Annual Meeting of the Society for Neuroscience edition:38 location:Washington DC, USA. date:15-19 November 2008
Research into the organization of high-level visual cortex has mainly focused on using discrete object classes like faces and buildings. In contrast, the large-scale retinotopic organization in lower levels of the visual cortex has been studied with techniques that allow continuous mapping, termed phase-encoding techniques. In this procedure, the stimuli typically consist of gradually changing rotating wedges and contracting/expanding rings. These traveling waves in retinal space produce traveling waves on the surface of the brain, and the time delay or phase of the response of each cortical location defines the most effective stimulus position. Here, we propose the extension of this method to higher-order properties of visual objects. We used continuously changing stimuli, forming a continuum between a face and a house. Both stimuli were constructed using a 3D program, and the subsequent morphs were created using clearly defined corresponding features and a point per point transformation. The resulting morphs were ambiguous stimuli with both house- and face-like features. The stimuli were shown from three possible angles (front, 45° left and 45° right orientation). In a separate experiment, we used contracting and expanding rings cut out from natural scenes. To validate the method, we focused initially on face-selective voxels (the fusiform face area or FFA) and building-selective voxels (the parahippocampal place area or PPA), as defined from localizer scans with face and building images presented in separate blocks. We found that the method does indeed work as intended. Activity in FFA was locked to the presentation of face or face-like images, while activity in PPA was locked to the presentation of building or building-like images. In addition, and again confirming the existing literature based on traditional blocked stimulus presentation, the activity of voxels in FFA and PPA was locked to the presentation of foveal and eccentric rings, respectively. These results indicate that the phase-encoding techniques used in retinotopic mapping can also be applied to stimuli with feature variations that are more suited to investigate the organization of high-level visual regions.