Before I make some comments, I want to highlight the original aspects of her account and show how it can explain how experience acquires a quality of “mineness” or “sense of presence”, that is of belonging to a self. We can then use the predictive coding framework to explain how the sense of presence can go missing. Loss of the sense of presence signals a prediction error which then requires a higher-level system to build a predictive model that fits that error.
The first point to note is that on the most radical interpretation of predictive coding ideas the veridicality of representation is a corollary of cognition not its primary goal. The primary goal of a cognitive system is to predict its own informational states consequent on its actions (broadly construed to include internal regulatory actions). The point is not just that the objects of experience are constructed and hence may be illusory or misrepresented. Rather veridicality of experience is secondary to the accuracy with which cognitive process predicts the flow of information in sensory systems. As she says in the case of perception this means that “instead of aiming to answer the question ‘what is this?’ perception studies should answer the question ‘… what does this resemble?’ ”(Lin this collection, p. 6). This formulation captures the idea that the visual system, for example, is not passively registering retinal information and constructing a representation of the external world, but using a model which predicts the flow of information coming from the retina.
The first step is to apply the same idea to interoception. We see that the mind is not passively registering changes in body state and constructing a model of the body accordingly but predicting the flow of bodily information in cognitive context. Those contexts range from maintenance of homeostasis to the use of affective experience to inform decision-making and reflective cognition. Thus when I think about the past or future these episodes of retrospection or prospection are infused with affective significance.
The radical import for the understanding of pathologies of self-representation is very elegantly brought out by her discussion. Ying-Tung Lin in effect argues that the experience of the self in autobiographical episodes is no more direct than experience of the world in perception or of past events in memory. In each case no object is directly represented or experienced. Rather the relevant object in each case (object of perception, remembered event, or self in the case of first person awareness) is inferred as a part of a process of optimizing predictive accuracy in specific cognitive contexts.
As many have argued the role of the Anterior Insular Cortex (AIC) is to integrate and represent affective information: i.e., those bodily states, which tell the organism how it is faring in the world, actual, imagine or remembered. The point to recall from Ying-Tung Lin’s account is that the AIC is not representing a self but constructing and optimizing a model that predicts the flow of affectively-charged bodily information.
This is why when AIC is hypoactive the subject feels a loss of subjective presence, reported as depersonalization. In particular the patient has a loss of subjective presence for her own body: she registers changes in body state but they do not feel affectively significant for her. Because that lack of feeling is not predicted she then reports it in the vocabulary of DPD.
Why does the DPD patient not proceed to something like the Cotard delusion? According to Ying-Tung Lin whether a delusion is formed depends on the degree of precision assigned to the information produced by hypoactivity in the AIC.
In the case of Cotard delusion developed from depersonalization, when one has the expectation of high precision, the system tends to be driven by the bottom-up predictive error of unexpected hypoactivity of the AIC, rather than the prior model. One is, therefore, more likely to revise the model in order to explain away the surprisal resulting from the mismatch between the actual and predicted activation level of the AIC; that is, the systems of patients suffering from CD are driven by an urge to modify their top-down predictive models in order to con- form to the loss of AIC activity. The construction of the model in CD is considered an attempt to minimize prediction error.