We have argued that not all BOLD responses across the amygdaloid region are related to the status of fearful faces as negatively valenced. If negative valence alone dictated amygdala response to facial expressions, response to angry faces would be similar to that observed to fear. But neither the human lesion data nor the early imaging data provide overwhelming support for this notion. We have hypothesized that fearful faces are particularly strong activators of the amygdala because of what they do not predict for the viewer. Both fearful and angry expressions predict an increased probability of threat. But while angry faces provide clear information about the source of threat, fearful faces do not.
The Figure presented here summarizes our view of how these expressions would be handled by the amygdaloid complex. Fearful and angry faces would both be detected by the basolateral complex (BLC) within the ventral amygdala as negatively valenced. Because angry faces are a clearer negative signal (i.e., source of threat known), they are handled via outputs at the BLC level and an appropriate response plan is enacted (e.g., are you fighting or fleeing)? But in the case of the fearful face (i.e., source of danger unknown), the to-be-potentiated response is not yet clear. The greater context-dependence of fearful faces (relative to anger) will require the central nucleus (Ce) within dorsal amygdala and its connections to neuromodulatory centers within the substantia innominata (SI) such as the nucleus basalis of Meynert (NBM) to call upon the cortex for additional information.
Consequently, response thresholds would be lowered throughout sensory cortex to increase vigilance and communicate additional information back to the BLC. This loop could remain in effect until adequate information enacted response potentiation via BLC outputs. While this scenario is admittedly speculative, it does make a clear prediction concerning amygdala response to fearful expressions relative to angry expressions. If directly compared during fMRI, fearful expressions should produce a significant increase in signal in the amygdaloid region compared to anger, but this difference should be confined to the dorsal amygdala/SI. Response to anger and fear will not be significantly different within the ventral amygdala. See Whalen et al (2001) in the Publications link for data consistent with this hypothesis.