[STUDYING THE INTERACTION OF RHESUS MONKEYS WITH THE TACTILE MONITOR DURING THEIR OBSERVATION OF LOW-FREQUENCY VISUAL TEST IMAGES].

In behavioral experiments rhesus macaque monkeys were trained to interact with the computer using a tactile display. We used grayscale Gabor patches of low spatial frequency as stimuli. Monkeys' task was to touch the screen with his hand in the area of the target stimulus, followed by automatic food or juice reinforcement. After two successive correct answers, stimulus contrast gradually decreased. Using a two-alternative forced choice method the contrast threshold was measured within which monkeys can detect the appearance of low-frequency images. It was shown that the contrast sensitivity decreased with the decrease of stimulus spatial frequency, while the reaction time increased. The findings extend our knowledge of the primates' activity in the virtual environment and open new possibilities for modeling and studying various human diseases.

[Latency of Evoked Potentials in the Tasks Involving Classification of Images after Wavelet Filtration.]

We studied the characteristics of evoked potentials recorded during the recognition test based on four types of series of images subjected to the wavelet filtration: images of living objects containing either low frequency or high frequency portion of the spatial frequency spectrum, and imaging of non-living objects in the same two spatial frequency bands. Each subject had to classify the image either by its semantic feature (live-not live), or by its physical feature (low-high frequency). The purpose of this study was to compare the time characteristics of evoked potentials in these two types of tasks, which provides information on the time characteristics of categorization mechanisms of visual images. Analysis of the latent periods and amplitudes of the components of evoked potentials enabled to detect the occipital areas of the leads where the early components (up to 170 ms) are associated with spatial and frequency characteristics of the image, the frontal and temporal areas where the components of 170-200 ms correspond with the process of categorization, and the later frontal, central and parietal areas (300-500 ms) correspond with the process of error detection and the organization of motor response.

[Psychophysiological analysis of the reading and text understanding].

Presented cognitive studies aimed to identifying the individual properties of macro saccades, one of the most important oculomotor functions during text, the fairy tales reading. We measure, in the form of abstracts, the reduced analogue of the "school essays", the capabilities of our subjects, the students of liberal arts colleges, to understand the presented text (morality of tales). Evaluation of this abstracts was conducted by independent experts using school 6 point scales. The preparing of the abstract wasn't limited in time. We found that there is not any relationship between reading speed and the quality of the interpretation of the text. No relationship of understanding the hidden morality with the number of characters being picked for the fixation and the mean duration of the macro saccades. Correlation analysis showed an inverse relationship between the number of returns and the understanding of the text, more macro saccades returns--the worse was the understanding of the text. There is an inverse relationship between the number of exciting characters and a number of macro saccades. Number of fixations (capture) increases the reading of the text. This means that the bandwidth of visual channel per time unit is determined by the number of captured characters. Less number of captured during fixation of signs--more time reading and less bandwidth of the eyes per time unit.

Studies of the perception of incomplete outline images of different sizes.

The aim of the present work was to assess the range of angular sizes of fragmented images of objects at which perception of the images was scale-independent. Measurements were made of human subjects' recognition thresholds for the shapes of the objects over a wide range of angular sizes (0.19-50 degrees). The experiments used the Gollin test--a method for studying the recognition of fragmented outline images of objects with which the observer is familiar. The results obtained demonstrated that there is a wide range of angular sizes, from 1.0 degrees to 50 degrees, over which the perception thresholds of incomplete outline images do not change with changes in size, along with a narrow range of stimulus sizes, 0.19-1.0 degrees, over which there is a significant size dependence. We suggest that the increase in thresholds and the failure to recognize images of small size occur as a result of an increased contribution of sampling noise at the level of the human retina.

Analysis of the studies of the perception of fragmented images: global description and perception using local features.

This report presents an analysis of studies of the visual perception of fragmented images. Two approaches to the description of brain functioning during the solution of this task are considered: studies of the perception of the statistical properties of global whole images and studies of the perception of images in terms of local higher-order informative features. These approaches describe two different mechanisms which the human brain may use for recognizing incomplete images. Studies performed using the Gollin test (measurement of recognition thresholds for fragmented outline images during gradual construction of the outline) are given most attention. In solving this task, the visual system extracts the statistical properties of the whole image. Local higher-order informative features are used by the brain as additional sources of information about the image. The role of this source increases on learning a given alphabet of stimuli. In accordance with a matched filtration model, the fragmented images used in the Gollin test are perceived as a whole structure, compared with a reference which is stored in memory and selected using the selective attention mechanism. At the primary filtration step and the matched filtration step, the recognition thresholds of images in the Gollin test reflect the processes of extracting the signal from noise. The Gollin test can be used as a differential tool for the diagnosis of different types of cognitive impairments.

