Depersonalisation Disorder (DPD) is characterized by “alteration in the perception or experience of the self so that one feels detached from and as if one is an outside observer of one’s own mental processes” (American Psychiatric Association 2000). Critchley points out that DPD is often accompanied by alexithymia, a condition in which conscious awareness of emotional states is compromised or absent. This is consistent with findings summarized by Medford that “de-affectualisation”, a reduction or absence of affective response, presents as a core feature of clinical cases. Depersonalisation is a separate disorder to derealisation (the feeling that the world is inanimate or unreal) but derealisation is often an important aspect of depersonalisation. Indeed, as Medford describes their relationship, depersonalisation can sometime be a response to derealisation (Sierra et al. 2002; Hunter et al. 2004).
Sethetal. (2011, p. 9; my emphasis) summarize a range of findings about DPD as follows: “In short, DPD can be summarized as a psychiatric condition marked by the selective diminution of the subjective reality of the self and world”. They explain this diminution as the result of the loss of “sense of presence”, the feeling of being engaged in experience. This is what they mean by subjective reality: the condition is not like an hallucination or delusion in which objective reality is misrepresented by faulty perception or belief fixation. In fact the patient correct represents “objective reality” but loses the sense of herself as the subject of experience.
In the attempt to explain the loss of the sense of presence cognitive neuroscience has developed a theoretical picture that considerably augments older theories. On those older theories DPD represented a suppression or inhibition of emotion as a response to trauma or distress. On this view DPD activates mechanisms which might in other circumstances be adaptive. For example, if the subject of violent attack deactivated those mechanisms which produce the experience of distress that would qualify as an adaptive response to trauma. Of course such a response is only adaptive in the short term. Inability to feel distress might also reduce avoidance behavior with disastrous consequences.
It seems that the deactivation is accomplished by inhibitory activity in the Ventrolateral Prefrontal Cortex (VLPFC). The VLPFC is a structure which plays a crucial role in the regulation of affective feeling, especially as part of a process of reappraisal (Füstös et al. 2013). The adaptive aspect here is that it allows the subject to redirect attention and divert cognitive resources to alternative interpretations of self-relevance and response behaviour by inhibiting an experience that would otherwise monopolise cognition. This role has been tested in tasks which involve the top down regulation of negative affect but, as Medford says, “In DPD such suppression is apparently involuntary (and largely resistant to volitional control), but it is reasonable to suppose that this will nevertheless engage similar inhibitory networks” (2012, p. 142). Thus the patient with DPD experiences the result of involuntary deactivation of systems that produce the bodily experience of emotion.
These ideas are consistent with the evidence from cognitive neuroscience about other primary neural correlates of DPD. Hyperactivity in VLPFC leads to hypoactivity in the Anterior Insular Cortex (AIC). That reduced activity in the AIC produces the loss of a sense of presence. This hypothesis results from findings that it has a primary role in higher order representation of interoceptive (visceral, autonomic, bodily) states. It generates the bodily feelings that signal how we are faring in the world moment to moment consequent on affective processing. Activity in the AIC produces what Damasio called the “core self” and what Critchley calls “the sense of presence”. As Critchley says,
evidence from a variety of sources converges to suggest a representation of autonomic and visceral responses within anterior insula cortex, where, particularly on the right side, this information is accessible to conscious awareness, influencing emotional feelings (2005, p. 162).
When Damasio made his contributions to the neurophilosophy of emotions and self-representation the computational theory in the field was less developed so that we can now make some additional observations about the role of the AIC.
To do so we first reiterate the distinction between being able to sense body state, which is the phenomenon baptized by Damasio interoception, (to distinguish it from exteroception [perception of the external world]), and sensing states of a self. The distinction is a subtle one of course but we can approach it intuitively by noting that there is a crucial difference between being able to sense heart rate, blood pressure or temperature as part of an illness and as part of an emotional episode. We observed earlier that the second kind of awareness is the one we describe as self-awareness in virtue of the fact that it reflects affective processing rather than pure bodily regulation. There is a difference in feeling state caused by raises in blood pressure generated by walking up stairs and by heated argument. This is so even though heart rate is heart rate, however caused. But the point of affective processing, as we saw, is to assess the self-relevance of unpredicted changes in things like heart rate and to indicate to the subject how and why they might matter in the cognitive context.
