352 control therapy psychopathological scales and perfor- mance tests were employed. These results suggest improvements in performance and decrease in basic disorders. In previous investigations, we have demonstrated chat patients with significantly more subjectively per- ceived basic disorders are discharged due to socio- therapeutic-psychopaedagogic approach alone (Land- siedel, 1986). Of course, the findings obtained need to be confirmed by control group design because drug effects and spontaneous improvements in the disease have to be taken into account. I-D MECHANISM ACCOUNTING FOR HYPERA- CUITY IN LOCATION PERCEPTION Petrauskas V., VaitkeviEius H. The human visual system is capable of judging relative position with hyperacuity precision (Westheimer, 1985; Barlow, 1979; Watt and Morgan, 1983) proposed two mechanisms of hyperacuity: (I) up to stimuli separation of 5-10 min arc, therelative positions of the stimuli are discriminated directly, i.e. by assessing their location; (2) at larger separations, the positions are discriminated in terms of the slope of the virtual line joining them. Our main interest will be focused on the dot (or other stimulus) location assessment involved in mechanism (I). Mat-r (1982), Watt and Morgan (1983) suggest that the location of zero-crossings in the second derivative of the retinal light distribution may encode the precise location. However, the mean of light distribution (MLD) in a small retinal area may encode the precise location as well. We would like to present such a mechanism. Let us have a line of photoreceptors (PhR in the figure). I(X) in the figure stands for light distribution. Our main goal will be to show that the mean of I(X) may be defined with the hyperacuity precision, i.e. with the threshold which is finer than the width of the photoreceptor. Let signals from photoreceptors be summed by two devices Cl and C2, having odd- and even-symmetric weighing functions (Wl and W2 in the figure) respec- tively. We suggest that the mean of I(X) is encoded by the orientation of a 2-D vector. E”= (Yl; Y2), where Y 1 and Y2 are the output signals of the devices Cl and c2. The dependence of the vector E orientation (cp on the ordinate) upon the location of the mean of I(X) (X on the abscissa) is shown in a lower part of the figure as a function q(X). Numerals on the abscissa denote photoreceptors. The continuous nature of q(x) show that the mean of I(X) has the location threshold finer than interphotoreceptor spacing. The presented mechanism is local, i.e. devices c2 and I;2 sum signals from a few photoreceptors (10 in our case) in a small region. We call this region the receptive field (RF) of the devices Cl and C2. In such a scheme, discrimination between two stimuli presented in one RF is not possible, because the location of MLD is assessed. To undergo this problem we propose whole visual field being processed by a great number of such mechanisms with overlapping RFs. According to such a scheme, continuous function I(X) is transformed in its sampled version S(X0): S(XO)=R(I(X))*60<-X0), where X0 are the locations of MLD in RFs; R(I(X)) stands for the response of photoreceptors to light distribution I(X); “*” denotes the operation of convolu- tion. To extend the above approach to the 2-D case one should make definitions of MLD in two orthogonal directions using two independent 1-D mechanisms. From the neurophysiological point of view the line of PhRs may be thought of as 1-D RF of two neurons Cl and C2. The odd-and even-symmetric weighing functions Wl and W2 resemble very much the ones of adjacent simple cells in striate cortex (Pollen and Ronner, 1983). In conclusion we would like to point out that it is not clear up to now, whether our visual system assesses locations in local visual field by (1) comparing relative responses of adjacent neurons, as it is done in our model, or (2) location detectors are used. If the second case is true, the output signals Y 1 and Y2 of neurones Cl and C2 can be used to feed location detectors (Fomin et. al, 1979).