A model for the estimate of local image velocity by cells in the visual cortex

Some computational theories of motion perception assume that the first stage en route to this perception is the local estimate of image velocity. However, this assumption is not supported by data from the primary visual cortex. Its motion sensitive cells are not selective to velocity, but rather are directionally selective and tuned to spatio-temporal frequencies. Accordingly, physiologically based theories start with filters selective to oriented spatio-temporal frequencies. This paper shows that computational and physiological theories do not necessarily conflict, because such filters may, as a population, compute velocity locally. To prove this point, we show how to combine the outputs of a class of frequency tuned filters to detect local image velocity. Furthermore, we show that the combination of filters may simulate 'Pattern' cells in the middle temporal area (MT), whereas each filter simulates primary visual cortex cells. These simulations include three properties of the primary cortex. First, the spatio-temporal frequency tuning curves of the individual filters display approximate space-time separability. Secondly, their direction-of-motion tuning curves depend on the distribution of orientations of the components of the Fourier decomposition and speed of the stimulus. Thirdly, the filters show facilitation and suppression for responses to apparent motions in the preferred and null directions, respectively. It is suggested that the MT's role is not to solve the aperture problem, but to estimate velocities from primary cortex information. The spatial integration that accounts for motion coherence may be postponed to a later cortical stage.     

Spatio-temporal visual properties in the ascending tectofugal system

Our study compares the spatio-temporal visual receptive field properties of different subcortical stages of the ascending tectofugal visual system. Extracellular single-cell recordings were performed in the superficial (SCs) and intermediate (SCi) layers of the superior colliculus (SC), the suprageniculate nucleus (Sg) of the posterior thalamus and the caudate nucleus (CN) of halothane-anesthetized cats. Neuronal responses to drifting gratings of various spatial and temporal frequencies were recorded. The neurons of each structure responded optimally to low spatial and high temporal frequencies and displayed narrow spatial and temporal frequency tuning. The detailed statistical analysis revealed that according to its stimulus preferences the SCs has markedly different spatio-temporal properties from the homogeneous group formed by the SCi, Sg and CN. The SCs neurons preferred higher spatial and lower temporal frequencies and had broader spatial tuning than the other structures. In contrast to the SCs the visually active SCi, as well as the Sg and the CN neurons possessed consequently similar spatio-temporal preferences. These data support our hypothesis that the visually active SCi, Sg and CN neurons form a homogeneous neuronal population given a similar spatio-temporal frequency preference and a common function in processing of dynamic visual information.     

Temporal adaptability and the inverse relationship to sensitivity: a parameter identification model

Following a prolonged period of visual adaptation to a temporally modulated sinusoidal luminance pattern, the threshold contrast of a similar visual pattern is elevated. The adaptive elevation in threshold contrast is selective for spatial frequency, may saturate at low adaptor contrast, and increases as a function of the spatio-temporal frequency of the adapting signal. A model for signal extraction that is capable of explaining these threshold contrast effects of adaptation is proposed. Contrast adaptation in the model is explained by the identification of the parameters of an environmental model: the autocorrelation function of the visualized signal. The proposed model predicts that the adaptability of threshold contrast is governed by unpredicted signal variations present in the visual signal, and thus represents an internal adjustment by the visual system that takes into account these unpredicted signal variations given the additional possibility for signal corruption by additive noise.     

Stereoscopic subsystems for position in depth and for motion in depth.

We describe psychophysical evidence that the human visual system contains information-processing channels for motion in depth in addition to those for position in depth. These motion-in-depth channels include some that are selectively sensitive to the relative velocities of the left and right retinal images. We propose that the visual pathway contains stereoscopic (cyclopean) motion filters that respond to only a narrow range of the directions of motion in depth. Turning to the single-neuron level we report that, in addition to neurons turned to position to depth, cat visual cortex contains neurons that emphasize information about the direction of motion at the expense of positional information. We describe psychophysical evidence for the existence of channels that are sensitive to change size, and are separate from the channels both for motion and for flicker. These changing-size channels respond independently of whether the stimulus is a bright square on a dark ground or a dark square on a bright ground. At the physiological level we report single neurons in cat visual cortex that respond selectively to increasing or to decreasing size independently of the sign of stimulus contrast. Adaptation to a changing-size stimulus produces two separable after-effects: an illusion of changing size, and an illusion of motion in depth. These after-effects have different decay time constants. We propose a psychophysical model in which changing-size filters feed a motion-in-depth stage, and suppose that the motion-in-depth after-effect is due to activity at the motion-in-depth stage, while the changing-size after-effect is due to to activity at the changing-size and more peripheral stages. The motion-in-depth after-effect can be cancelled either by a changing-size test stimulus or by relative motion of the left and right retinal images. Opposition of these two cues can also cancel the impression of motion in depth produced by the adapting stimulus. These findings link the stereoscopic (cyclopean) motion filters and the changing-size filters: both feed the same motion-in-depth stage.     

