Interoception
The example statements and sayings used in this article are an attempt to ‘bring alive’ the concepts they are illustrating. However, obviously, within themselves, the statements and sayings can hold a broader meaning that may challenge the concepts they are attempting to illustrate here, but that is not their intention.
What is Interoception
Interoception is the conscious and non-conscious sensing, interpreting, integrating and regulation of internal bodily signals to provide the individual with a momentary map of their body’s internal landscape and their relationship to the outside world. This is fundamentally important to facilitate the sort of adaptive change that Darwin believed was critical for survival (Paulus et al 2019) regulating homeostasis through the visceral, immune, and autonomic systems, using nociceptive, chemosensory and thermoregulatory information (Bohlen et al 2021). For example, it regulates, temperature, pain, immune, hormonal, and cardiovascular activities, touch, hunger and thirst (Di Lernia et al 2016).
In a dynamically changing world these homeostatic and allostatic mechanisms are relatively stable as violating set boundaries e.g. body temperature can have serious health consequences. Also, there are usually no accurate objective representations of subtle physiological fluctuations e.g. if I wanted an accurate representation of my hand I can look at it, but a representation of subtler fluctuations in my physiology that has to be constantly ‘tweaked’ before any overt symptoms occur has to be ‘felt’ by guessing our current physiological state and predicting our future state as a consequence of any internal or external action (Seth & Friston 2016); this makes us “pro-active survival-enabled prediction machines” (Ciaunica et al 2021). However, the individual can stand back and more objectively monitor the overall performance of these homeostatic and allostatic mechanisms. For instance, taking into account current, and what I predict or anticipate to be the future states of both my body’s internal environment and the external environment (Unalac et al 2021) is important when regulating metabolically important molecules.
Our brain has a constant 24/7 bombardment of every single interoceptive stimuli that obviously can’t be consciously appraised and acted on. Therefore, in order for our interoceptive awareness to infer the causes of its sensory input, explain past events, control present happenings, and prepare for the future, it has to avoid surprise (Brown & Brune 2012) in order to establish continuity and pave a predictable path in order to be adaptive in maintaining optimal physiological and psychological integrity and identity (Seth & Tskaris 2018). It does this by (i) to avoid uncertainty, use previous experience to predict, and based on that prediction, instigate what it believes to be the most appropriate physiological response; (ii) prioritise unexpected, surprising, stimuli (hence most likely to be informative) when deciding how to allocate resources (Fradkina et al 2020). This is how pain, as a helpful interoceptive stimuli, can teach us to adaptively respond to real physical threat and avoid potential harm in upcoming situations (Lim et al 2020).
Interoceptive information is transmitted to the brain for processing via the vagus and glossopharyngeal nerves and via viscerosensory, somatosensory, chemosensory, and lamina I spinothalamic pathways (C-tactile afferents activated through low force 3cm/sec, stroking) (Bohlen et al 2021).
The origin of interoceptive input
Interoceptive input is afferent input generated in response to an activity. What motor activity we decide to initiate depends on what we predict, or anticipate, the outcome or sensory consequences of it to be e.g. we predict, when we want to speak, that the sensory consequences of our muscles moving are the sound of our own voice. These predictions also generate rapid corrective movements when we experience sensory stimulation we were not expecting (prediction errors) e.g. when sitting in a chair and surprised it’s deeper than we predicted, this error may cause our quads to contract to slow the rate of descent. These predictions, guiding motor movement, occurs through two different mechanisms (Liesner & Kunde 2021):
Predictive based models. In predictive based models it is the intention of achieving a certain goal (e.g. grasping an object) that triggers the motor plan to achieve this intention. Based on this motor plan, one predicts how it should “feel” once this motor plan is achieved and the object has been grasped. The actual sensation whilst grasping the object is then compared with how one thought it would’ve felt in the predicted state and any deviations from this predicted state are bought to our attention; for example, you intend to reach for a cup without looking and this initiates a motor response. You’ve predict what a cup should feel like when your hand touches the object it doesn’t feel like a cup. This mismatch between what you predicted to feel and what you are actually feeling (‘prediction error’) gets bought to our attention so we can either initiate corrective movements, or, depending on the evidence, it should update how we represent what a cup should feel like.
