurized region was monitored by means of electrodes which remained in the animal after the operation. The pressure was adjusted to selectively eliminate conduction in the largest fibers (Y-type) but not in the medium-size fibers (X-type). The conduction block is probably due to a demyelination and remains complete for about 3 weeks. Within 2 weeks after the pressure-block operation, recordings were made from single neurons in the striate cortex (area 17, area VI) of the cat anesthetized with N2O/O2 mixture supplemented by continuous intravenous infusion of barbiturate. Neurons were activated visually via the normal eye and via the eye with the pressure-blocked optic nerve (“Y-blocked eye”). Several properties of the receptive fields of single neurons in area 17 such as S (simple) or C (complex) type of receptive-field organization, size of discharge fields, orientation tuning, direction-selectivity indices, and end-zone inhibition appear to be unaffected by removal of the Y-type input. On the other hand, the peak discharge rates to stimuli presented via the Y-blocked eye were significantly lower than those to stimuli presented via the normal eye. As a result, the eye-dominance histogram was shifted markedly towards the normal eye implying that there is a significant excitatory Y-type input to area 17. In a substantial proportion of area 17 neurons, this input converges onto the cells which receive also non-Y-type inputs. In one respect, velocity sensitivity, removal of the Y input had a weak but significant effect. In particular, C (but not S) cells when activated via the normal eye responded optimally at slightly higher stimulus velocities than when activated via the Y-blocked eye. These results suggest that the Y input makes a distinct contribution to velocity sensitivity in area 17 but only in C-type neurons. Overall, our results lead us to the conclusion that the Y-type input to the striate cortex of the cat makes a significant contribution to the strength of the excitatory response of many neurons in this area. However, the contributions of Y-type input to the mechanism(s) underlying many of the receptive-field properties of neurons in this area are not distinguishable from those of the non-Y-type visual inputs.

We showed high-contrast, second-order motion stimuli to subjects whilst recording their horizontal eye movements. These stimuli were very poor at evoking optokinetic nystagmus. Smooth-pursuit eye movements and fixation were reduced by a masking band ±2.5 deg above and below an imaginary fixation point. First-order stimuli evoked vigorous optokinetic nystagmus (OKN) under identical conditions and also when matched for apparent contrast. These findings are discussed in terms of the site of detection of second-order motion.