Receptive field structure of neurons in monkey primary visual cortex revealed by stimulation with natural image sequences

Probing the visual system with the ensemble of signals that occur in the natural environment may reveal aspects of processing that are not evident in the neural responses to artificial stimulus sets, such as conventional bars and sinusoidal gratings. However, unsolved is the question of how to use complex natural stimulation, many aspects of which the experimenter cannot completely specify, to study neural processing. Here a method is presented to investigate the structure of a neuron's receptive field based on its response to movie clips and other stimulus ensembles. As a particular case, the technique provides an estimate of the conventional first-order receptive field of a neuron, similar to what can be obtained with other reverse-correlation schemes. This is demonstrated experimentally and with computer simulations. Our analysis also revealed that the receptive fields of both simple and complex cells had regions where image boundaries, independent of their contrast sign, would enhance or suppress the cell's response. In some cases, these signals were tuned for the orientation of the boundary. This demonstrates for the first time that it might be feasible to investigate the receptive field structure of visual neurons from their responses to natural image sequences.

Contrast-dependent changes in spatial frequency tuning of macaque V1 neurons: effects of a changing receptive field size

Previous studies on single neurons in primary visual cortex have reported that selectivity for orientation and spatial frequency tuning do not change with stimulus contrast. The prevailing hypothesis is that contrast scales the response magnitude but does not differentially affect particular stimuli. Models where responses are normalized over contrast to maintain constant tuning for parameters such as orientation and spatial frequency have been proposed to explain these results. However, our results indicate that a fundamental property of receptive field organization, spatial summation, is not contrast invariant. We examined the spatial frequency tuning of cells that show contrast-dependent changes in spatial summation and have found that spatial frequency selectivity also depends on stimulus contrast. These results indicate that contrast changes in the spatial frequency tuning curves result from spatial reorganization of the receptive field.

Suppression of neural responses to nonoptimal stimuli correlates with tuning selectivity in macaque V1

Neural responses in primary visual cortex (area V1) are selective for the orientation and spatial frequency of luminance-modulated sinusoidal gratings. Selectivity could arise from enhancement of the cell's response by preferred stimuli, suppression by nonoptimal stimuli, or both. Here, we report that the majority of V1 neurons do not only elevate their activity in response to preferred stimuli, but their firing rates are also suppressed by nonoptimal stimuli. The magnitude of suppression is similar to that of enhancement. There is a tendency for net response suppression to peak at orientations near orthogonal to the optimal for the cell, but cases where suppression peaks at oblique orientations are observed as well. Interestingly, selectivity and suppression correlate in V1: orientation and spatial frequency selectivity are higher for neurons that are suppressed by nonoptimal stimuli than for cells that are not. This finding is consistent with the idea that suppression plays an important role in the generation of sharp cortical selectivity. We show that nonlinear suppression is required to account for the data. However, the precise structure of the neural circuitry generating the suppressive signal remains unresolved. Our results are consistent with both feedback and (nonlinear) feed-forward inhibition.

Pursuit eye movements to second-order motion targets

We studied smooth-pursuit eye movements elicited by first- and second-order motion stimuli. Stimuli were random dot fields whose contrast was modulated by a Gaussian window with a space constant of 0.5 degrees. For the first-order stimuli, the random dots simply moved across the screen at the same speed as the window; for the second-order stimuli the window moved across stationary or randomly flickering dots. Additional stimuli which combined first- and second-order motion cues were used to determine the degree and type of interaction found between the two types of motion stimuli. Measurements were made at slow (1 degrees/s) and moderate (6 degrees/s) target speeds. At a velocity of 1 degrees/s the initiation, transition, and steady-state phases of smooth pursuit in response to second-order motion targets are severely affected when compared with the smooth pursuit of first-order motion targets. At a velocity of 6 degrees/s there is a small but significant deficit in steady-state pursuit of second-order motion targets but not much effect on pursuit initiation.

