From psychophysics to neural activity: an inference about the motion pathway in visual perception

MARTÍN A.; BARRAZA J.F.

Córdoba

Congreso; 12 Taller Argentino de Neurociencia - Primera Reunión Conjunta Taller-SAN; 2009

Taller - SAN (Sociedad Argentina de Neurociencia)

Recently, we showed that the spatial layout of a stimulus affects our ability to locate the border defined by two adjacent surfaces with different speeds.  However, this effect does not appear or is very weak when the task is to discriminate a transparent stimulus [Martin, Barraza, and Colombo (2009)]. Both segmentation tasks would be performed by V1 and/or V2 cells. In addition, it has been shown that the case of segmentation of adjacent surfaces would be mediated by a feedback from MT to these lower areas (Hupe et al,, 2001). Because the effect of the spatial layout would reflect an integrative mechanism occurring in MT such was proposed by Verghese and Stone (1995), then we may hypothesize that the difference between the two kinds of segmentation reported in our previous work could be due to the fact that transparency is not mediated by this feedback. In the case of two adjoining surfaces with different speeds, an integration process may improve the speed estimation of each surface by keeping the information of the two different speeds but, in the case of transparency, an integration in the scale of MT would blend the local signals and thus, losing the information given by the speed difference. An alternative hypothesis is that the results are just reflecting differences in the psychophysical tasks. In order to disentangle these two alternatives we devised an experiment in which we test the effect of stimulus layout on speed threshold as a function of target size. Results show that increasing the angular distance among patches reduces the speed threshold only for targets sizes larger than 0.4 deg, which suggests that transparent stimuli are not affected by the spatial integration occurring after V1/V2.