Linear and Nonlinear Properties of Feature Selectivity in V4 Neurons

• Main Authors: Jonathan O. Touryan, James A. Mazer
• Format: Article
• Language: English
• Published: Frontiers Media S.A. 2015-05-01
• Series: Frontiers in Systems Neuroscience
• Subjects: extrastriate cortex; natural scenes; feature-selectivity; Natural Scene; receptive field; V4
• Online Access: http://journal.frontiersin.org/Journal/10.3389/fnsys.2015.00082/full

Extrastriate area V4 is a critical cortical component of visual form processing in both humans and non-human primates. The tuning of V4 neurons shows an intermediate level of complexity that lies between the narrow band orientation and spatial frequency tuning of neurons in primary visual cortex and the highly complex object selectivity seen in inferotemporal neurons. Single neuron recording studies in the monkey have demonstrated that V4 neurons can be highly selective for complex properties of visual stimuli, like contour curvature (Pasupathy and Connor, 1999) and the relative positions of object features in the receptive field (Pasupathy and Connor, 2001). However, the origins of complex feature selectivity and the specific circuits that transform the relatively simple wavelet-like encoding seen in primary visual cortex into the more complex selectivity profiles observed in V4 are not well understood. This is especially true in the case of selectivity for features occurring in natural scene stimuli. Little is known about how the selectivity of V4 neurons to isolated stimuli is altered when those stimuli appear in the context of a spectrally complex natural scene. Previous work using quasi-linear system identification methods has shown that some, but not all, of the responses of V4 neurons to natural stimuli can be accounted for by a neuron’s orientation and spatial frequency tuning (David et al., 2006). In this study we assessed the degree to which preferences for natural images can really be inferred from classical orientation and spatial frequency tuning functions. Using a psychophysically-inspired method we isolated and identified the specific visual driving features occurring in natural scene photographs that reliably elicit firing from single V4 neurons. We then compared the measured driving features to those predicted to drive each cell based on the linear spectral receptive field (SRF), which was estimated from responses to narrowband sinusoidal grating