Purchase this article with an account.
Daniel Leeds, John Pyles, Michael Tarr; Real-time fMRI search for the visual components of object perception. Journal of Vision 2014;14(10):1302. doi: 10.1167/14.10.1302.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The nature of visual properties used for object perception in mid- and high-level vision areas of the brain is poorly understood. Past studies have employed simplistic stimuli limited in the visual details representative of real objects. Unfortunately, pursuit of more complex stimuli and properties requires searching through a wide, unknown space of models and images. The difficulty of this pursuit is exacerbated in neuroimaging by the limited number of samples that can be collected for a given human subject over an experiment. To more rapidly identify complex visual features underlying cortical object perception, we developed and tested a novel method in which visual stimuli are selected in real-time based on BOLD responses to recently shown stimuli. A variation of the simplex method (Cardoso, 1996) controlled continuous stimulus selection within a real-time search through visual image space designed to maximize neural responses across a pre-determined 1 cm3 brain region within ventral cortex. This method was applied using two different stimulus sets: photographs of real-world objects and 3D synthetic "Fribble" objects (Williams 2000). To assess the value of this search method for the understanding cortical object encoding, we examined both the behavior of the method itself and the complex visual properties the method identified as highly activating the selected brain regions within the ventral visual pathway. While further technical and biological challenges remain, our results demonstrated convergence for complex visual properties in the majority of our subjects for a subset of the searches we conducted. More specifically, we were able to identify ventral regions selective for both holistic and component object shapes and for a variety of surface properties, providing evidence that these methods may yield more precise measures of selectivity within the broad classes of visual features associated with cortical object representation (Hung 2012, Tanaka 2003, Vogels 1999).
Meeting abstract presented at VSS 2014
This PDF is available to Subscribers Only