The human magno and parvo systems and selective impairments of their functions.

The magno and parvo systems, consisting of large and small cells respectively and showing the greatest differences in temporospatial characteristics, are two of the most interesting channels in the organization of retinotopic input pathways of the human visual system. We tested the hypothesis that selective lesions of fibers of a specified diameter occurring in pathology of the conducting pathways, i.e., multiple sclerosis, could be used as a "key" for studies of the different functional properties of the magno and parvo systems. Measurements were made of spatial-frequency contrast sensitivity in healthy subjects and patients in the early stages of multiple sclerosis. Decreased sensitivity at low and intermediate or at high spatial frequencies provided indirect reflections of changes in the activities of the magno and parvo systems. These psychophysical data are compared with electrophysiological results. The amplitudes of the early components of visual evoked potentials reflect activation of the two systems in healthy people and patients with multiple sclerosis. We identified two groups of patients, with statistically significantly dominant lesions of the magno and parvo systems respectively.

Localization of human cortical areas activated on perception of ordered and chaotic images.

The aims of this study were to identify the locations of areas in the human cortex responsible for describing fragmented test images of different degrees of ordering and to identify the areas taking decisions regarding stimuli of this type. The locations of higher visual functions were determined by functional magnetic resonance imaging (fMRI) using a scanner fitted with a superconducting magnet and a field strength of 1.5 T. The blood oxygen level-dependent (BOLD) method was based on measurements of the level of hemoglobin oxygenation in the blood supplied to the brain. This level was taken to be proportional to the extent of neuron activation in the corresponding part of the gray matter. Stimuli were matrixes consisting of Gabor elements of different orientations. The measure of matrix ordering was the ratio of the number of Gabor elements with identical orientations to the total number of elements in the image. Brain neurons were activated by simultaneous changes in the orientations of all the elements, leading to substitution of one matrix by another. Substitution of the orientation was perceived by observers as rotation of the elements in the matrix. Stimulation by matrixes with a high level of ordering was found to activate the occipital areas of the cortex, V1 and V2 (BA17-BA18), while presentation of matrixes with random element orientations also activated the parietal-temporal cortex, V3, V4, V5 (BA19), and the parietal area (BA7). Brain zones responsible for taking decisions regarding the level of order or chaos in the organization of the stimuli are located in different but close areas of the prefrontal and frontal cortex of the brain, including BA6, BA9, and BA10. The results are assessed in terms of concepts of the roles and interactions of different areas of the human brain during recognition of fragmented images of different degrees of complexity.

Electrophysiological studies of texture recognition mechanisms.

We report here our electrophysiological and psychophysiological studies of the mechanisms by which the visual system recognizes structured images with different levels of ordering. Visual stimuli consisted of textures, i.e., a set of matrixes consisting of Gabor grids. Matrixes differed in terms of the degree of ordering resulting from changes in the probability that grids with the same orientation would appear. The subject's task was to identify the dominant orientation in the stimulus. The relationship between response accuracy, reaction time, and the main characteristics of evoked potentials on the one hand, and the number of identical grids in the matrix on the other was identified. The proportion of correct responses increased and the reaction time decreased as the degree of ordering of stimuli increased. Visual evoked potentials recorded in the occipital areas showed a relationship between the amplitudes of the N2, P2, and P3 waves, with latent periods of 180, 260, and 400 msec, respectively, and matrix parameters. The amplitudes of the P3 component and the positive component recorded in the frontal leads, with a latent period of 250 msec, increased gradually as the task became simpler. The amplitude of the N2 wave also increased with increases in the number of identically oriented elements in the matrix, though this relationship was S-shaped. The magnitude of the P2 component, conversely, was maximal in response to presentation of those matrixes which were most complex to recognize and gradually decreased as the content of identically oriented grids in the matrix increased. These relationships were compared with the statistical characteristics of the stimuli and assessed in terms of the view that the visual system contains two mechanisms, i.e., local and integral image descriptions.