Visual spatial characterization of macaque V1 neurons

This study characterizes the spatial organization of excitation and inhibition that influences the visual responses of neurons in macaque monkey's primary visual cortex (V1). To understand the spatial extent of excitatory and inhibitory influences on V1 neurons, we performed area-summation experiments with suprathreshold contrast stimulation. The extent of spatial summation and the magnitude of surround suppression were estimated quantitatively by analyzing the spatial summation experiments with a difference of Gaussians (DOG) model. The average extent of spatial summation is approximately the same across layers except for layer 6 cells, which tend to sum more extensively than cells in the other layers. On average, the extent of length and width summation is approximately equal. Across the population, surround suppression is greatest in layer 4B and weakest in layer 6. Estimates of summation and suppression are compared for the DOG (subtractive) model and a normalization (divisive) model. The two models yield quantitatively similar estimates of the extent of excitation and inhibition. However, the normalization (divisive) model predicts weaker surround strength than the DOG model.

Velocity constancy in a virtual reality environment

During everyday life the brain is continuously integrating multiple perceptual cues in order to allow us to make decisions and to guide our actions. In this study we have used a simulated (virtual reality--VR) visual environment to investigate how cues to speed judgments are integrated. There are two sources that could be used to provide signals for velocity constancy: temporal-frequency or distance cues. However, evidence from most psychophysical studies favours temporal-frequency cues. Here we report that two depth cues that provide a relative object--object distance--disparity and motion parallax--can provide a significant input to velocity-constancy judgments, particularly when combined. This result indicates that the second mechanism can also play a significant role in generating velocity constancy. Furthermore, we show that cognitive factors, such as familiar size, can influence the perception of object speed. The results suggest that both low-level cues to spatiotemporal structure and depth, and high-level cues, such as object familiarity, are integrated by the brain during velocity estimation in real-world viewing.

The spatial transformation of color in the primary visual cortex of the macaque monkey

Perceptually, color is used to discriminate objects by hue and to identify color boundaries. The primate retina and the lateral geniculate nucleus (LGN) have cell populations sensitive to color modulation, but the role of the primary visual cortex (V1) in color signal processing is uncertain. We re-evaluated color processing in V1 by studying single-neuron responses to luminance and to equiluminant color patterns equated for cone contrast. Many neurons respond robustly to both equiluminant color and luminance modulation (color-luminance cells). Also, there are neurons that prefer luminance (luminance cells), and a few neurons that prefer color (color cells). Surprisingly, most color-luminance cells are spatial-frequency tuned, with approximately equal selectivity for chromatic and achromatic patterns. Therefore, V1 retains the color sensitivity provided by the LGN, and adds spatial selectivity for color boundaries.

Contrast's effect on spatial summation by macaque V1 neurons

Stimulation outside the receptive field of a primary visual cortical (V1) neuron reveals intracortical neural interactions. However, previous investigators implicitly or explicitly considered the extent of cortical spatial summation and, therefore, the size of the classical receptive field to be fixed and independent of stimulus characteristics or of surrounding context. On the contrary, we found that the extent of spatial summation in macaque V1 neurons depended on contrast, and was on average 2.3-fold greater at low contrast. This adaptive increase in spatial summation at low contrast was seen in cells throughout V1 and was independent of surround inhibition.

Sensitivity of LGN Neurons in Infant Macaque Monkeys

To understand the neuronal factors limiting visual sensitivity in infant primates, we studied the responses of neurons recorded in parts of the LGN representing the central visual fields in paralysed, opiate-anesthetised 1-week-old and 4-weeks-old macaque monkeys; comparison data were taken from animals older than 6 months. We tested each neuron with achromatic sinusoidal gratings varying in spatial and temporal frequency and contrast, and we also studied the effects of added spatiotemporal white noise.

In agreement with earlier reports, we found that neurons in the infant monkeys had relatively poor spatial resolution; sensitivity to high temporal frequencies was also lower than in adults. When tested with gratings of near-optimal spatiotemporal frequency, however, most LGN neurons in the infant monkeys gave brisk and reliable visual responses that were qualitatively similar to those seen in older animals. Spontaneous and evoked response rates and contrast gain were modestly lower in the infants, but response variability was also lower, and therefore statistical measures of sensitivity and susceptibility to masking noise showed little difference between infant and adult neurons. Especially, in the 1-week-old animals, a substantial fraction of neurons lacked spontaneous activity. This resulted in ‘hard’ contrast thresholds not seen in adult animals. As a consequence, masking noise often paradoxically enhanced visual responses in these animals by subthreshold summation, a result not seen in the adults.

