Disintegrating and Reintegrating the Self - (In)Flexible Self-Models in Depersonalisation and Psychedelic Experiences

This preprint paper (2022) explores the contrast between the feeling of ‘losing’ the sense of familiarity with one’s self and body in Depersonalisation experiences (DP) and psychedelics (with some consideration of meditative experiences) using the lens of the Active Inference Framework (AIF). It is suggested that such experiences can involve a stance with radically altered prior expectations, so providing opportunities for flexibly modulating self-and world models and that controlled acquisition of new self-and world models may enhance the plasticity of one’s perceptual and sensorimotor experiences.

Authors

Ciauncia, A.
Safron, A.

Published

March 13, 2022

Psyarxiv

meta Study

Links

Read Paper DOI Google Scholar

Abstract

Across times and cultures, humans constantly and intentionally tried to ‘lose’ or to ‘escape’ their familiar, ordinary self, to ‘self-detach’ and to radically change the ways of perceiving oneself and the world. In this paper, we explore the contrast between the feeling of ‘losing’ the sense of familiarity with one’s self and body in Depersonalisation experiences (DP) and psychedelics (with some consideration of meditative experiences). We explore these radical changes in self-experiences through the lens of the Active Inference Framework (AIF). AIF is a process theory aiming to capture the capacity of biological organisms (e.g. living human bodies) to survive and thrive in volatile and uncertain environments. In line with previous work on depersonalisation and psychedelic mechanisms, we suggest that such experiences can involve a stance with radically altered prior expectations, so providing opportunities for flexibly modulating self-and world models. Specifically, we suggest that controlled acquisition of new self-and world models may enhance the plasticity of one’s perceptual and sensorimotor experiences. This newly gained flexibility, we claim, may allow the individual to ‘leave behind’ certain habits, perceptual rigidities that holds him/her ‘stuck’ in certain behavioural patterns. And to open to new ways of perceiving and integrating self-and world-related information. By contrast, depersonalisation experiences point to an uncontrolled phenomenon of non-flexible (rigid) (dis)integration of ordinary/habitual self-models, and a consequent feeling of being ‘stuck’ in one’s mind. While controlled (dis)integration of habitual self-experiences and consequent re-integration may have positive effects, uncontrolled (dis)integration of habitual self-experiences triggered by unpredictable life events may be overwhelming and lead to self-detachment and potentially adverse clinical outcomes. Contrasting these two modes of alteration will allow us to outline the importance of the controlled ability to flexibly integrate, disintegrate and reintegrate multisensory bodily signals, and its impact on the human sense of self and agency.

Unlocked with Blossom Pro

Research Summary of 'Disintegrating and Reintegrating the Self - (In)Flexible Self-Models in Depersonalisation and Psychedelic Experiences'

Introduction

The paper situates the sense of a familiar, embodied self within a large literature in philosophy, psychology and neuroscience that emphasises multisensory integration as the scaffold of selfhood. Everyday experiences of bodily familiarity and agency are contrasted with culturally widespread practices and technologies that intentionally disrupt ordinary self-experience (for example, meditation, dance, virtual reality or psychedelics), as well as with cases in which radical changes in self- and world-perception occur involuntarily (for example following trauma or severe stress). The authors foreground depersonalisation (DP) as a clinical phenomenon in which individuals feel estranged from their bodies and selves and note that forms of DP-like experience can also arise during psychedelic and meditative states, sometimes therapeutically and sometimes harmfully. Ciaunica and Safron set out to compare and contrast depersonalisation and psychedelic experiences through the conceptual lens of the Active Inference Framework (AIF). They propose that both classes of altered self-experience can be understood in terms of changes to hierarchical predictive models and to the precision-weighting that balances prior expectations versus incoming sensory evidence. The paper aims to show how controlled, flexible (dis)integration and reintegration of self-models (as may occur in guided psychedelic or meditative contexts) differs from the uncontrolled, rigid disintegration characteristic of pathological depersonalisation, and why that distinction matters for wellbeing and clinical outcomes.

Methods

This is a theoretical and integrative paper rather than an empirical study. The authors develop their argument by synthesising conceptual work and empirical findings from multiple domains (predictive processing/active inference theory, neuroimaging, neurophysiology, phenomenology, and clinical descriptions) to build a mechanistic account of altered self-experience. The core methodological approach is conceptual: AIF and Predictive Processing are used as the organising framework to interpret prior literature on interoception, exteroception, precision-weighting, large-scale brain networks (salience network, default mode network, frontoparietal control network) and clinical phenomenology of DP and psychedelic states. The extracted text does not report systematic review procedures, databases searched, eligibility criteria, or quantitative meta-analytic methods, and no original primary data collection, experimental protocol, or statistical analyses are described. Where neuroimaging and psychopharmacological findings are cited, they are integrated illustratively to support model components rather than presented as results from a prespecified empirical method.

Results

Working from AIF/PP, the authors elaborate a set of linked theoretical claims. First, selfhood emerges from hierarchical generative models that predict multisensory inputs (interoceptive, exteroceptive and proprioceptive) and minimise prediction error through perceptual updating and action. A central computational variable is precision (inverse variance), which determines the relative influence of top-down priors versus bottom-up prediction errors. Flexible, context-sensitive precision-weighting allows adaptive switching between internally- and externally-directed processing and supports skilful embodied agency. They argue that depersonalisation can be modelled as a pathological disruption of precision-weighting such that self-models become opaque and over-attended. This excessive self-objectification or ‘‘self-grasping’’ moves the self from a transparent background to a disturbing foreground, producing classic DP symptoms (feeling like a stranger to oneself, loss of agency, disembodiment, and somatosensory distortions). The paper reproduces Sierra's taxonomy of four anomalous body experiences in DP: lack of body ownership, loss of agency, disembodiment, and somatosensory distortions. Clinical and physiological correlates mentioned include altered autonomic and somatic responses, vestibular-system involvement, and disrupted activity in retrosplenial/posterior cingulate regions implicated in egocentric perspective. Population figures cited in the text indicate a DP prevalence of around 1-2% and transient depersonalisation episodes in 34–70% of the general population. By contrast, psychedelics are characterised in the paper by models that reduce top-down filtering and/or relax prior precision, thereby enabling novel combinations of inferences and greater updating from prediction errors. The authors review three related mechanistic proposals: thalamic gating (reduced sensory filtering), REBUS (‘‘RElaxed Beliefs Under pSychedelics’’; relaxation of high-level priors), and ALBUS (a broader account in which 5-HT2A receptor stimulation may either relax or strengthen beliefs depending on dose, set and setting). Empirical signatures used to support these ideas include reports of increased neural entropy/complexity, altered large-scale network dynamics, connectome-harmonic changes, and increases in measures such as fractal dimension. The authors note empirical findings from studies of psilocybin and dissociatives (dextromethorphan) that are consistent with enhanced network and psychological flexibility, while also noting at least one report of reduced psychological flexibility following lysergic acid diethylamide. A central conceptual outcome is the distinction between flexible, controllable (dis)integration and reintegration of self-models (which the authors ascribe to therapeutic forms of psychedelic or meditative experience) and rigid, uncontrolled disintegration (as in pathological DP). The paper introduces the idea of ‘‘cohesive flexibility’’ — a dynamic network property indexing the capacity to reorganise connectomic community structure coherently over time — as a candidate bridge between momentary state changes and longer-term trait-like adaptations. The authors also speculate about potential neural ‘‘hubs’’ (posterior cingulate, precuneus, temporoparietal junction, dorsal anterior cingulate, dorsomedial and dorsolateral prefrontal cortices) whose modulation might mediate shifts between adaptive and maladaptive self-experience.

