Effects of high and low spatial filtering and spatial location of fearful faces on amygdala and fusiform gyrus activity

Abstract

8were presented within an oval frame excluding the hair and non-facial contours. Faces were bandpass filtered using a radially symmetric filter and normalized in luminance. Spatial frequency (SF) content in the original stimuli (broad-band, BSF) was filtered (high-pass cut-off: 24 cycles/image for HSF stimuli; low-pass cut-off: 8 cycles/image for the LSF stimuli). Pairs of images, one face (LSF, HSF or BSF) and one Fourier transformed image, were presented either 1.7° (height: 3.8°; central location) or 9.5° (height: 8.2°; peripheral location) from central fixation for 500ms. Subjects had to indicate via button press on which side the face appeared. As control condition two Fourier images (F) were displayed, one containing a black oval-shaped circle as target. The photographs were scaled by the human cortical magnification factor to activate an approximately equivalent portion of early visual cortex at all stimulated eccentricities 9 . The four task conditions were implemented in a blocked design and were separated from each other by a fixation condition (fixation cross in the middle of a white screen; Fix). The fusiform gyrus was localized in a separate run by presenting blocks of greyscale face or house images interleaved with Fourier images, which were all fit behind an oval mask. BOLD fMRI was performed at 3 Tesla (Siemens TRIO, whole-brain EPI, TR 2s, TE 36ms, 2x2x2mm³). Analysis was performed using the general linear model approach (BrainVoyager QX). Regions of interest (ROIs) were determined on a single subject level either anatomically (amygdala) or functionally (fusiform gyrus) on the basis of the localizer experiment contrasting the face versus the house stimuli. Results The ROI-analysis revealed a different modulation of activity in the amygdala (Fig. A) and the fusiform gyrus (Fig. B). The spatial frequency of the emotional stimuli had an effect on the responses in the amygdala and the fusiform gyrus whereas the latter one was also affected by the eccentricity of the faces. A decrease in response was observed in the amygdala for all task conditions in contrast to fixation (Fix), regardless of the position of the stimuli. However, contrasting the faces against the control condition (F) was associated with an increase in activity in the amygdala. BSF faces activated the amygdala significantly stronger than LSF and HSF faces. Corresponding t-values of all contrasts are depicted in the bar graph in Figure A. The location of the emotional stimuli had no significant effect on amygdala responses (not shown). In the fusiform gyrus increased activity was found during all task conditions compared to fixation (Fix). The face tasks lead to significantly higher activity in the fusiform gyrus than the control task (F). The BSF face task resulted in significantly higher responses in the fusiform gyrus than the face stimuli only containing either LSF or HSF visual information. Corresponding t-values of all contrasts are depicted in the bar graph in Figure B. The spatial location of the faces also had an effect on the activity in the fusiform gyrus as faces presented in the central location rather than the peripheral one were associated with a significantly stronger response (not shown). Discussion