[The meaning of the Weber-Fechner Law IV. The psychometric curve, the interaction between inter hemisphere and the interaction in intrahemisphere].

The subject was presented with a pair of lines. One line was an objective referent line Ro and did not change its length. The other was a test line and changed the its length. The observer must report--where the longer line was--above or below in the pair. On the basis neurophysiological, psychophysical data and new type of the psychometrical curve a complex neuronal construction was built. The comparison of action of the neuronal construction with the psychophysical data on different stages of introduction the visual information showed that in both cases at first stages of introduction the information referent subjective stimulus Rs is not equal to Ro. But at the last stages, when Weber's fraction is obtained Rs = Ro. This result allows to suggest that the suggested neuronal contraction is near to the real constriction of the visual brain.

Separate training of hemispheres to perform visual discrimination in conditions of blocking of interhemisphere transmission by masking.

Subjects were trained to discriminate three figures presented in the left field of vision and three other figures presented in the right field of vision. In these conditions, the two hemispheres usually show identical learning to discriminate the sets of stimuli because the hemispheres in healthy humans can exchange information. In the present study, training was performed in conditions in which, during presentation of stimuli, the opposite visual hemifield was covered by a mask. After training, the recognition of all six figures was compared by presenting them to the left and right visual fields. Each hemisphere recognized figures presented to the cognate hemifield but completely or very nearly failed to recognize figures learned by the other hemisphere. The mask would thus appear to block (completely or partially) the transmission of information from one hemisphere to the other. Thus, it was possible to train the hemispheres separately to recognize different sets of images in healthy subjects.

[Recognition of visual images by separate hemispheres in conditions of masked blocking of interhemispheric transmission]. / Razdel'noe obuchenie polusharii zritel'nomu raspoznavaniiu v usloviakh blokirovaniia maskirovkoi mezhnolusharnogo perenosa.

The subjects learned to recognize three figures presented in the left visual hemifield and three figures presented in the right visual hemifield. During presentation of a stimulus, the contralateral hemifield was overlapped by a mask. After the training, recognition of all six figures presented in the right and left visual hemifields, was compared. Each hemisphere recognizes figures which were learned in the corresponding visual hemifield, but the recognition of figures learned in the opposite visual hemifield was poor. Thus, the ability of the hemispheres to act separately in recognizing different sets of visual images, was established.

Linear and nonlinear properties of simple cells of the striate cortex of the cat: two types of nonlinearity.

In a proportion of simple cells of the striate cortex, the weighting functions of the receptive fields (RFs) had more periods than could be established by mapping using responses to light bars and dark bars. In these multiperiodical cells, side subfields do not respond to single bars, as they have lower weights than central zones and the excitation is under the threshold of impulse response if a single bar is applied. This fact has been established by different methods: conditioning and testing stimuli, grating patches, and inverse Fourier transform of the amplitude-phase characteristic, combining them in one cell. We assume that this type of nonlinarity can be used in analyzing the image, as it acts as a spatial-frequency filter of the area overlapped by the RF. The responses to complex gratings composed by two sinusoidal gratings of different frequency, contrast, and phase shift were compared with the sum of the responses to the gratings when they were presented separately. The results show that the principle of superposition holds a reasonable approximation even if the response is evoked from the side subzones. Some simple cells have nonlinear properties beyond the classic zone of RF (2nd type of nonlinearity). Linear cells have a tendency to be localized in layer 4 of striate cortex, cells with a nonlinear surround in layers 2, 3, 5, and 6. The significance of both types of nonlinearities in simple cells is discussed.

Algorithm of invariant image description by the use of a modified Gabor transform.

A scale-invariant algorithm of image description that is derived from the investigations of the neural structure of striate complex modules is proposed. First an image is preprocessed by difference-of-Gaussians functions, then the Gabor expansion in modules of different scales is used, and a group of the most excited modules is looked for with the help of scale weighting functions (SWF's). Finally, the image code is calculated as a linear combination of the output signals from the Gabor modules. The coefficients of this linear combination correspond to the SWF selected in the previous step. The scales of Gabor modules are chosen to be a logarithmic sequence with a half-octave step. Because of the SWF method, the described algorithm takes into account all the image sizes on a continuum between the minimum and the maximum scales of the Gabor modules.

Harmonic basis functions for spatial coding in the cat striate cortex.