Discussion

Ciaunica and Safron interpret their synthesis to mean that altering self-models can be either adaptive or maladaptive depending on control, context and the dynamics of precision-weighting. Controlled deviations from habitual self-world models — for example in carefully guided psychedelic or meditative contexts — may temporarily relax rigid priors and increase multisensory and inferential plasticity, permitting people to ‘‘leave behind’’ maladaptive habits and discover alternative ways of perceiving and acting. Conversely, traumatic or unpredictable life events may provoke uncontrolled, rigid disintegration of self-models that leaves individuals ‘‘stuck’’ and leads to chronic depersonalisation and impaired social and affective functioning. They position these claims relative to earlier work by linking phenomenology (self-opacity, self-objectification) to specific computational mechanisms (precision misallocation) and to putative neurobiological substrates (salience, DMN, posterior cingulate cortices and vestibular pathways). The authors acknowledge important uncertainties: several neural-level claims remain speculative and the extracted text explicitly notes insufficient evidence to identify which specific neural systems are most relevant for therapeutic ‘‘resets’’ under psychedelics. They also highlight mixed empirical results (for example, a study reporting decreased psychological flexibility with LSD), and caution that a single construct of ‘‘flexibility’’ may manifest differently across biological and psychological levels. In terms of implications, the paper advocates for nuance: psychedelic and meditative interventions might be harnessed to increase adaptive flexibility, but require attention to individual characteristics, dosing, and ‘‘set and setting’’. The authors suggest the value of precision psychiatry approaches that target particular neural systems and tailor interventions, and they emphasise the ethical and clinical importance of preserving agency and providing skillful integration to avoid harmful, uncontrolled depersonalisation. Limitations acknowledged in the text include the conceptual and speculative nature of the synthesis and the need for empirical research to test the proposed mechanisms and to clarify boundary conditions for benefit versus harm.

Conclusion

The conclusion reiterates the central claim that one does not simply ‘‘lose’’ the self but can radically change its organisation in ways that are either controlled and self-strengthening or uncontrolled and self-undermining. Acquisition of new self- and world-models may increase perceptual and sensorimotor plasticity and enable people to leave behind maladaptive rigidities, whereas uncontrolled (dis)integration — as in depersonalisation — can be overwhelming and lead to chronic clinical harm. Maintaining a flexible balance between familiar/exploitative and exploratory/unfamiliar modes of self- and world-modelling, together with sufficient agency over those transitions, is presented as crucial for healthy personhood and wellbeing.