The maturation of visual responses in macaque LGN consists largely of changes in spatial and temporal scale, accompanied by modest changes in responsiveness and little or no change in sensitivity.

Dynamics of orientation tuning in macaque primary visual cortex

Orientation tuning of neurons is one of the chief emergent characteristics of the primary visual cortex, V1. Neurons of the lateral geniculate nucleus, which comprise the thalamic input to V1, are not orientation-tuned, but the majority of V1 neurons are quite selective. How orientation tuning arises within V1 is still controversial. To study this problem, we measured how the orientation tuning of neurons evolves with time using a new method: reverse correlation in the orientation domain. Orientation tuning develops after a delay of 30-45 milliseconds and persists for 40-85 ms. Neurons in layers 4C alpha or 4C beta, which receive direct input from the thalamus, show a single orientation preference which remains unchanged throughout the response period. In contrast, the preferred orientations of output layer neurons (in layers 2, 3, 4B, 5 or 6) usually change with time, and in many cases the orientation tuning may have more than one peak. This difference in dynamics is accompanied by a change in the sharpness of orientation tuning; cells in the input layers are more broadly tuned than cells in the output layers. Many of these observed properties of output layer neurons cannot be explained by simple feedforward models, whereas they arise naturally in feedback networks. Our results indicate that V1 is more than a bank of static oriented filters; the dynamics of output layer cells appear to be shaped by intracortical feedback.

Development of contrast sensitivity and temporal-frequency selectivity in primate lateral geniculate nucleus

We studied the development of spatial contrast-sensitivity and temporal-frequency selectivity for neurons in the monkey lateral geniculate nucleus. During postnatal week 1, the spatial properties of P-cells and M-cells are hardly distinguishable, with low contrast-sensitivity, sluggish responses, and poor spatial resolution. The acuity of P-cells improves progressively until at least 8 months, but there is no obvious increase in their maximum contrast-sensitivity with age. The contrast sensitivity of M-cells is already clearly higher than that of P-cells by 2 months, and at 8 months of age this characteristic difference between M- and P-cells approaches the adult pattern. There is a major increase in responsiveness during the first 2 postnatal months, especially for M-cells, the peak firing rate of which rises fivefold, on average, between birth and 2 months. Many P-cells in the neonatal and 2-month-old animals did not give statistically reliable responses to achromatic gratings, even at the highest contrasts: this unresponsiveness of P-cells might result from low gain and/or chromatic opponency. The upper limit of temporal resolution in the neonate is low--about one-third of that in the adult. Among M-cells, the improvement in temporal resolution, like that in contrast sensitivity, is rapid over the first 2 months, followed by a slower change approaching the adult value by 8 months of age. The development of contrast sensitivity, responsiveness and temporal tuning are little affected, if at all, by binocular deprivation of pattern vision from birth for even a prolonged period.

Interaction of motion and color in the visual pathways

In recent years the idea of parallel and independent processing streams for different visual attributes has become a guiding principle for linking the organization, architecture and function of the visual system. Findings concerning the segregation of motion and color information have been at the forefront of the evidence in favor of the parallel processing scheme. A number of studies have shown that motion perception is impaired for isoluminant stimuli, which are thought to isolate the color system. However, there are now many studies, the results of which are incompatible with the simple idea of segregated pathways. We propose two processing streams for motion that differ mostly in their temporal characteristics. Although neither of the two motion streams is color-blind, as was originally suggested, they differ radically in the way they process color information. The view that we propose provides a framework that reconciles a number of seemingly contradictory results. Evidence to support the new framework comes from psychophysical, physiological and lesion studies.

