Abstract
Previous work in our lab used a backward masking paradigm to demonstrate that activation of the fusiform face area (FFA) reflects not only the contents of the visual image, but also the subject's perception of the image (i.e., with identical masking parameters, FFA activation is significantly greater when a face is detected than when it is not). In the present study, we used a rapid-presentation event-related fMRI design and methods of signal detection theory to examine the role of FFA while subjects made perceptual decisions requiring the discrimination of face from non-face images. Grayscale stimuli were constructed from photographs of human faces and scrambled versions of the same, with contrast adjusted in a staircase manner so that subjects were able to detect the images with an accuracy of approximately 79%. Subjects were then required to discriminate between these stimuli, when presented briefly (40 ms) in a 2-alternative forced choice paradigm. fMR images of FFA in the right hemisphere were acquired at 4T (3×3×5 mm, 1.5 s TR, 10 parasagittal slices) using a 12 cm single-loop surface coil. For all subjects, activation of the FFA during trials in which faces were correctly reported as faces (i.e., hits) was found to be significantly greater (p < 0.05) than during trials in which these same images were incorrectly reported as being scrambled (i.e., misses). Furthermore, the subjects' sensitivity (d') in the discrimination task was directly correlated (r = 0.893) with the difference in the levels of FFA activation for the intact face and scrambled images. This provides evidence that the activation of the FFA reflects decision processes critical to perceptual discrimination.
Supported by NSF grant BCS — 9996264