Spatial frequency analysis in early visual processing

The existence of multiple channels, or multiple receptive field sizes, in the visual system does not commit us to any particular theory of spatial encoding in vision. However, distortions of apparent spatial frequency and width in a wide variety of conditions favour the idea that each channel carries a width- or frequency-related code or 'label' rather than a 'local sign' or positional label. When distortions of spatial frequency occur without prior adaptation (e.g. at low contrast or low luminance) they are associated with lowered sensitivity, and may be due to a mismatch between the perceptual labels and the actual tuning of the channels. A low-level representation of retinal space could be constructed from the spatial information encoded by the channels, rather than being projected intact from the retina.     

The physics of visual perception

By measuring the contrast threshold for gratings of different waveform and spatial frequency, Campbell & Robson suggested in 1968 that there may be 'channels' tuned to different spatial frequencies. By using the technique of adapting to a high contrast grating, it was possible to measure the band-pass characteristics of these channels. Similar techniques were used to establish the orientational tuning of the channels. Reasons are put forward why it is advantageous to organize the visual system in this manner.     

Independent component analysis of natural image sequences yields spatio-temporal filters similar to simple cells in primary visual cortex.

Simple cells in the primary visual cortex process incoming visual information with receptive fields localized in space and time, bandpass in spatial and temporal frequency, tuned in orientation, and commonly selective for the direction of movement. It is shown that performing independent component analysis (ICA) on video sequences of natural scenes produces results with qualitatively similar spatio-temporal properties. Whereas the independent components of video resemble moving edges or bars, the independent component filters, i.e. the analogues of receptive fields, resemble moving sinusoids windowed by steady Gaussian envelopes. Contrary to earlier ICA results on static images, which gave only filters at the finest possible spatial scale, the spatio-temporal analysis yields filters at a range of spatial and temporal scales. Filters centred at low spatial frequencies are generally tuned to faster movement than those at high spatial frequencies.     

Characteristics of detection of the general direction of movement of visual objects by preschool-age children with typical and atypical cognitive development

The specifics of functioning of two channels for processing visual information, the magnocellular and parvocellular ones, in preschool children with normal and impaired cognitive development have been studied. Children with delayed psychological development have been found to have a diminished activity in the magnocellular channel due to weakening of the mechanism responsible for assessing the temporal parameters of the visual input. At the same time, children displaying a spectrum of infantile autism syndromes complicated by a delayed psychological development and mental retardation have a pronounced deficit in information processing in both channels of the visual system, manifested in an impaired ability to assess the direction of movement in terms of both temporal and spatial characteristics of the objects. The degree of this impairment is correlated with both the severity of the particular neurological disorder and the level of speech development of the child.     

The interpretation of a moving retinal image

It is shown that from a monocular view of a rigid, textured, curved surface it is possible, in principle, to determine the gradient of the surface at any point, and the motion of the eye relative to it, from the velocity field of the changing retinal image, and its first and second spatial derivatives. The relevant equations are redundant, thus providing a test of the rigidity assumption. They involve, among other observable quantities, the components of shear of the retinal velocity field, suggesting that the visual system may possess specialized channels for computing these components.     

综述: 初级视皮层的亮度信息加工

亮度(luminance)是最基本的视觉信息.与其他视觉特征相比,由于视神经元对亮度刺激的反应较弱,并且许多神经元对均匀亮度无反应,对亮度信息编码的神经机制知之甚少.初级视皮层部分神经元对亮度的反应要慢于对比度反应,被认为是由边界对比度诱导的亮度知觉(brightness)的神经基础.我们的研究表明,初级视皮层许多神经元的亮度反应要快于对比度反应,并且这些神经元偏好低的空间频率、高的时间频率和高的运动速度,提示皮层下具有低空间频率和高运动速度通路的信息输入对产生初级视皮层神经元的亮度反应有贡献.已经知道初级视皮层神经元对空间频率反应的时间过程是从低空间频率到高空间频率,我们发现的早期亮度反应是对极低空间频率的反应,与这一时间过程是一致的,是这一从粗到细的视觉信息加工过程的第一步,揭示了处理最早的粗的视觉信息的神经基础.另外,初级视皮层含有偏好亮度下降和高运动速度的神经元,这群神经元的活动有助于在光照差的环境中检测高速运动的低亮度物体.     