Binocular eye movements caused by the perception of three-dimensional structure from motion

We report that the perception of three-dimensional structure from monocular two-dimensional images changing over time--the kinetic depth effect (KDE)--can evoke binocular eye movements consistent with a three-dimensional percept. We used a monocular KDE stimulus that induced a vivid perception of a rigid three-dimensional sphere rotating in space. The gaze directions of both eyes were measured while observers pursued the motion of a patch on the surface of the perceived sphere as it went through a complete revolution. We found that the eyes converged when the patch was perceived on the front surface of the KDE sphere and diverged when the patch was perceived in the back. The pattern, magnitude and dynamics of binocular eye movements observed in the KDE experiment resembled those obtained when subjects viewed binocularly a light-emitting diode (LED) rotating in space and to the responses obtained with a dynamic stereogram simulating a rotating random dot sphere. Thus, the perception of three-dimensional structure from motion, stereopsis, or motion and stereopsis combined, were effective in guiding binocular eye movements.

Temporal-frequency selectivity in monkey visual cortex

We investigated the dynamics of neurons in the striate cortex (V1) and the lateral geniculate nucleus (LGN) to study the transformation in temporal-frequency tuning between the LGN and V1. Furthermore, we compared the temporal-frequency tuning of simple with that of complex cells and direction-selective cells with nondirection-selective cells, in order to determine whether there are significant differences in temporal-frequency tuning among distinct functional classes of cells within V1. In addition, we compared the cells in the primary input layers of V1 (4a, 4c alpha, and 4c beta) with cells in the layers that are predominantly second and higher order (2, 3, 4b, 5, and 6). We measured temporal-frequency responses to drifting sinusoidal gratings. For LGN neurons and simple cells, we used the amplitude and phase of the fundamental response. For complex cells, the elevation of impulse rate (F0) to a drifting grating was the response measure. There is significant low-pass filtering between the LGN and the input layers of V1 accompanied by a small, 3-ms increase in visual delay. There is further low-pass filtering between V1 input layers and the second- and higher-order neurons in V1. This results in an average decrease in high cutoff temporal-frequency between the LGN and V1 output layers of about 20 Hz and an increase in average visual latency of about 12-14 ms. One of the most salient results is the increased diversity of the dynamic properties seen in V1 when compared to the cells of the lateral geniculate, possibly reflecting specialization of function among cells in V1. Simple and complex cells had distributions of temporal-frequency tuning properties that were similar to each other. Direction-selective and nondirection-selective cells had similar preferred and high cutoff temporal frequencies, but direction-selective cells were almost exclusively band-pass while nondirection-selective cells distributed equally between band-pass and low-pass categories. Integration time, a measure of visual delay, was about 10 ms longer for V1 than LGN. In V1 there was a relatively broad distribution of integration times from 40-80 ms for simple cells and 60-100 ms for complex cells while in the LGN the distribution was narrower.

Perceived velocity of luminance, chromatic and non-fourier stimuli: influence of contrast and temporal frequency

We measured perceived velocity as a function of contrast for luminance and isoluminant sinusoidal gratings, luminance and isoluminant plaids, and second-order, amplitude-modulated, drift-balanced stimuli. For all types of stimuli perceived velocity was contrast-invariant for fast moving patterns at or above 4 deg/sec. For slowly moving stimuli the log of perceived velocity was a linear function of the log of the contrast. The slope of this perceived velocity-vs-contrast line (velocity gain) was relatively shallow for luminance gratings and luminance plaids, but was steep for isoluminant gratings and isoluminant plaids, as well as for drift-balanced stimuli. Independent variation of spatial and temporal frequency showed that these variables, and not velocity alone, determine the velocity gain. Overall, the results indicate that slow moving stimuli defined by chromaticity or by second-order statistics are processed in a different manner from luminance defined stimuli. We propose that there are a number of independent mechanisms processing motion targets and it is the interplay of these mechanisms that is responsible for the final percept.

Temporal and chromatic properties of motion mechanisms

We measured threshold contours in color space for detecting drifting sinusoidal gratings over a range of temporal frequencies, and for identifying their direction of motion. Observers were able to correctly identify the direction of motion in all directions of color space, given a sufficiently high contrast. At low temporal frequencies we found differences between luminance and isoluminance conditions; for isoluminance there was a marked threshold elevation for identification when compared to detection. The threshold elevation for identification is dependent on eccentricity as well as on temporal frequency. At high temporal frequencies there were no differences between detection and identification thresholds, or between thresholds for luminance and isoluminance. A quantitative analysis of the threshold contours allowed us to identify two mechanisms contributing to motion: a color-opponent mechanism with a high sensitivity at low temporal frequencies and a luminance mechanism whose relative sensitivity increases with temporal frequency. An analysis of the cone contributions to motion detection and identification showed that L-cones dominated threshold behavior for both detection and identification at high temporal frequencies. There was a weak S-cone input to motion detection and identification at high temporal frequencies.

