Reading over this account I wonder if there is an alternative interpretation available consistent with the predictive coding account. It is consistent with the view that patterns of activity in the AIC are abnormal in CD, but unlike DPD those patterns are not the result of VLPFC-induced hypoactivity.
Ex hypothesi the CD patient is extremely depressed. Evidence suggests that circuitry centred on the amygdala is affected, which means that online affective responses are flattened.
The role of the AIC is to monitor for changes driven by affective processing. It thus predicts for example that a typically positive event would be processed as positive. Thus, when that event is processed as negative or neutral, the AIC detects an error, signaled in the form of an anomalous experience. The patient is in the position being able to detect and signal changes in her affective responses, which take the form of unpredicted absences in bodily response. Thus her lack of felt bodily response is processed as affectively significant in the Cotard delusion with the result that she experiences it. Thus she does not feel neutral she feels miserable. Or as we might put it she feels metamisery because the role the AIC is to enable the person to feel the affective significance of bodily changes including the absence of predicted changes. In Cotard delusion the patient feels the affective significance the unpredicted absence of positive changes.
In DPD, by contrast, the patient does not feel the significance of bodily information because her AIC is inhibited and hypoactive.
Thus the difference between the two conditions may consist, not only in the fact that the Cotard delusion is a response to lower level prediction error, but the level in the predictive processing hierarchy at which predictions about bodily information are violated.