View full paper sections

INTRODUCTION

A robust body of work in philosophy, psychology and neuroscience outlined that integration of the multisensory bodily inputs scaffolds our sense of self, i.e. the subjective experience of a familiar, or ordinary 'self' or 'I', bound to my body and distinct from the world and others; Park and Blake 2019). For example, when you wake up in the morning, you expect to see your hands having a certain shape, size and colour. You also expect when you look in the mirror to see a face that is 'familiar' to you: your face. Everyday experience also seems to involve experiences of agency; namely, the feeling that 'I am in control of my own bodily actions, that I can leverage them to access and change the external world. Across times and cultures however, humans constantly and intentionally tried to 'lose' or to 'escape' their familiar, ordinary self, to 'self-detach' and to radically change the ways of perceiving oneself and the world. The so-called "altered" or non-ordinary states of consciousness involve dramatic changes in sensory, time, and space perception, leading to novel perspectives on familiar phenomena. For example, some people engage in training with intense contemplative practices with the intention of 'losing' or 'leaving behind' their daily selves. Others make use of intense bodily movements and dance techniques, creative activities, or even virtual realityin order to drastically modify the perception of their own bodies and the surroundings. Others attempt to use various chemical substances in order to achieve such modifications of experienced realities for recreational, personal/spiritual, and sometimes professional purposes. These are cases where people actively seek to alter their self-experiences and voluntarily engage in behaviours designed to radically change their ordinary senses of self and world perception. Indeed, novelty seeking and curiosity-based behaviours may provide significant evolutionary advantages over rigid and limited ones. However, there are also cases where these radical changes in self and world perception occur without the deliberate will of the individual (e.g. traumatic life events, exposure to severe stress, depression or physical and mental illness). In this paper we explore the contrast between the feeling of 'losing' the sense of familiarity with one's self and body in Depersonalisation experiences (DP) and psychedelics (with some consideration of meditative experiences). As we will see below, DP is a condition that makes people feel like 'strangers' to themselves, unfamiliar and detached from their ordinary self, body and the world. Perhaps surprisingly, DP-like altered states (as well as their opposites) are not uncommonly sought out as part of psychedelic experiences, while also being considered to be a source of potentially adverse outcomes associated with psychedelic use. The ability of psychedelics to so radically alter fundamental processes of selfhood speaks to their transformative potential, allowing individual to see/experience the world differently, potentially teaching them how to "change their minds" (and lives). In what follows, we explore these issues through the lens of the Active Inference Framework (AIF). As we explain in greater detail below, AIF is a theory aiming to characterize the processes by which biological organisms (e.g. living human bodies) survive and thrive in volatile and uncertain environments. In line with previous suggestions regarding (i) DP mechanisms) and (ii) psychedelic mechanisms, we suggest that such experiences can involve a stance with radically altered prior expectations, so providing opportunities for moreor-less flexible modulation of models for the self and world. Indeed, enhanced flexibility may be a primary mediator of positive outcomes associated with psychedelic psychotherapy. The inductions of controlled deviations from normal (or "default") modes of operation may alter the typical multisensory integration of bodily signals underlying our ordinary self-and worldexperiences. Such embodied processes may constitute the core (or roots) of the generative models that control (or constitute) individuals, which when altered may lead, in turn, to the subjective phenomenology of 'losing' 'familiar'-a potentially strange word for describing DP states in which everything might feel unfamiliar or "unreal" -ways of perceiving oneself and the world, so giving rise to non-ordinary states of consciousness. Carefully controlled deviations from normal/default biophysical functioning may enhance flexibility on multiple levels, ranging from psychological processes to neural dynamics, potentially 'tuning' body-brain-mind towards a regime that is more conducive to playful exploration and open-ended evolution. The fact that these changes are intentional (e.g. sense of agency intact) may further contribute to the feeling of widening the horizon of one's experiences and ensuing increases in wellbeing. In depersonalisation experiences however, individuals are unintentionally faced with high levels of uncertainty and consequent radical changes in their self-and world-models. These dramatic alterations with respect to the multisensory integration of bodily signals can also have the effect of causing deviations from ordinary ways of perceiving oneself and the world. In the following discussion, contrasting these two modes of alteration will allow us to outline the importance of the controlled ability to flexibly integrate, disintegrate, and reintegrate multisensory bodily signals, and their impacts on sense of agentic selfhood. Indeed, such flexible transitions between modes of person/selfstructured states of mind appears to be a core aspect of normative functioning with respect to dynamic brain networksin humans, potentially in all mammals (e.g. mice), and perhaps even all life. This flexibility, we suggest, subserves the process of self-regulation via homeostatic and allostatic control within a volatile environment in which all is subject to constant change. Since this information is necessarily greater than the computing capacity of the human brain (particularly with respect to environmental changes for which prediction is not possible), higher flexibility allows better adaptability to unforeseen events. Higher flexibility of self-and world models enhances novelty-seeking and curiosity-driven exploratory behaviours, which in turn, gradually expand personal comfort zones and wellbeing. Importantly, it is not only the world that constantly changes, but the body as well. Through various organs such as our mouths, skin, and lungs, we literally change our cells every moment of our lives as we continually exchange matter and energy (and information) with our environments. We feel that we are the same 'old' or 'familiar' self, yet our body is never exactly the same as it was one second ago. One can never step into the same stream of experience twice. Yet Seth and colleagues have compellingly described insensitivity to such alterations as a kind of "change blindness" that may be an essential component of the phenomenology of selfhood, or "being you". However, while such self-coarse-graining could be understood as a source of necessary stability for selfhood, it may also result in situations in which individuals find themselves stuck in modes of being with inflexible maladaptive self-processes, as observed with DP (Sierra 2019; Perkins 2021). Below we will describe some of the neurophenomenology of DP, and also discuss ways in which altered psychedelic and meditative states may allow for the cultivation of new states (and potentially traits) of mind in which agency and aliveness are rediscovered/created. In Section 1, we briefly introduce the conceptual toolbox concept of self-models, minimal self, allostasis as control. In Section 2, we extend these discussions to consider DP. In Section 3, we discuss how psychedelics and meditative experiences may allow such processes to be altered in what may be either therapeutic or pathological directions. In section 4, we speculate on ways that such changes may be reflected in terms of more flexible (or dynamic) brain networks. Finally, we will consider these phenomena through the lens of the philosophy of intersubjectivity, showing how such visceral experiences may help individuals to get "out of their heads, and into their [shared] lives".

