Article

Faces are multidimensional stimuli that convey information for complex social and
emotional functions. Separate neural systems have been implicated in the
recognition of facial identity (mainly extrastriate visual cortex) and emotional
expression (limbic areas and the superior temporal sulcus). Working-memory (WM)
studies with faces have shown different but partly overlapping activation
patterns in comparison to spatial WM in parietal and prefrontal areas. However,
little is known about the neural representations of the different facial
dimensions during WM. In the present study 22 subjects performed a face-identity
or face-emotion WM task at different load levels during functional magnetic
resonance imaging. We found a fronto-parietal-visual WM-network for both tasks
during maintenance, including fusiform gyrus. Limbic areas in the amygdala and
parahippocampal gyrus demonstrated a stronger activation for the identity than
the emotion condition. One explanation for this finding is that the repetitive
presentation of faces with different identities but the same emotional expression
during the identity-task is responsible for the stronger increase in BOLD signal
in the amygdala.
These results raise the question how different emotional
expressions are coded in WM. Our findings suggest that emotional expressions are
re-coded in an abstract representation that is supported at the neural level by
the canonical fronto-parietal WM network.
CI - Copyright (c) 2010 Elsevier Ltd. All rights reserved.

Langue : ANGLAIS