The experiential differences between heart rate per se and heart rate consequent on affective processing can be explained in terms of the principle of hierarchical computational organization, reflected in cortical organization (Craig 2009, 2010; Dunn et al. 2010). The insular cortex is hierarchically organized to map body state at different levels of abstraction and integration. Posterior sections map body state directly and integrate those representations to co-ordinate reflexive regulatory functions. Thus the Posterior Insular Cortex (PIC), for example, represents things like blood pressure and departures from homeostasis and integrates that information to enable reflexive regulatory processing. More anterior regions re-represent and integrate this information in formats available for higher levels of cognitive control. If we sense raised blood pressure the PIC is primary in the representation of that information. When, however, we are deciding how to respond, we need to integrate that information with current and long term goals, representations of contextual information, memory, planning and inference. We may have to inhibit or reprogram automatic behavioral tendencies (not punch the boss) and perhaps reappraise the situation. Thus we need a way, not just to feel raised blood pressure, but to feel its significance in order to program a suitable response. This is the role of the AIC.
This explains a recent finding which seems paradoxical on the “somatic” James-Lange view of emotions revived by Damasio. On that view emotions are representations of body state simpliciter. The feeling of fear is the feeling of being primed to take avoidance action, for example. Michal and collaborators compared the “interoceptive accuracy”, that is ability of patients to judge body state (using heart rate as a proxy), of patients with DPD to normal patients. Strikingly they found that “[there] was no correlation of the severity of ‘anomalous body experiences’ and depersonalization with measures of interoceptive accuracy.” They explained this finding as follows: “[The] findings highlight a striking discrepancy of normal interoception with overwhelming experiences of disembodiment in DPD. This may reflect difficulties of DPD patients to integrate their visceral and bodily perceptions into a sense of their selves” (Michal & Reuchlein 2014, p. 1; my emphasis).
The AIC can only integrate currently available bodily feeling. As Craig says, it “represents the sentient self at one moment of time [and] provides the basis for the continuity of subjective emotional awareness in a finite present” (2009, p. 67). However we can extend the temporal range of information represented by those feelings by integrating them with representation of past and future episodes of experience and/or semantic knowledge. Simulations involved in planning and episodic memory are associated with activation of the AIC to provide sense of extended self. In other words it is the integration of the metarepresentations of body state produced by the AIC with representations of episodes of a temporally extended autobiography that produces the feeling that we are a self with a past and future, rather than a series of disconnected selves, moment to moment.
Nothing in what I have said refutes skepticism about the self, or episodic theories of first person experience (Strawson 2004). It is in fact consistent with the idea that experience is a series of episodes. Whether we feel those episodes are ours depends on how they are integrated. There is no suggestion that everyone integrates them the same way or that integration evokes an equally strong sense of presence in each person. All I have suggested is that there are mechanisms which can create self-awareness moment to moment and mechanisms which integrate those moments of self-awareness with higher level forms of cognitive control that represent past and future actions and outcomes in order for the organism to assess the self-relevance of actual and potential actions. The explanation of awareness of self-relevance in different contexts is a sufficient explanation of the phenomenon of self-awareness that was our initial quarry.
Craig adds a subtle but important qualification to this account. He (and others) remind us that if the predictive coding account of the mind is correct then we are never directly aware of objects, including the body (Craig 2009, 2010; Seth et al. 2011; Garfinkel & Critchley 2013). Rather representations of objects are computed on the basis of discrepancy between their predicted informational effects on us and actual incoming information. It is fluctuations and discrepancies measured against expectations computed at different levels in the control hierarchy that determine the information that becomes consciously available. “An expected event does not need to be explicitly represented or communicated to higher cortical areas which have processed all of its relevant features prior to its occurrence” (Bubic et al. 2010, p. 10; Clark 2013, p. 199; my emphasis.)
The same should be true of neural activation in the AIC, and hence of moments of self-awareness. We are aware of what is relevant to us via unpredicted changes in bodily feeling consequent on affective processing.
This latter feature is the key to understanding the link between “de-affectualisation”, as Medford called it, and depersonalization (Medford 2012). It is not the fact that affect is suppressed that matters, but that affect which was predicted to occur does not in virtue of the involuntary inhibition of the AIC by the VLPFC. When people engage in voluntary or effortful inhibition of affect they do not feel depersonalized. We noted earlier the role of expectation in post-natal depression, but there the expectation is of affective response to a specific object, a baby. In depersonalization it seems that almost all expected affective feelings are absent because of hyperactivity in the VLPFC.