(IN)FLEXIBLE SELF MODELS THROUGH AN ACTIVE INFERENCE LENS

Bodily self-preservation and reproduction are the centre preoccupation of all living organisms. To achieve this vital task, like any other living system, human bodies and brains needs to reach two targets "with one stone" (so to speak). On the one hand, sensory processing needs to be stable and self-centred enough to keep my body within a given range for survival. On the other hand, information processing needs to be flexible and worldoriented enough to allow self-centered organisms (e.g. me and you) to engage in active behaviours necessary for survival. We do not self-produce our own energy, and hence we cannot survive in isolation without resources from our surroundings, including from others. As such, self-preservation crucially depends on our ability to constantly move both internally (e.g. my heart needs to keep beating); and externally (e.g. inaction and sedentarism may atrophy my muscles and impair capacities for obtaining food). Since the moment we are born, and even before birth in the womb, our bodies move constantly. Bodily movements can be either autonomic, phenotypebased hardwired (e.g. heartbeats, breathing, reflexes) or intentional and goal-oriented (e.g. we may open our hands or engage in food-seeking behaviours). To ensure the survival of the body, the brain continuously processes and integrates a cascade of sensory signals coming from inside and outside our bodies, as well as from our surroundings. Interoception typically refers to the processing of sensory signals from innervated internal organs (e.g. heart), but also applies to nociceptive, temperature and C-tactile afferent mediated (affective touch). Recently, the term has been broadened to include signals from the endocrine system (Chen et al. 2021), the immune system), and digestive system (de. All these interoceptive signals can be addressed under the umbrella of "visceral" signals. Exteroception designates the processing of sensory signal from outside the body, while proprioception refers to the signals indicating the position of the static and dynamic body in space. Importantly, while the perception of some exteroceptive inputs can be intentionally blocked so as to be ignored (e.g. we wear headphones to block auditory inputs from outside the body), signals coming from inside the body need to be constantly treated and processed by our brains. At the level of explicit conscious awareness, we may not necessarily pay attention to our heartbeats or breathing movements (although we can reflectively choose to do so, for example in meditation). Yet, at the implicit pre-reflective level, the brain needs to "pay attention" to these signals, even when we are asleep. The question of the nature and development of implicit or pre-reflective forms of selfawareness, especially in relation to their bodily foundations, has attracted a significant number of empirical studies and theoretical accounts. Indeed, a growing number of authors defend several versions of the "minimal mineness" thesis, i.e. the idea that our everyday phenomenology is characterized by a prereflective sense of self, referred to as the "minimal" or the "core" self. Despite different perspective on the nature of the minimal self, there is overall agreement over the basic idea that actions in the world and movement of the body subserves regulation of bodily states, keeping them within an optimal range. How exactly our brain performs this delicate balancing act(ive inference) between self-centred regulation and world-oriented exploratory perception is an area of active investigation. Some evidence suggests that the salience network (SN) may help to guide degrees of internally-and/or externally-directed processing via orchestrating degrees of predominance for respective default mode networks (DMN) and frontoparietal control networks (FPN). This sort of capacity for intelligent switching is essential for being able to adaptively move through the world in ways that harmonize internal goals with external contexts, with this flexible orchestration of inner and outer dynamics applying to phenomena ranging from skillful motor control to our abilities to connect with others through shared activities and over abstract meanings. While specific implementational details for this kind of vital self-control continue to be explored, the Active Inference Framework (AIF) and Predictive Processing (PP) represent increasingly influential models for both philosophy and the mind sciences. AIF and PP have been heralded as central to understanding how alterations in low-level bodily sensory processes could drive the emergence of high-level mental phenomena. According to PP, the brain processes hierarchical and dynamically shifting generative models of what is causing incoming sensory events, based on Bayesian probabilistic "guesses" or beliefs about the likely causes of sensory information (exteroceptive, interoceptive, proprioceptive). These signals are matched against either learned or innate patterns (or priors), constituting what are called "predictions" about typical sensory inputs, given "the kind of creature that I am". When predictions (or Bayesian prior 'expectations') do not match ongoing sensory inputs, then "prediction-errors" result, which have the effect of updating priors/predictions. Thus, the brain constantly and dynamically adjusts its prior expectations to minimise overall prediction error in the context of regulating life management via both perceptual updating and actively changing relationships with the world (hence, active inference). Such predictions are continuously constructed and updated moment-by-moment at multiple levels of hierarchical processing by the generative models of the brain-body-environment system, the sum-total of which we may understand as the biophysical (and informational) realization of adaptive intelligence. In these ways, AIF provides both a descriptive and normative account of what systems must do if they are to persist. The energetic and informational openness of persisting systems is nicely captured in the expression "The first law of psychology is the second law of thermodynamics". The second law holds that closed physical processes tend toward a state of "increasing (statistical) probability and decreasing order, leading to thermodynamic equilibrium". Organisms, in contrast, "maintain themselves in a state of high statistical improbability, of order and organization" far from thermodynamic equilibrium (ibid.). Organisms do this by importing 'order' (energy in the form of matter) from the environment, chemically transforming it via metabolic processes into useable forms of fuel for the synthesis of biologically important molecules, and exporting 'disorder' in the form of various waste products. This continual exchange of matter and energy-which makes organisms thermodynamically-open systems-takes place within an environment in which both internal and external parameters are continually in flux, so necessitating ongoing open-ended learning to assure adaptive functioning. According to AIF, the experience of being a self is the result of ongoing generative modeling in the context of PP that centers on individual organisms. From this point of view, embodied agents act as "self-evidencing systems" in that they act in ways that tend to maximize evidence for their self-models, in which their (extended) phenotypes constitute implicit predictive models (Hohwy 2014). Self-evidencing is formally equivalent to tracking and minimizing the (variational) free-energy (or cumulative precision-weighted prediction errors) generated by discrepancies between predicted and sensed data, which under AIF means predicting ourselves as (fundamentally) embodied, (environmentally) embedded, (functionally) extended, enactive (or interactive/niche-constructive) (and affective) systems. From this perspective, we experience ourselves as being endowed with a persisting self because we reliably infer that the existence of such a self is the most probable (common) cause of our sensory observations. For instance, the fact that my experiences seem to have a zero point or origin that I carry along with me as I move my body is best explained by postulating that there exists a causal factor-a perspectival "I", "me", and other forms of modes of being (or 'selfing') that are ultimately attached to a body and are distinct from other selves-that is experiencing the experience. Crucially, this kind of explanation also functions as an essential-and foundational-"selffulfilling prophecy" in that such self-inferences go on to enable modes of enactment that both serve and are structured by various models of selfhood (and intersubjectivity). This self/world-modelling (S/W-M) is organised in a multi-layered hierarchical fashion. Prior beliefs 1 (empirical) about the self and world are conveyed by the top-down (predictive) descending connections between hierarchical levels. Bottom-up sensory evidence, in contrast, is mediated by ascending connections that pass predictions errors up the hierarchy until they are "explained away" by sufficiently informative (and precise) predictions. These predictions take the form of implicit (and sometimes explicit) Bayesian posterior beliefs, or hypotheses concerned with the causes of sensory input, which at higher hierarchical levels involve the integration and updating of increasingly complex, temporally deep, and counterfactually rich models, with the updating process at every level derived from the moment-to-moment processing of (1) prior beliefs about the self and world and (2) current sensory evidence. A crucial aspect of the process of updating S/W-M depends on the "precision" of (i) prior predictions and (ii) prediction errors caused by incoming sensations. Prior beliefs and sensory data are represented as probability distributions with (a) mean values (expectations) and (b) precisions (inverse variance). Precision captures the confidence of modelled expectations with respect to associated variance parameters. In each case, precision allows generative models to arbitrate how best to balance the relative influences of either prior beliefs or sensory evidence. Crucially for our discussion here, prediction errors that convey precise, reliable, high-quality information are afforded greater precision or weight, such as that they have a greater influence on perception. This equates to "attending to the right sort of newsworthy information and attenuating imprecise or 'fake' news. It is this mechanistic bridge between a fundamental computational imperative to properly balance sensory evidence against (hierarchical) previous expectations and plausible neurobiological mechanisms that links psychopathology and pathophysiology.". Heuristically, if the prediction error emerges from the contrast between precise sensory data and relatively imprecise prior beliefs, the mean of the posterior will be closer to the mean of the sensory data. Conversely, if it is instead the sensory information that is believed to be less reliable, posterior beliefs will be much closer to prior beliefs. This multi-layered system needs to decide 'on the fly' the most skillful balance with respect to assigning precision-weighting (the relative 'gain' of the updating process) for either the (a) sensory evidence coming from various sensory modality, or (b) prior beliefs (or expectations) that the model predicts for that sense data. Context-sensitive and flexible switching of relative influences from prior beliefs or sensory evidence (i.e., prediction errors) is required for reliable inferences (and learning) via adaptively well-tuned "precision weighting." Such dynamic precision adjustments may have a particularly fundamental role in the construction of complex/abstract representations related to selfhood as an integrative process for cybernetic control. Notable work has also significantly broadened the concept of attention as precision weighting to take into account inner bodily states (interoception) as well as dynamic brain-body coupling. In short, self-modelling (and adaptive self-evidencing) in active inference crucially relies on the flexible balance of precision with respect to emphasizing either prior expectations or present sense data, with small differences in this balancing act having potentially non-linearly impactful consequences for overall self-world modeling. Crucially for our discussion here, such S/W-M must necessarily be flexible to account for a world of constant change and limited predictability. But even within a relatively static environment, capacities for flexibly adjusting (or attenuating) the precision of sensory prediction-errors is required for free energy to be minimized via overt enaction, rather than updating of internal models to reflect current estimated system-world configurations (potentially leading to pathological stasis/inaction). However, if allocation of precision weighting to generative models lacks flexibility, then this could engender maladaptive belief dynamics with respect to self and world, as we will now discuss with respect to DP.

