Key points are not available for this paper at this time.
Abstract Neural systems that participate in working memory for faces were investigated in an experiment designed to distinguish face perception areas from working memory areas. Regional cerebral blood flow (rCBF) was measured using positron emission tomography (PET) while subjects performed a sensorimotor control task, a face perception control task, and five working memory tasks with parametrically varied retention intervals, ranging from 1 to 21 sec. Striate and ventral occipitotemporal extrastriate areas demonstrated a simple negative correlation between rCBF and retention delay, indicating that these areas participate principally in perceptual operations performed during visual stimulation. By contrast, right and left frontal areas demonstrated rCBF increases that were significantly more sustained across delays than were increases in ventral extrastriate areas, but the relation between rCBF and retention interval differed significantly by hemisphere. Whereas right frontal rCBF showed a nonsignificant tendency to diminish at longer delays, left inferior frontal, middle frontal, and anterior cingulate cortex, as well as left parietal and inferior temporal cortex, demonstrated their largest rCBF increases at the longest delays. These results indicate that right frontal and left frontal, parietal, and temporal areas all participate in face working memory, but that left hemisphere areas are associated with a more durable working memory representation or strategy that subjects rely on increasingly with longer retention intervals. One possible explanation for this hemispheric difference is that left hemisphere activity is associated with a face representation that embodies the result of more analysis and elaboration, whereas right frontal activity is associated with a simpler, icon‐like image of a face that is harder to maintain in working memory. © 1995 Wiley‐Liss, Inc.
Haxby et al. (Sun,) studied this question.