Visual control of cursorial prey pursuit by tiger beetles (Cicindelidae)

Target detection poses problems for moving animals, such as tiger beetles, that track targets visually. The pursuer's movements degrade target image contrast and induce reafferent image movement that confounds continuous detection of prey. In nature, beetles pursue prey discontinuously with several iterations of stop-and-go running. The beetle's dynamics were analyzed by filming pursuits of prey or experimenter-controlled dummies. Durations of stops are inversely related to prey visual angular velocity; as the prey image moves between neighboring ommatidial fields, the beetle relocalizes it and renews running. During subsequent runs, translation and rotation depend upon prey visual angular velocity and position, respectively, seen during the previous stop. The beetle runs, then stops, while prey continues moving. After two to three iterations of stop-and-go the beetle catches its prey, suggesting open-loop control of running. Computational model simulations produce good qualitative spatio-temporal fit with actual pursuits. However, when pursuing prey dummies, beetles track continuously and quickly follow changes in target position, suggesting closed-loop control using a position-sensitive servo mechanism. Differences between these types of pursuit control system are discussed with respect to limitations in signal detection, particularly spatio-temporal contrast, that may force beetles to use an open-loop system. Accepted: 7 April 1997     

Optogenetics: a new enlightenment age for zebrafish neurobiology

Zebrafish became a model of choice for neurobiology because of the transparency of its brain and because of its amenability to genetic manipulation. In particular, at early stages of development the intact larva is an ideal system to apply optical techniques for deep imaging in the nervous system, as well as genetically encoded tools for targeting subsets of neurons and monitoring and manipulating their activity. For these applications,new genetically encoded optical tools, fluorescent sensors, and light-gated channels have been generated,creating the field of \optogenetics." It is now possible to monitor and control neuronal activity with minimal perturbation and unprecedented spatio-temporal resolution.We describe here the main achievements that have occurred in the last decade in imaging and manipulating neuronal activity in intact zebrafish larvae. We provide also examples of functional dissection of neuronal circuits achieved with the applications of these techniques in the visual and locomotor systems.     

Two spatio-temporal filters in human vision

1. We have studied visual detection of a circular target moving across a spatially and/or temporally modulated background. Illumination, I _t , for threshold detection of the target has been measured as a function of background modulation frequency and changes in I _t associated with background modulation provide a means of determining the frequency response characteristics of visual channels. 2. Temporal frequency responses obtained with temporally modulated, spatially uniform backgrounds have pass-band characteristics and the temporal frequency for peak response increases with increase in mean background illumination. These temporal frequency responses resemble those of the de Lange (1954) filter, but the latter incorporates the incremental thresholds for steady backgrounds. 3. The amplitude of this temporal response saturates at low (40%) background modulation, decreases to zero as the target velocity falls to zero, and is maximum for a circular target of diameter 2°. 4. The spatial characteristics of this temporal filter were measured with a background field consisting of alternate steady and flickering bars. The resulting spatial frequency curve peaks at 1 cycle deg^-1 for all background illuminations and is independent of the background grating orientation. This spatial response differs significantly from the IMG spatial functions observed with a background grating (Barbur and Ruddock, 1980). 5. The spatial and temporal responses reviewed above exhibit similar parametric variations and we therefore associate them with a single spatiotemporal filter, ST2. 6. A second temporal response, with low-pass frequency characteristics, was observed with a background field consisting of two matched gratings, presented in spatial and temporal antiphase. This response has parametric properties similar to those of the IMG spatial response described previously by Barbur and Ruddock (1980), thus we associated the two sets of data with a single spatio-temporal filter, ST1. 7. We show that the ST2 responses can be obtained by combining ST1 responses, and we present a network incorporating the two filters. 8. We review other psychophysical studies which imply the activity of two spatio-temporal filters with properties of the kind revealed in our studies. We argue that filter ST1 has properties equivalent to those of X-type and filter ST2 has properties equivalent to those of Y-type electrophysiological mechanisms.     