LOSING THE 'FAMILIAR' SENSE OF SELF -THE CASE OF DEPERSONALISATION

What happens if the process of precision estimation of a self-model gets disrupted? For example, what happens if the precision weighting balance gets 'stuck' and excessively leans (or tilts) systematically towards one of the two branches of either bottom-up sensory evidence or top-down prior beliefs? Such systems will automatically lose the flexibility afforded by an adaptive modulation to the incoming sensory information, typically associated with optimal gain and successful model updating. More importantly, as we will see below, disruptions of precision-estimation means that systems will afford less of the 'luxury' of forgetting about certain aspect of self-models in ways that allow for "transparent" processing. Aberrant precision of either sensory evidence or expectations will lead the system, as a sideeffect, to 'over-attend' to its own self-modelling which becomes more 'visible' or overly "opaque." To provide an illustrative metaphor framing): if there is a crack in a window, obstructing opacity may occur, rather than the desired transparency by which windows serve their primary function as portals through which one can look. Considerable (potentially distracting and distressing) computational resources are now allocated to deal with these alterations, which result in shifting the self-model from a 'transparent' invisible background to the visible and 'opaque' foreground. The resulting self-model in this case will most likely trigger altered self-experiences and consequently enhanced self-focus, so undermining the epistemic permissibility of selfprocesses as windows onto (or that bring forth) an experienced world for a living being. In brief, well-tuned self-processes should be able to "get out of their own way" and not attract excessive attention to themselves in order to serve their primary functions as high-level control systems whereby persons may adaptively couple with the world on multiple scales across diverse contexts. In these ways, the sense-making capacities provided by self-processes may be so extensive that one could say that experiences arise in "the field of me". When smoothly functioning, these self-processes provide a source of contextualization for perception and action so ubiquitous as to potentially go unnoticed (cf. fish not noticing the water in which they immersed). Yet the very ubiquity of such organization-via-selfhood may end up being the very thing that renders it transparent to introspection, potentially via mechanisms of efficient "explaining away" via hierarchical predictive processing. This may be the most profound of all ways in which we observe a seeming paradox with respect to the experiential invisibility of the most fundamental aspects of being, potentially even contributing to some of the enduring problems of consciousness. However, well-tuned 'selfing' should also involve various kinds of self-reflection (as mental actions) to enable various forms of metacognitive awareness (and more sophisticated control). In this way, adaptive self-control (and self-construction) requires a capacity for flexibly moving between transparency and opacity, with moment-to-moment, context-sensitive adaptive orchestration of dynamics based on (hopefully accurate) predictions of what will allow for skillful and context-sensitive responding. At the experiential level, the process of rendering selfhood opaque may correspond to what phenomenologists call "self-objectification": by allocating extra resources to the processing of its own model, the self related to itself as an object to be controlled and 'grasped', and thereby disrupts the transparency of one's sense of self. Hence the agent may experience herself less as a subject of an experience and more like an object of an experience. Indeed, previous research has illustrated that negative effects such as anxiety and depression "cause somatic symptoms, as well as increasing selffocused attentionthrough general arousaland ruminations)". Indeed, the notion of (hopefully flexible/adaptive) self-grasping speaks to enactivist notions of "optimal grips" and Heideggerian phenomenology. That is, much of skillful engagement with the world involves "ready to hand" situations in which individuals find themselves in a state of effortless mastery of sensorimotor/conceptual contingencies, largely governed by unconscious habits. However, when an individual encounters circumstances of a surprising nature (i.e., "present at hand" situations), then more elaborate modes of being may become involved as individuals try to cope with these challenges and obtain more secure epistemic/pragmatic grips on matters, potentially including selves as control processes. This temporary pulling out from the ongoing flow of experience may be essential for allowing an individual to escape local optima and discover more adaptive ways of being. Indeed, this may be part of the functional significance of threat-sensitive serotonin systems and their wellestablished opponency with dopaminergic signalling, potentially involving increased dominance of more interoceptive (and self-focused) default mode over more externallyoriented frontoparietal control networks. However, if the phenomenal opacity accompanying such self-reflection/gripping interferes with skillful engagement with the world, it could be experienced as a kind of "choking", potentially including with respect to the intersubjective coupling and cooperative sensemaking that constitutes the most meaningful aspects of our lives. These challenges may then go on to trigger further attempts at regripping via self-reflection, potentially giving rise to a vicious cycle of disconnection in which individuals may find themselves "trapped within themselves." Taken to extremes, these positive feedback cycles could potentially trigger various dissociative states as more complex (and hierarchically higher) modeling is compromised, such as the kinds of (active) inferences with temporal depth and counterfactual richness that would be required for maintaining a sense of self as a coherent locus of control and source of influence on (personally meaningful) world dynamics. With respect to meditative practices, this kind of maladaptive disconnection could be a source of "dark night" experiences sometimes identified as potential adverse outcomes for vulnerable individuals. Similar dynamics could be at play for some adverse outcomes associated with psychedelics as well, but as of yet there is insufficient research to make strong claims on this matter. But with respect to both meditation and psychedelics, it is notable that selfless experiences (cf. experiences of "ego death") are often times viewed as one of their main benefits with respect to promoting a potentially desirable, radical shift of perspective on self and world. Most people can alter their experiences by stepping in and out of detached states at will, returning to the 'ordinary' world of shared experiences when needed. Yet, the capacity to separate the self from the present ordinary or 'familiar' experiences can go wrong in multiple ways. Depersonalisation (Sierra & Berrios 1998) is a very common subjective experience that makes people feel as if they are detached and disembodied observers of one's self, body or even of reality itself: "When I'm having an episode of depersonalisation, it feels more like I'm watching myself doing things, but I'm not present for it. I'm witnessing myself… I 'know' I'm in control, but I'm not 'feeling' in control" (Perkins 2021:44). People with DP find themselves trapped in the detached state and unable to return back to their familiar selves whereby they may connect with others, and potentially create even deeper forms of self/other connection through virtuous cycles involving shared valuing. Feelings of DP are the third most common psychological symptom reported, triggered by severe stress, traumatic life events or drug use), and has highly distressing effects on the quality of one's mental and social life (Simeon & Abugel 2006; Perkins 2021). The experience of a 'split' between the self and the body at the basic sensory level-strikingly described as feeling trapped in one's head (mind) and outside one's body-is one of the most frequently cited symptoms in DP. This split is responsible for the ensuing sense of (a) self-detachment, of looking to oneself from the outside, from the "back of one's eye sockets"; and (b) unrealness, often and strikingly reported by DP patients as 'having a pane of glass' interposed between one's self, body and the world. This foregrounding of an illusory point of view may speak to the centrality of coherent egocentric perspective for both somatic selfhood and visuospatial awareness, as has been suggested by theories of consciousness based on projective geometric modeling, attention schemas, and integrated world models. It is also central to intuitions in which (quasi-homuncular) selves view themselves as observers of experiential fields, as if viewing within a kind of (quasi-Cartesian) theatre. DP has a prevalence of around 1-2% in the general population)-roughly comparable to base rates for schizophrenia. Also, the prevalence of transient episodes of depersonalization is between 34 and 70% in the general population. Everyday phenomena such as fatigue, sleep deprivation (van Heugten-van der, or even travelling (or making a 'trip') to new places may also trigger transient depersonalization feelings). As will be discussed in greater detail below, the triggering of depersonalization by intensely stressful episodes suggests potential overlap with mechanisms contributing to psychedelic phenomenology, which have been suggested to correspond to "active coping" responses to challenge, and which may involve major reconfigurations of self and world models with both therapeutic and psychopathological potential.