The number of subregions in the activity profiles of simple cells varies in different cells from 2-8; that is, the number of cycles in the weighting function varies from 1-4. The distribution of receptive-field (RF) sizes at eccentricities of 0-6 deg are clustered at half-octave intervals and form a discrete distribution with maxima at 0.62, 0.9, 1.24, 1.8, 2.48, and 3.4 deg. The spatial frequencies to which the cells are tuned are also clustered at half-octave intervals, forming a discrete distribution peaking at 0.45, 0.69, 0.9, 1.35, 1.88, 2.7, 3.8, and 5.6 cycles/deg. If we divide the RF sizes by the size of the period of the subregions, then the average indices of complexity (really existing) or the number of cycles in the weighting function form (after normalization) the sequences: 1, 1.41, 2.0, 2.9, 4.15. The relation between the bandwidth of the spatial-frequency characteristic and the optimal spatial frequency is in accordance with predictions of the Fourier hypothesis. The absolute bandwidth does not change with the number of cycles/module. This means that inside the module the absolute bandwidth does not change with the number of the harmonic. The results allow us to suggest the following. A module of the striate cortex, which is a group of cells with RFs of equal size projected onto the same area of central visual field, accounts for the Fourier description of the image. The basis functions of the module are composed of four harmonics only, irrespective of size and position of the module. Besides linear cells (sinusoidal and cosinusoidal elements), the module contains nonlinear cells, performing a nonlinear summation of the responses of sinusoidal and cosinusoidal elements. Such cells are characterized by an index of complexity which is more than the number of cycles in the weighting function and by marked overlap of ON and OFF zones. The analysis of organization suggests that the cells can measure the amplitude and phase of the stimulus.

Spatial organization of subfields in receptive fields of cells in cat striate cortex.

Spatial organization of receptive fields (RF) of cells in cat striate cortex was investigated with moving and flashing light and dark bars and with grating-patterns of a varying number of cycles. It was shown that the maximum number of subfields in a simple cell is equal to eight or the number of periods in weighting function is equal to four. Quantitative comparison of the data with the results of seven other studies allows us to suggest that the number of periods in linear component of some complex cells is close to this value. The discrepancies between the results of different authors in estimation of the number of subfields are explained by the experimental data.

[Principle of uncertainty in vision]. / Printsip neopredelennosti v zrenii.

When investigating the dependence of the bandwidth of the spatial frequency characteristic (delta F) of the receptive fields of the cat striate cortex on the size (D) of the receptive field it was shown that delta FD = 1.2. Thus in the system of elements performing the processing of information in the visual cortex the inaccuracy of the signal representation in the spatial frequency and in space is related by an uncertainty principle. The size of the constant evidences that the weighting functions of elements are sinusoids or cosinusoids modulated by squarewave impulse rather than by the Gabor elements. The performing of uncertainty principle and the size of the constant are arguments in favour of the hypothesis that the receptive fields of the visual cortex organize a quasilinear basis.

Relationship between spatial and spatial-frequency characteristics of receptive fields of cat visual cortex.

The spatial (magnitude and eccentricity) and spatial-frequency (optimum frequency and width of pass band) characteristics of the receptive fields of the cat visual cortex were investigated. It was shown that in accordance with the predictions of the theory of piecewise Fourier analysis, linear and quasilinear receptive fields of a single size comprise a modulus in each of the fields of which the index of complexity (ratio of size of field to number of periods of its optimum frequency) equals the optimum frequency multiplied by a coefficient that is constant for the given modulus. Five moduli were found with field sizes of 2.6, 3.8, 5.2, 6.2, and 7.0 degrees, shifting with increase in the size of the modulus towards the periphery of the field of view. In accordance with predictions, when the index of complexity is fixed the width of the pass band declines inversely proportionately to the size of the fields. The obtained data directly support the hypothesis according to which the receptive fields effect a piecewise quasi-Fourier expansion of the image.

[2 types of generalization in the evolution of the visual brain]. / Dva vida obobshcheniia v évoliutsii zritel'nogo mozga.

It has been shown that mechanism of generalization with respect to the size (shift of differentiation of the size from objects of one form to other forms) is located in dogs within the suprasylvian convolution, in cats--within the median part of the lateral suprasylvian region. After removal of these parts of the brain, other visual functions including the invariant description of the image, remain unaffected. The latter is disturbed after extirpation of the field 21; however, the shift with respect to the size does not undergo any significant changes. Therefore, two types of visual generalization, i.e. phylogenetically less ancient generalization of the detected properties and evolutionary more ancient subject generalization (invariant image), are located in different parts of the visual brain.