Located in the fusiform gyrus in the extrastriate cortex.
Specialized for face perception.
fMRI studies showed higher activation when viewing faces compared to objects.
Methods
fMRI was used to measure brain activity while subjects viewed faces and objects.
Regions of interest (ROIs) were defined based on face-selective activation.
Multiple stimulus comparisons were conducted within the same ROIs.
Key Findings
FFA responds more strongly to faces than to various control stimuli.
FFA is not solely driven by low-level visual features.
FFA is not simply involved in subordinate-level classification or general processing of animate objects.
The study suggests that face recognition may involve distinct neural mechanisms compared to object recognition.
Specificity Tests
Intact vs. Scrambled Faces: FFA responds more to intact faces, ruling out low-level feature extraction.
Faces vs. Houses: FFA responds more to faces than houses, suggesting face-specific processing.
Three-Quarter-View Faces vs. Hands: FFA responds more to faces, indicating it's not just animate objects or body parts.
1-Back Matching Task: FFA activation during face matching is higher than during hand matching, controlling for attention and subordinate-level classification.
Alternative Interpretations Rejected
Visual Attention: Ruled out as the primary driver of FFA activity.
Subordinate-Level Classification: FFA activity is specific to faces, not general object recognition.
General Animate Object Processing: Hands were used as a control to rule out general animate object processing.
Location and Consistency
The right fusiform gyrus showed consistent face activation across subjects.
активация была очень последовательной в разных сеансах тестирования у одного и того же субъекта.
Relation to Other Brain Areas
Middle temporal gyrus/superior temporal sulcus (MT gyrus/STS) may also be involved in face processing.
The FFA might be the human homolog of the macaque inferotemporal region.
Implications
The existence of FFA supports the idea of specialized cortical machinery for face perception.
Suggests that face recognition may involve qualitatively different computations than object recognition.