SIERRA (2009) LISTS FOUR PROMINENT TYPES OF ANOMALOUS BODY EXPERIENCES IN DPD:

(1) lack of body ownership, (2) feelings of loss of agency, (3) disembodiment feelings, and (4) somatosensory distortions. Previous theoretical and empirical work has demonstrated disrupted physiological responses in patients with DP, compared to healthy participants. DP has also been linked to disrupted activity in neuronal regions underlying somatic processing) and the vestibular system, which is responsible for providing information about the body's position in space. Notably, efferent signals from the vestibular apparatus converge with neck-stretch receptor information from the mamillary bodies at the anterolateral thalamic nucleus. This jointly integrated head-direction information is projected to retrosplenial posterior cingulate cortices in helping to establish a coherent egocentric perspective, as well as facilitating flexible transformations between reference frames, such as those that would be involved with adopting the perspectives of others, including counterfactual selfviews. Strikingly, these areas also host the same systems/processes whose alterations have been heavily implicated in self-transcendent experiences from psychedelic and advanced meditative states. However, being a person is more than just having the sense of a body with a point of view on the world, but is inherently intersubjective in involving other persons as feeling beings with lived worlds from their own perspectives. Rather than an unmoving tethering to a singular experiential mode, the experience (and skill) of being a person involves a dynamic motion between various perspectives, so allowing for synergistic inferential power in the intersubjective modeling of self and other. Yet, if an individual becomes fused with a particular point of view, while in some circumstances this could correspond to a "flow"-like deep absorption with experience, in other cases it could lead to the severing of connections to the world, others, and the meanings we may co-create/share. Indeed, in such ways, DP phenomena may not only be of potentially increasing clinical relevance, but it may also speak to the fundamental nature(s) of forms of consciousness related to personhood and embodied agency. For example, building upon the James-Lange and Schachter-Singer (1962) theories of emotion, Seth and colleagues developed an interoceptive predictive coding model of conscious presence according to which disorders of selfhood (such as depersonalisation) arise from pathologically imprecise interoceptive predictive signals (which consequently fail to update predictions, and thereby perpetuate a sense of unreality). In this way, attending to objectified selfhood may outcompete attending to the fundaments of life management that grounds individuals in vital signals from living bodies. Such affective blunting-and potentially alexithymia-may not only sever individuals from the experience of being governed by value-driven intentions, but also from being able to empathize with other feeling/valuing persons, potentially furthering such pathological disconnections from self and/or/via others. Theoretically, such disconnection could constitute a means of handling stressful situations that may be useful for short-term coping, or may function as an occasionally adaptive 'defence mechanism.' But when chronic, such an inability to connect with individual and shared experiences can have devastating consequences for oneself and others. Importantly, DP is most straightforwardly associated with altered self-modelling / phenomenology: "I look in the mirror and it doesn't feel like myself I'm looking at. It's like I'm floating, not actually experiencing the world, and slowly fading away into nothing. It's like I'm on autopilot in somebody's else body". However, DP seem to be almost paradoxically (given excessive self-focus) accompanied by a lack of 'plot' or 'narrative' in one's life: "If I quieten my mind, I can still almost taste the colour and richness of life as I knew it before depersonalisation happened; the feeling of being your own agent of change, the feeling of plotting a course through life, and the sense of expectation.". Such accounts are consistent with narrative selfhood requiring some degree of self-objectification, or leveraging 2 nd -and 3 rd -person self-perspectives, which have been speculated to involve the establishment of fictitious 1 st -person points of view via "mirroring" with others. In these ways, the problem of DP may not just stem from an inability to access certain bodily feelings, but may also involve a lack of perspectival flexibility. By this account, while attempting to 'grasp' onto the self may enhance presence and world-immersion by virtue of increasing the gain on somatic sensations, excessive and rigid fixation on selfhood may have the opposite effect, paradoxically resulting in a situation of experiential disembodiment and distancing from the vibrant immediacy of what is possible for the lived world of a conscious being. In the remainder of this paper, we suggest that acquisition of new self-and world models may enhance the plasticity of one's perceptual and sensorimotor experiences. This newly gained flexibility, we claim, may allow individuals to 'leave behind' certain habits-perceptual rigidities that hold them 'stuck' in certain behavioural patterns-and so open to new ways of perceiving and integrating self-and world-related information. This flexible modelling may be achieved during psychedelic (and potentially meditative) experiences via the controlled/flexible (dis)integration of ordinary/habitual self/world-beliefs, and consequent reintegration or re-organisation of such internal working models via modulatory multisensory information. Depersonalisation experiences, in contrast, point to a phenomenon of non-flexible (rigid) (dis)integration of ordinary/habitual self-models, and a consequent feeling of being 'stuck' in one's mind. While controlled (dis)integration of habitual self-experiences and consequent reintegration may have positive effects, uncontrolled (dis)integration of habitual selfexperiences triggered by unpredictable life events may be overwhelming and lead to selfdetachment and potentially adverse clinical outcomes. Indeed, if framed unskilfully, either involuntarily experiencing or seeking after transcendent states via "ego death" could involve a wide range of undesirable consequences ranging from reduced conscientiousness to the loss of personality stability more generally. The key claim here is that one does not simply 'lose' the self, but can only 'merely' change it. When these changes are radical, then what might be lost is the sense of (potentially excessive/entrapping/devitalizing) familiarity in life, for both better and worse. People have attempted to induce such changes via meditative practices for 1000s of years, with psychedelics having the potential to accelerate such transformations. While we are potentially witnessing an unprecedented renaissance in psychedelic research, we are only beginning to ask the questions of which interventions (and their "sets and settings") are most appropriate for helping different individuals to transform themselves, hopefully in ways that are consistent with achieving our most deeply valued goals. Towards this end, before concluding, we will briefly discuss models in which psychedelics may facilitate potentially therapeutic (transient) depersonalized states, thereby enhancing flexibility on multiple levels. 3 Losing (and finding) the 'familiar' self -the case of psychedelics Within the Active Inference Framework (AIF), the most well-known (non-mutually exclusive) models of psychedelic action are the "thalamic gating" and "RElaxed Beliefs Under pSychedelics" (REBUS) models. With thalamic gating, psychedelics have the effects of reducing top-down filtering of sense data and altering normal modes of synchronous binding, so allowing new patterns to make their way through the "doors of perception" with creative combinations. With REBUS, psychedelics are thought to have the effect of changing predictive coding mechanisms in ways that "relax" the precision of priors from higher (or deeper) levels of cortical hierarchies, so affording both a) greater updating from prediction-errors, and b) novel combinations of inferences. More recently, a broader model of altered beliefs under psychedelics (ALBUS) has been introduced in which 5-HT2a receptor stimulation may result in either relaxing or strengthening of beliefs (and potentially simultaneously at different hierarchical levels) depending on the dosing, sets, and settings associated with use. By all these accounts, psychedelics could provide means of (potentially radically) altering internal working models of self and world via a combination of atypical processing and elevated plasticity. Functionally, this would correspond to restructuring deep beliefs and their impacts on default modes of thinking and action. Algorithmically, this could be understood in terms of altered hierarchical message passing leading to atypical patterns of inference and updating. Mechanistically, this could correspond to altered and potentially more flexible-or "entropic")-dynamics for biophysical processes. Intriguing neuroimaging work has found support for this model of altered and potentially more complex/entropic patterns of neural activity for both meditative and psychedelic states, and even during virtual reality experiences that attempt to recapitulate aspects of psychedelic phenomenology. Evidence for enhanced network (and psychological) flexibility has also been obtained in studies of dextromethorphan (a dissociative with psychedelic properties) and more recently psilocybin. It may seem naïve to postulate that a single construct of "flexibility" could be at play in similar ways across multiple levels of analysis. Depending on how flexibility is defined, we may even see either too much or too little flexibility on mechanistic levels corresponding to either increased or decreased modes of flexibility on psychological levels in potentially nonlinear and possibly pathological ways. Indeed, a recent study found decreased psychological flexibility associated with lysergic acid diethylamide in the form of increased tendencies towards cognitive perseveration-which could speculatively be a consequence of increased experiential absorption. However, if flexibility is understood as a capacity for controllable reconfiguration, then it is conceivable that some measures could reliably indicate normative functioning across scales. It has been suggested that a dynamic network measure of "cohesive flexibility" might have such broadly beneficial properties(Figure), by virtue of indexing the capacity of systems to alter their connectomic community structures across time in a coherently organized-and potentially controllable-fashion. Perhaps intuitively, we may consider cohesive flexibility to be a proxy for harmonious flow capacity, so to speak, as complex adaptive systems minimize their informational (and potentially thermodynamic) free energy. Indeed, cohesive flexibility may index the meaningful degrees of freedom for a system, and in some ways could even be thought of as the primary capacity required for intelligent adaptivity, agency, or "free will". Theoretically, this kind of measure may correspond to a well-tuned active inferential system capable of exhibiting "self-organized criticality", or "edge of chaos" dynamics that allow for maximally efficient exploration of state spaces in ways that balance stability and plasticity. Such nearcritical organization may be the only (inter-)regime under which evolution (and/or/as learning) is possible in terms of both generating variation while preserving sufficient stability for the accumulation of adaptive complexity. The significance of this might be difficult to overstate, as (generalized) natural selection is the only process we have discovered that is capable of (temporarily) locally resisting the second law of thermodynamics-so providing foundations for the "first law of psychology". Figure(Taken with permission from: Cohesive flexibility may provide a functional bridge between moment-to-moment changes in psychological states, more enduring psychological traits, and the attracting states by which individuals evolve through time as cybernetic (free energy minimizing) systems. At each level of organization, flexible (potentially self-organized critical) dynamic processes allow for intelligent responding, learning, and evolution toward increasing degrees of adaptive complexity. Can psychedelics actually help to "tune" brains and minds towards increased criticality, and thereby flexibility/adaptivity? Intriguing evidence involving "connectome harmonics" suggests this may be the case, with psychedelics being observed to increase power law character across different eigenmodes (i.e., harmonic functions) of functional connectivity. Power law distributions-believed to reflect both "edge of chaos" nonlinearities and capacities for integrated/sensitive responding across scales-are one of the more disputed hallmarks of criticality. However, such findings have also received convergent support from other more wellestablished measures such as multiscale complexity as assessed by fractal dimension. One could even interpret views of depersonalization with respect to optimal grips on self and worldas reflecting a loss of cohesive flexibility via disrupted criticality, whether unbalanced in the direction of either excessive looseness or excessive rigidity. That is, a well-tuned complex adaptive system is only likely to obtain harmonious internal and external coupling if it can balance exploration and exploitation with respect realizing desired extrinsic outcomes (and hence affording allostatic coupling with the world) and foraging for the intrinsic value of obtaining novel information (and hence affording opportunities for learning/evolution). More specifically, self-processes that experience "psychological fusion" with respect to themselves could be considered to have an overly tight grip, and may be vulnerable to depersonalization as self-models become opaque via excessive self-grasping. For any thing/process of any appreciable degree of complexity, an optimal grip is necessarily a flexible grip. And with their capacities for selfreflection and for recursively-amplified combinatorics among potential meanings, selfprocesses may constitute the pinnacle of complex domains that require flexibility for sustainable harmonious functioning. Theoretically, dissociative such as ketamine and high-dose classic psychedelics may be beneficial in such contexts by helping such rigidly grasping selves to relax their maladaptive attempts at excessive control, which if unchecked may (apparently) paradoxically result in the loss of influence with respect to the very thing they desire to control. In trying to hold onto selves overly tightly, the experience of personhood may end up being lost as we lose connections with meaningful internal and external dynamics, which requires attending to their moment-to-moment unfolding/evolution. Such severed connections could take the form of disrupted neurovisceral coupling-potentially involving both internal and external entrainment of vital signals-alexithymia, and even Cotard's/Capgras syndrome-like experiences. Could psychedelics help recalibrate or "reset" pathological self-processes by helping them to relax (Carhart-Harris and Friston, 2019)-or potentially strengthen-their grips on the world? At present there is insufficient evidence to establish which neural systems are most relevant with respect to the potential "quantum change" associated with psychedelic states. Although modifying the previously discussed core 'circuits' underlying egocentric perspective-e.g. posterior cingulate cortices-may be particularly impactful as a source of therapeutic self-modifications via their temporary alterations. However, while excessive rigidity can lead to depersonalization through self-fusion, insufficient stability may lead to somewhat similar depersonalized places through a lack of engagement with both individual and shared meanings. The work (and play) of sustainably being a person can be complicated. Although not being a person can be even more complicated. That is, we need well-tuned personhood if we are to flexibly orchestrate internal and external dynamics-and/or "harmonics"-in the process of navigating between order and chaos, as we attempt to become who we want to be and realize our deepest values for ourselves and others. Along these lines, and in the spirit of precision psychiatry, it may be desirable to customize psychedelic, meditative, and even other biofeedback and brain stimulation protocols to target particular neural systems in specific ways based on individual characteristics (and sets and settings). Speculatively, modifications of dorsal anterior cingulate and dorsomedial prefrontal cortices (and perhaps the temporoparietal junction) could directly influence depersonalization by altering (or re-tuning) the functioning of attention/intention schemas for self and other. While disrupted posterior cingulate functioning might be more associated with dissociation (and potentially depersonalization thereby), modifying these other systems may more straightforwardly impact intersubjective/social-cognition schemas and sources of intention/will/agency, all of which are essential aspects of personhood (and co-personhood). Perhaps similarly, the extensive association of dorsolateral prefrontal cortices with executive functions could be enactively recast in terms of "optimal grips" via an action hierarchy that developmentally emerges (or grows) from experiences with physical gripping. Can exogenous interventions teach individuals to more adaptively/flexibly adjust these (generalized) grips in different circumstances? Alternatively, the neural hypnosis literature and lesion studies suggest that the precuneus and temporoparietal junction may be central for action (and potentially self) ownership via visualizing outcomes in the mind's eye and simultaneously feeling them in the mind's body, which may provide sources of prediction/control-energy when smoothly functioning, and potentially depersonalization when disrupted. These areas would all constitute "hubs" within the brain's connectomic networks, and would be essential for generative self-world modeling with temporal depth and counterfactual richness by which various forms of conscious presence and selfhood are realized. Theoretically, models of "pivotal mental states", goalhierarchy "collapse" via "allostatic overload", and "hub resetting"suggest it could even be the case that natural selection may have 'designed' such centralized control structures-potentially mediating various forms of selfhood-such that they could be temporarily disrupted, so opening up novel areas of state/inference space and affording opportunities for potentially radical regime shifts in minds. In light of these considerations, perhaps it is less surprising that DP-like altered states (as well as their opposites) are often sought out as part of psychedelic experiences, while also being considered as a source of potentially adverse outcomes associated with psychedelic use. The ability of psychedelics to so radically alter fundamental processes of selfhood speaks to their transformative potential, allowing individual to see/experience the world differently, potentially teaching them how to "change their minds", and lives, and perhaps the broader world.