The predictive coding framework also allows us to finesse explanations of the role of anxiety in the experience of derealisation. We noted that Cotard described anxiety as part of the aetiology of the depersonalization experience in Cotard delusion. Medford, in an early discussion of DPD, also postulated a role for anxiety in order to explain an apparent paradox of DPD: the distress experienced by the patient at the absence of affective response. It is not merely that the patient has no emotions, but, as a patient of Medford’s said, “I don’t have any emotions—it makes me so unhappy.” Medford (2012) pointed out that this is only slightly paradoxical: the distress is at the lack of internal affect, the inability to feel rather than at the derealisation of the external world. Medford related this specifically to the anxiety component of the syndrome. The patient expects that the world will induce positive affect but when it does not an expectation is violated and the patient anxiously attends to that absence of affect. On this view highly anxious patients are hyperattentive to their experience and encounter, not the normal bodily experience, which represents how they are faring in the world, but a strange absence of such experience, in combination with intact exteroception which tells them that the world is unchanged (Paulus & Stein 2010; Garfinkel & Critchley 2013; Seth 2013; Terasawa et al. 2013). Their problem is that they no longer feel the relevance of changes in their own bodies and the world to themselves and this inability to feel the world increases their anxiety. Medford quotes an earlier theorist (Ackner 1954) who noted “increased responsiveness for anxiety of internal origin, whereas that of external origin [is] reduced” (Medford 2012, p. 141).
This perhaps explains the differences in casual aetiology between depersonalisation arising in the Cotard syndrome and in DPD. In the Cotard syndrome something is amiss with the mechanisms that appraise perceptual and interoceptive information for self-relevance. The AIC is not getting any information from affective systems to integrate and relay to higher order cognition. Thus felt significance disappears. When the depressive patient then focuses on her experience she feels alienated from the world and depersonalised. In the case of DPD it appears that the AIC is hypoactive for another reason: its activity is inhibited by the VLPFC.
In both cases the patient attends to her experience and tries to interpret it in order to respond. This is consistent with the role postulated by predictive coding theories for attention: the attempt to interpret and sharpen the informational content of a signal by improving the signal to noise ratio. Unfortunately an increase in attention does not provide any increase in precision, it only makes the absence of predicted response more salient. Since those predictions are, in effect, representations of expected self-relevance that normally provide the experience of self-awareness, the patient concludes that the self does not exist. After all, the information necessary to generate self-awareness is still in place. The body, the world and first order representations of their interaction are all represented in experience. What is lost is a sense of the significance of those interactions for the body that mediates them.
The explanation has become complicated so at this point it is useful to situate it in terms of the conceptual architecture (points (i)-(iii) below) outlined in the introduction. On this view DPD arises in the following way as a personal level response to the absence of predicted affective experience.
Appraisal systems normally represent the significance of information for the organism. The primary way of experiencing the result of those appraisals is via activation in the AIC. This is because the AIC is specialised for informing the subject, via bodily experience, of the affective significance of its encounters with the world. These experiences are not the same as experience of body state per se but the emotional significance of that body state.
Those experiences can be rehearsed offline in planning and deliberation to extend the temporal horizon of affective experience. We feel like temporally integrated selves because memory and prospection have affective significance.
Predictive coding architecture has the effect of making discrepancy between anticipated and actual affective feeling highly salient.
In DPD activity in the AIC is inhibited most likely as a result of the involuntary activity of the VLPFC.
Consequently the patient has normal perceptual and sensory responses to the world but those responses are not integrated into a bodily representation which informs her of their significance. The world feels derealised or as Medford puts it de-affectualised
However, given the way predictive coding works, the patient actually has a model of the world that predicts activity in the AIC as a result of her perceptual encounters. Thus absence of AIC-produced experience is a prediction error that drives metacognitive responses.
Those responses include increased attention (driven by sub personal mechanisms of resource allocation) to the experience itself as the patient tries to extract further information from it. However, being produced by subpersonal mechanisms the experience is both intractable and inscrutable.
Highly anxious people cannot divert attention from the experience, since anxiety is driven by the need to resolve uncertainty. But the experience is inexplicable and irresolvable.
The patient’s personal level interpretation of the experience is of depersonalisation “it feels like it is not happening to me”. The interpretation is not a direct report of the experience, which I have argued is more like a total deaffectualisation. It amplifies it.
However the form that amplification takes, depersonalisation, is explained by the role such experiences have in creating the normal sense of being a self. We feel we are selves precisely because the significance of the world for our organismic goals is normally computed by appraisal systems and represented in characteristic forms of bodily experience.