In his target paper, Philip Gerrans proposes a theory of self-awareness that integrates the predictive coding framework, the appraisal theory, and the simulation model. It accounts for the loss of self-awareness in DPD and CD, and provides a new understanding of patients’ anxiety. In this commentary, I have proposed (1) that the simulation model should be considered a hierarchical model involving multiple memory systems—namely, it is constituted by procedural, semantic, and episodic memory and prospection (section 2); and (2) that the function of memory systems or simulation models, under the predictive coding framework, is to update the knowledge required for successful prediction (section 3). This implies that memory function and malfunction are independent from the generation of a predictive model that succeeds or fails in representing the world, since it is possible that misrepresentation rather than veridical representation leads to a generative model that minimizes prediction error. Based on such view of the simulation model, CD can be regarded as the modification of top-down prediction in an attempt to explain away the prediction error resulting from unexpected hypoactivity of the AIC. I also suggested (3) that a combination of two factors is necessary for the occurrence of CD from depersonalization: the compromised decontextualized supervision system and the expectation of high precision of interoceptive signals (section 4).
If both the general framework and my suggestions are correct, there are a number of issues worthy of further investigation: First, if the model that explains the symptoms of CD is created by the system in order to minimize prediction error from hypoactivity of the AIC, with the aim of affording relief from anxiety, it is expected that the change of prediction may be accompanied by minimized prediction error or/and prediction error from other unpredicted activities. In the case of Cotard delusion, the new model—the model of the organism’s death or non-existence—would encounter new kinds of prediction error due to information about bodily states, instead of a lack of emotional significance. This may as well be the kind of prediction error that cannot be cancelled top-down and which can be expected to lead to anxiety based on Gerrans’ theory. Therefore, the anxiety characteristic of the Cotard delusion is speculated to be the result of different prediction errors from patients suffering from Cotard syndrome. Studies on the difference between the anxiety present in DPD and that in CD can the support or refutation of the framework proposed. Furthermore, it is worth noting that not all patients with the CD suffer from anxiety. For example, in Berrios & Luque’s (1995) analysis of 100 cases, anxiety is reported in only 65% of subjects, and patients were categorized: Cotard type I patients showed no affective component, whereas type II patients showed depression and anxiety. Can the proposed framework account for both types of patients?
Another interesting question for future research is whether we can better understand the relation between the simulation model and affective processing within the predictive coding framework, and whether an explanation of this would be consistent with the existing evidence relating to emotional memory (e.g., LaBar & Cabeza 2006). Affective processing can influence encoding and retrieval of memories, whereas simulating possible episodes is thought to help rehearse affective responses. One possible avenue might be the investigation of the influence of different forms of simulation on affective processing (e.g., memory retrieval from a field or an observer perspective; Berntsen & Rubin 2006), and further on one’s awareness of one’s future and past (Wilson & Ross 2003): How can this be accounted for by the principle of prediction error minimization? Does the simulation of potential affective responses optimize prediction and reduce potential error in the future? The simulation and integration of future potential changes into the model of one’s autobiography is thought to potentially contribute to the prevention of dramatic changes in one’s model at higher levels, and to maintain mental autobiographies that are more consistent across time.
Acknowledgments
I am grateful to Thomas Metzinger and Jennifer M. Windt, as well as two reviewers, for their critical and constructive comments.