Ichasus Llamas-Cornejo¹ (illamas@ucm.es), Ignacio Serrano-Pedraza¹; ¹Universidad Complutense de Madrid

There is a well-stablished anisotropy in stereovision: sensitivity for detecting sinusoidal disparity corrugations of low spatial frequency is higher for horizontal than for vertical corrugations. Despite the strong evidence for this stereoscopic anisotropy, its cortical locus, underlying mechanisms, and functional role of this phenomenon remain unclear. Several explanations have been proposed, including the number of underlying channels or the size of the summation fields in stereovision, but none provide a clear explanation of the effect. In the present study, we used a masking paradigm to estimate the internal noise and efficciency for detecting vertical and horizontal sinusoidal corrugations. Using a Bayesian staircase procedure, we measured disparity thresholds at three spatial frequencies (0.2, 0.,4, and 0.8 cpd) masked by disparity white noises of different power. As expected, our results show that thresholds increased with noise power. We then fit an equivalent noise model to the threshold vs. noise power data. The model has two parameters: internal noise and efficiency. The fitting results indicate that the mechanism responsible for detecting horizontal corrugations exhibits lower internal noise and higher efficiency than the mechanism detecting vertical corrugations. Interestingly, while internal noise increased with spatial frequency, efficiency remained approximately constant across spatial frequencies for both orientations. These findings support the existence of different orientation mechanisms underlying disparity processing and suggest that differences in internal noise and efficiency contribute to the stereoscopic anisotropy.

Acknowledgements: Supported by grant PID2024-155243NB-I00 from Ministerio de Ciencia, Innovación y Universidades (Spain) to ISP.