Local spectral analysis in the visual cortex

On the basis of experimental evidence presented earlier a model of local spectral analysis of the image performed by the complex receptive fields of the visual cortex has been proposed. An essential feature of the model is that the generalized piece-wise Fourier transform is performed not over the image luminance function but over the logarithm contrast function resulted from analysis of the image by the round receptive fields of the preceding levels. Such an assumption removes a number of experimental objections offered against the hypothesis of two-dimensional Fourier transform in the visual system. The consequencies from the piece-wise expansion in a series of basic functions have been considered and among them: the channel frequency characteristics which can have more than one maximum; the possibility of describing the image by a limited number of channels with overlapping frequency characteristics; the existence of mechanisms for estimation of phase shift between frequencies.     

Effect of the distance of optokinetic stimuli from the eyes on certain parameters of the optokinetic nystagmus.

The above effect was studied in 65 subjects with normal vision (mean age 20 years) in investigations in which the following factors were successively changed: distance of optokinetic stimuli from the eyes; this distance and angular velocity of stimuli; distance and frequency of stimuli or finally distance and accommodation level. The angular velocity of the pursuit nystagmus phase was found to be by far the highest and simultaneously the closest to the angular velocity of optokinetic stimuli when the latter are 1.5m from the eyes. With shorter distances, the velocity of the pursuit movements lags steadily behind that of stimulus velocity. This change is conditioned by changes in OKN amplitude since its frequency as a whole does not change. Even though the accommodation level significantly affects the velocity of the pursuit nystagmus phase, the dependence on the distance of optokinetic stimuli from the eyes persists even after atropinization. The interpretation of these findings must take into account sepcifically the demands on accommodation, convergence, and on visual attention which are increased with shorter distances.     

Seeing things in motion: models, circuits, and mechanisms

Motion vision provides essential cues for navigation and course control as well as for mate, prey, or predator detection. Consequently, neurons responding to visual motion in a direction-selective way are found in almost all species that see. However, directional information is not explicitly encoded at the level of a single photoreceptor. Rather, it has to be computed from the spatio-temporal excitation level of at least two photoreceptors. How this computation is done and how this computation is implemented in terms of neural circuitry and membrane biophysics have remained the focus of intense research over many decades. Here, we review recent progress made in this area with an emphasis on insects and the vertebrate retina.     

Visual signals in an optomotor reflex: systems and information theoretic analysis

Compensatory optomotor reflexes were examined in crayfish (Procambarus clarkii) with oscillating sine wave gratings and step displacements of a single stripe. A capacitance transducer was used to measure the rotation of the eyestalk about its longitudinal axis. System studies reveal a spatial frequency response independent of velocity and stimulus amplitude and linear contrast sensitivity similar to that of neurons in the visual pathway. The reflex operates at low temporal frequencies (<0.002 Hz to 0.5 Hz) and exhibits a low-pass temporal frequency response with cut-off frequency of 0.1 Hz. Eyestalk rotation increases as a saturable function of the angular stimulus displacement. When compared to the oscillatory response, transient responses are faster, and they exhibit a lower gain for large stimulus displacements. These differences may reflect system nonlinearity and/or the presence of at least two classes of afferents in the visual pathway. Our metric for information transmission is the Kullback-Leibler (K-L) distance, which is inversely proportional to the probability of an error in distinguishing two stimuli. K-L distances are related to differences in responsiveness for variations in spatial frequency, contrast, and angular displacement. The results are interpreted in terms of the neural filters that shape the system response and the constraints that the K-L distances place on information transmission in the afferent visual pathway.     

Latency of visually evoked saccadic eye movements

The paper deals with the initiation of visually guided saccades, in order to break down the saccadic reaction time into functionally different periods of time. It takes into account that spatial processing of information is so basic that modelling of saccadic control properties should include spatio-temporal arrangements. The output signal of the saccadic system was measured in response to visual stimuli in which the time between the appearance of a visual stimulus in the peripheral field and the disappearance of the central fixation point was varied. The variation of the mean saccadic latency time, measured with respect to the onset of the peripheral stimulus, as a function of stimulus asynchrony was highly significant. This variation may be represented by a so-called gap-overlap curve, which is characterized here by means of five parameters. A facilitation model is introduced to fit the results of the gap-overlap experiments. The facilitation model for the initiation of visually evoked saccades incorporates a mechanism which governs the efficiency of processing of signals that arise from a stimulus presented at a particular position in space. It explains how visual information may be affected by other sensory information before it is used to command further saccades. It allows determination of saccadic system parameters, such as the peripheral and the foveal afferent processing time, the central processing time for a saccade and the degree of facilitation. These quantities were found to be characteristic for the given test subjects, and where these data could be compared with neurophysiological data, the agreement was within the experimental error.