CONCLUSION

In these explorations we have considered the experience of 'losing' one's familiar sense of self in Depersonalisation (DP) and psychedelic experiences through the (perhaps often overly opaque) lens of Active Inference. Depersonalisation is characterized by feelings of being detached from one's self, body, and world. We suggested that acquisition of new self-and world models may enhance the plasticity of one's perceptual and sensorimotor experiences. This new gained flexibility, we posited, may allow the individual to 'leave behind' certain habits, perceptual rigidities that holds him/her 'stuck' in certain behavioural patterns, and open to new ways of perceiving and integrating self-and world-related information. This adaptive modelling may be achieved during psychedelic (and potentially meditative) experiences via a flexibly controllable (dis)integration of ordinary/habitual self-models, and a consequent re-integration or re-organisation of the latter via modulatory multisensory information. By contrast, depersonalisation experiences point to a phenomenon of non-flexible (rigid) (dis)integration of ordinary/habitual self-models, and a consequent feeling of being 'stuck' in one's mind. While controlled (dis)integration of habitual self-experiences and consequent reintegration may have positive effects, uncontrolled (dis)integration of habitual selfexperiences triggered by unpredictable life events may be overwhelming and lead to selfdetachment and potentially adverse clinical outcomes such as depersonalisation disorder. If these models are accurate, then one may not simply 'lose' the self, but may nonetheless change it radically in a fashion that is either (a) controlled, intentional, self-strengthening, or (b) uncontrolled, unintentional, and self-undermining. While either mode of re-tuning the self may be more or less desirable for different individuals in different contexts, the latter may constitute a more dangerous 'intervention,' potentially being associated with a legacy of trauma (Van Der Kolk, 2015)-i.e., unnavigable free energy landscapes. Keeping a flexible balance between more exploitative or 'familiar' and more exploratory or 'unfamiliar' ways of perceiving self and world may be the key for a fulfilling life in biological and experiential domains. However, one must also be able to keep sufficient agency (i.e., or controllable degrees of freedom) over this fine-tuning between the stable maintenance of order and the plastic changes that open one to the (hopefully) fabulous new. Without such agency, one may find themselves 'lost' in a world where they are unreal and unable to connect to meanings, feeling like strangers to themselves and others. But without skillful discernment-and perhaps compassion-one may find oneself similarly forlorn if attempts at being in control undermine the very preconditions by which meaningful influence is possible: healthy personhood and wellbeing requires a delicate balance of flexibility in being able to hold on more tightly when desired, but also knowing how to let go when needed.