Células simples e complexas no wulst visual da coruja

Abstract

The visual wulst is the avian telencephalic target of the retinothalamofugal pathway. In owls, this area is known to have a close functional resemblance with mammalian primary visual cortex (V1). In his pioneering study, Pettigrew reports that this resemblance also extends to two basic cell types initially described by Hubel and Wiesel: simple cells, characterized by approximately linear spatial selectivity to stimulus contrast polarity, and a much more diverse population of so-called complex cells, which do not exhibit this response property. Here, we reassess Pettigrew's finding using a more quantitative classification method derived from linear system theory. The method computes a linearity index of spatial summation based on the spiking response of a cell to moving sinusoidal gratings optimized in terms of spatial as well as temporal frequencies, orientation and direction of motion. The index ($F_{1}/F_{0}$) is obtained by dividing the amplitude of response oscillation at the fundamental frequency of the stimulus ($F_{1}$) by the mean response of the cell above spontaneous discharge ($F_{0}$). Cells yielding a $F_{1}/F_{0}$ $>$ 1 are classified as simple cells; the others as complex cells. The reliability of this classification method is supported by two lines of evidence: 1) general agreement with other mapping methods; 2) Bimodal distribution of $F_{1}/F_{0}$ with a pronounced dip close to unity across V1 neurons of most mammalian species investigated to date. We analyzed a total of 369 neurons randomly sampled along different depths of the visual wulst of sixteen awake burrowing owls. Unexpectedly, the $F_{1}/F_{0}$ frequency distribution was found to be unimodal (median $F_{1}/F_{0}$ = 0.43) with no clear evidence of local minimum at unity (P = 0.997, Hartigan's dip test). Adopting the classification criterion conventionally used in mammals (see above), 83\% of our sampled cells were classified as complex (median $F_{1}/F_{0}$ = 0.36). The proportion of simple cells (median $F_{1}/F_{0}$ = 1.33) was however only slightly lower than that reported by Pettigrew in anesthetized barn owls (21\%, 126/579). Neurons classified as simple and complex had similar median spontaneous rates (3.43 and 4.24, respectively) and $F_{1}/F_{0}$ was not significantly correlated with spontaneous activity. According to our data, it was found a considerable variability of $F_{1}/F_{0}$ for low and high contrast, 7.5\% and 98\%, respectively. Some wulst's cells (about 1/4 of evaluated cells, n = 120) showed different classification for both contrasts. Most wulst's neurons exhibit marked response suppression when a drifting grating is expanded beyond their classical receptive field (CRF). Here, we found that this suppressive effect is actually often accompanied by a significant increase in $F_{1}/F_{0}$ (about 2/3 of evaluated cells, n = 86). Receptive field subunits mapping methodologies were developed and applied using the reverse correlation in order to characterize its internal structures. The receptive field subunits maps showed a dynamic behavior in the spatial localization of receptive field subunits ON and OFF during time and a receptive field reconfigurability depending on the stimulus statistics. The receptive field was more 'complex' like when stimulated by a sparse noise and more 'simple' like when stimulated by a dense noise. Altogether, our results do not support the existence of two discrete simple/complex cell classes within the visual wulst. They further suggest that the degree with which a wulst neuron operates as a spatiotemporal linear filter dependents on specific combinations of stimulus attributes.