Cerebral regions and hemispheric specialization for processing spatial frequencies during natural scene recognition. An event-related fMRI study

It has been suggested that visual scene recognition is mainly based on spatial frequency (Fourier) analysis of the image. This analysis starts with processing low spatial frequencies (LSF), followed by processing high spatial frequencies (HSF). Within the framework of the spatial frequency analysis, the right/left hemisphere would be predominantly involved in LSF/HSF analysis, respectively. The aim of this event-related fMRI study was to evaluate neural correlates and hemispheric specialization of spatial frequency analysis during recognition of nonfiltered (NF) and filtered, either in LSF or HSF, natural scenes. Comparing LSF or NF to HSF scene recognition, significant activation was obtained within right anterior temporal cortex and right parahippocampal gyrus. As these regions are known to be involved in scene processing, we interpret this result as suggesting that scene recognition is mainly based on LSF extraction and analysis. When LSF scene was compared to HSF scene recognition, supplementary activation was obtained within the right inferior parietal lobule that likely reflects attentional modulation on spatial frequency processing. A direct interhemispheric comparison for each particular band of spatial frequencies highlighted predominance within the early visual areas (such as the middle occipital gyros) to the right for LSF processing and to the left for HSF processing. This result provides supplementary evidence for hemispheric specialization at early levels of visual analysis when spatial frequencies are processed. (C) 2004 Elsevier Inc. All rights reserved.