Mechanisms of recognition of the outlines of "vanishing" optotypes.

The aim of the present work was to study the interaction between the optical properties of images of " disappearing" optotypes and their recognition thresholds. The "disappearing" optotypes were figures with complex outlines and had a unique property - they were close to the threshold of recognition and observation, which increases the accuracy of measurement of visual acuity and the subjects' attention to them. The recognition distances of "disappearing" optotypes were measured. A relationship was found between the recognition distance of "disappearing" optotypes and different optical density profiles on the one hand and the spatial and spatial frequency characteristics of the stimuli on the other. The decisive factor determining the threshold of recognition of optotypes in spatial frequency terms is its spatial frequency spectrum; that in spatial terms is the width of the black/white pair or black-white triad in the complex outline. Regardless of the shape of the optotype, one of the most important limiting factors was the concordance of this test with the scattering function of the subject's eye optics.

Objective measurement of human visual acuity by visual evoked potentials.

Electrophysiological studies were performed to measure the threshold (upper end of range) spatial frequency using visual evoked potentials and comparison with visual acuity neuron 26 healthy subjects. The aim of the present work was to create a method for objective measurement of visual acuity. This was addressed by initial measurements using a universally accepted method of visual stimulation and processing of electroencephalograms, which allows errors due to individual differences in visual system function to be minimized. These experiments yielded a strong correlation between the threshold spatial frequency of the test grid yielding an evoked potential on presentation and visual acuity, in degrees, expressed as the resolving ability of the visual system for this optotype. A logarithmic relationship was found between these values and an equation allowing automated calculation of visual acuity (resolving ability) from electrophysiological data was derived. The results were independent of the subject's responses and therefore provides a maximally objective assessment of visual acuity.

The threshold signal:noise ratio in the perception of fragmented figures.

Perception thresholds were measured for fragmented outline figures (the Gollin test). A new approach to the question of the perception of incomplete images was developed. In this approach, figure fragmentation consisted of masking with multiplicative texture-like noise--this interference was termed "invisible" masking. The first series of studies established that the "similarity" between the amplitude-frequency spectra of test figures and "invisible" masks, expressed as a linear correlation coefficient, had significant effects on the recognition thresholds of these figures. The second series of experiments showed that progressing formation of the figures was accompanied by increases in the correlation between their spatial-frequency characteristics and the corresponding characteristics of the incomplete figure, while the correlation with the "invisible" mask decreased. It is suggested that the ratio of the correlation coefficients, characterizing the "similarity" of the fragmented figure with the intact figure and the "invisible" mask, corresponds to the signal:noise ratio. The psychophysical recognition threshold for figures for naive subjects not familiar with the test image alphabet was reached after the particular level of fragmentation at which this ratio was unity.

The optical-geometrical characteristics and thresholds of perception of fragmented outline figures.

Measurements were made of the threshold of recognition of cumulatively forming line figures. The threshold value of the outline, expressed in pixels, depended on the length of the outline of the whole unfragmented figure. Relative threshold values were constant, and for the measures of figure fragments used in the present study, averaged 12.5%. A spatial frequency analysis of the test images was performed. Variation of the amplitude-frequency parameters of the spectra of the images of various figures with threshold fragmentation was minimal as compared with the variation of these parameters in figures with subthreshold or suprathreshold levels of fragmentation.

Resolving ability and image discretization in the visual system.

Psychophysiological studies were performed to measure the spatial threshold for resolution of two "points" and the thresholds for discriminating their orientations depending on the distance between the two points. Data were compared with the scattering of the "point" by the eye's optics, the packing density of cones in the fovea, and the characteristics of the receptive fields of ganglion cells in the foveal area of the retina and neurons in the corresponding projection zones of the primary visual cortex. The effective zone was shown to have to contain a scattering function for several receptors, as this allowed preliminary blurring of the image by the eye's optics to decrease the subsequent (at the level of receptors) discretization noise created by a matrix of receptors. The concordance of these parameters supports the optical operation of the spatial elements of the neural network determining the resolving ability of the visual system at different levels of visual information processing. It is suggested that the special geometry of the receptive fields of neurons in the striate cortex, which are concordant with the statistics of natural scenes, results in a further increase in the signal:noise ratio.