Contrast dependence of colour and luminance motion mechanisms in human vision

Conventional views of visual perception propose a colour-blind pathway conveying motion information and a motion-blind pathway carrying colour information. Recent studies show that motion perception is not always colour blind, is partially dependent on attention, can show considerable perceptual slowing around isoluminance and is contrast-dependent. If there is a single motion pathway, receiving luminance and chromatic input, then the dependence of relative perceived velocity on relative stimulus contrast should be the same for both luminance and chromatic targets. Here we provide a distinctive characterization of the motion mechanisms using a robust velocity-matching task. A relative contrast scale allows direct comparison of the performance with luminance and chromatic targets. The results show that the perceived speed of slowly moving coloured targets at isoluminance has a steep contrast dependence. The perceived speed of slowly moving luminance targets shows a much lower contrast dependence. At high speeds the contrast dependence is low for both luminance and isoluminant stimuli, although the behaviour is unlike either of the slow mechanisms. The results suggest two independent pathways that perceive slowly moving targets: one is luminance-sensitive and the other is colour-sensitive. Fast movement is signalled via a single motion pathway that is contrast-invariant and not colour blind.

Macaque V1 neurons can signal 'illusory' contours

We describe here a new view of primary visual cortex (V1) based on measurements of neural responses in V1 to patterns called 'illusory contours' (Fig. 1a, b). Detection of an object's boundary contours is a fundamental visual task. Boundary contours are defined by discontinuities not only in luminance and colour, but also in texture, disparity and motion. Two theoretical approaches can account for illusory contour perception. The cognitive approach emphasizes top-down processes. An alternative emphasizes bottom-up processing. This latter view is supported by (1) stimulus constraints for illusory contour perception and (2) the discovery by von der Heydt and Peterhans of neurons in extrastriate visual area V2 (but not in V1) of macaque monkeys that respond to illusory contours. Using stimuli different from those used previously, we found illusory contour responses in about half the neurons studied in V1 of macaque monkeys. Therefore, there are neurons as early as V1 with the computational power to detect illusory contours and to help distinguish figure from ground.

Local circuit neurons of macaque monkey striate cortex: II. Neurons of laminae 5B and 6

This study investigates the intrinsic organization of axons and dendrites of aspinous, local circuit neurons of the macaque monkey visual striate cortex. These investigations use Golgi Rapid preparations of cortical tissue from monkey aged 3 weeks postnatal to adult. We have earlier (Lund, '87) described local circuit neurons found within laminae 5A and 4C; this present account is of neurons found in the infragranular laminae 5B and 6. Since the majority of such neurons are GABAergic and therefore believed to be inhibitory, their role in laminae 5B and 6, the principal sources of efferent projections to subcortical regions, is of considerable importance. We find laminae 5B and 6 to have in common at least one general class of local circuit neuron-the "basket" neuron. However, a major difference is seen in the axonal projections to the superficial layers made by these and other local circuit neurons in the two laminae; lamina 5B has local circuit neurons with principal rising axon projections to lamina 2/3A, areas whereas lamina 6 has local circuit neurons with principal rising axon projections to divisions of 4C, 4A, and 3B. These local circuit neuron axon projections mimic the different patterns of apical dendritic and recurrent axon projections of pyramidal neurons lying within laminae 5B and 6, which are linked together by both dendritic and axonal arbors of local circuit neurons in their neuropils extending between the two laminae. The border zone between 5B and 6 is a specialized region with its own variety of horizontally oriented local circuit neurons, and it also serves as a special focus for pericellular axon arrays from a particular variety of local circuit neuron lying within lamina 6. These pericellular axon "baskets" surround the somata and initial dendritic segments of the largest pyramidal neurons of layer 6, which are known to project both to cortical area MT (V5) and to the superior colliculus (Fries et al., '85). Many of the local circuit neurons of layer 5B send axon trunks into the white matter, and we therefore, suspect them of providing efferent projections. The axons of lamina 6 local circuit neurons have not been found to make such clear-cut contributions to the white matter.

Laminar organization and contrast sensitivity of direction-selective cells in the striate cortex of the Old World monkey

The directional preference of neurons sampled from all layers of the striate cortex was determined using the responses to drifting grating stimuli of optimal spatial and temporal frequency. In addition, contrast sensitivity as a function of spatial frequency was measured and from the resulting spatial contrast sensitivity function the peak contrast sensitivity and optimal spatial frequency were obtained. The distribution of directionally selective cells showed a distinct laminar pattern. Upper layer 4 (4a, 4b, and 4c alpha) and layer 6 were the only cortical layers with neurons that showed a pronounced preference for the direction of stimulus motion. The directionally selective cells in these layers are among those with the highest contrast sensitivities but had optimal spatial frequencies that were confined to the low and middle range of the optimal spatial frequency distribution. These findings suggest that the directionally selective cells may fall into at least 2 distinct populations, which may be the first stages in the visual pathway that correspond to those channels, inferred from psychophysical experiments, that underlie the detection of movement.

Vibratory detection thresholds following a digital nerve lesion

Vibratory detection thresholds were measured at a number of frequencies between 5 and 320 Hz following a lesion of the lateral digital nerve innervating the terminal phalanx of the left index finger. Thresholds measurements began approximately 4 weeks after the nerve was repaired. A staircase method was used to determine thresholds on both the injured fingerpad and the intact fingerpad of the opposite hand. There was a large increase in thresholds on the injured fingerpad in the lower range of frequencies (5-40 Hz) while at higher frequencies (80-250 Hz) there was no significant difference between the thresholds on the injured fingerpad and those on the intact fingerpad. It is suggested that the differential effect of the nerve lesion on vibratory thresholds reflects the spread of the vibratory stimulus through the skin and the spatial characteristics of functionally intact receptor/afferent groups innervating neighbouring skin.

Spatial properties of neurons in the monkey striate cortex

Contrast sensitivity as a function of spatial frequency was determined for 138 neurons in the foveal region of primate striate cortex. The accuracy of three models in describing these functions was assessed by the method of least squares. Models based on difference-of-Gaussians (DOG) functions where shown to be superior to those based on the Gabor function or the second differential of a Gaussian. In the most general case of the DOG models, each subregion of a simple cell's receptive field was constructed from a single DOG function. All the models are compatible with the classical observation that the receptive fields of simple cells are made up of spatially discrete 'on' and 'off' regions. Although the DOG-based models have more free parameters, they can account better for the variety of shapes of spatial contrast sensitivity functions observed in cortical cells and, unlike other models, they provide a detailed description of the organization of subregions of the receptive field that is consistent with the physiological constraints imposed by earlier stages in the visual pathway. Despite the fact that the DOG-based models have spatially discrete components, the resulting amplitude spectra in the frequency domain describe complex cells just as well as simple cells. The superiority of the DOG-based models as a primary spatial filter is discussed in relation to popular models of visual processing that use the Gabor function or the second differential of a Gaussian.

The postnatal reduction of the uncrossed projection from the nasal retina in the cat

We have investigated the postnatal reduction of the uncrossed projection from the nasal retina in the cat by injecting horseradish peroxidase into one optic tract of kittens and cats and retrogradely labeling the cells in the ipsilateral retina that have an uncrossed projection to the brain. The newborn kitten has over 600 uncrossed cells in the nasal retina. The number is reduced to about one-quarter of that value by postnatal day 10. The two adult cats examined had 75 and 100 of these ipsilaterally projecting nasal cells. They are distributed all across the nasal retina, and most have the morphology characteristic of gamma cells. A lesion in one optic tract in the newborn kitten results in an increase in the number of cells from the nasal retina with an ipsilateral projection at maturity. There are more of these cells in the region that has been depleted of ganglion cells by the lesion. This excess consists mostly of gamma and epsilon cells. These findings indicate that competitive factors play a role in the elimination of inappropriate ganglion cell projections in the cat, and that this process contributes to the precision of the nasotemporal division of the retina.