Ideomotor models. In contrast, ideomotor reflexes and learning start with ‘motor babbling’. Motor babbling is characterised by purely random movements and behaviours that aim to practice or explore promoting learning centred around a self-rewarding curiosity (Haar et al 2020) (e.g. when babies explore their voice by saying “bababa”). Therefore, motor babbling is not orientated towards a predetermined goal or anticipated outcome, a baby doesn’t say “bababa” because it’s goal oriented towards refining its voice for public speaking, but it’s more of an exploration into the unknown ‘prod it and see what happens’. Over time associations are learnt between these motor actions and what is felt as a consequence of performing these actions which in effect tees up and primes the motor system e.g. learnt associations develop between the motor movement of saying ‘bababa’ and felt experiences such as noise, vibrations and received attention from other people.
Once these links have been established and any conflicting ideas have been removed (Massen & Prinz 2009) this process can be reversed; therefore, instead of motor actions instigating a felt consequence (‘I do something —> I feel something’), simply using our imagination, our mental representation, to recollect the consistently produced consequences of these motor actions activates motion in order to achieve these desired consequences or ‘goals’ (I feel something —> my autopilot elicits learnt motor responses’). An example of a central idea releasing, triggering and giving life to a teed up muscular system is whereby simply anticipating or predicting how it “feels” to grasp an object would activate the motor response; but because the strength of the induced action will depend on how strong the link is between what is being perceived and what are the perceivable effects of that action are (Massen & Prinz 2009) grasping a falling baby will drive a stronger motor movement than grasping a falling piece of paper.
Three dimensions of Interoception
There are three distinct dimensions of interoception (Garfinkel et al 2015):
Interoceptive accuracy: is the ability to accurately detect and track internal bodily sensations e.g.“Can you accurately report when your heart is beating?” . This can be objectively measured by e.g. taking your pulse. Individuals with high levels of interoceptive accuracy experience higher levels of arousal, but are also sensitive to cues informing them how they can positively self-regulate their behaviour and follow their intuition (Tsakiris 2017).
Interoceptive accuracy anchors an individual in their body awareness being able to block out, and not get swept away, with exteroceptive signals from the rapidly moving outside world (Tsakiris 2017). This is illustrated in the saying “and breath” whereby predicting, controlling and experiencing ones internal breath anchors oneself against the distractions from the outside world.
In chronic pain the nervous system is saturated with the constant physical and emotional aspects of pain. This constant deafening ‘noise’ reduces interoceptive accuracy by drawing our attention towards pain related information and away from other interoceptive information. This makes the nervous system reliant on multisensory integration of external inputs (e.g. vision, touch, etc) to create a picture of behavioural and autonomic interoceptive functions; it holds higher value to exteroceptive input when it appeals to our mental representations (De Lernia et al 2016). For example, low levels of interoceptive accuracy during flexion is seen when “I know my own body” and flexion is disproportionately predicted to have catastrophic consequences. This heightens the value of sensing pain as a warning sign to prevent further damage; misinterprets or simply ignores other interoceptive inputs (e.g. stretching sensation); normal proprioceptive input is interpreted as something ‘moving out of place’ and there is an over-reliance on external input that appeals to our mental representations of an ensuing catastrophic injury e.g. touch “I can feel my muscle spasm when I bend forward” (when they’re just feeling a normal muscle being stretched).
In contrast, acute pain, that has a practical value in protecting against further damage, has a high interoceptive accuracy, so we can accurately feel our internal bodily sensations as to avoid further tissue trauma (De Lernia et al 2016).
Interoceptive awareness: how we access interoceptive information, by how we explicitly think about, and appraise it; this insight determines how we approach or withdrawal from this stimuli “Do you ‘know’ whether you are accurately or inaccurately assessing your heart-timing?”; “do you ‘know’ whether your beliefs around bending forward are catastrophic or physically warranted?”
Increased awareness of interoceptive information makes it easier to reappraise, learn from, promote empathy and manage in order to challenge unhelpful mental representations e.g. catastrophic belief systems. Any type of self focused attention improves an individual’s interoceptive accuracy, for example (i) low-level perceptual and bodily aspects of the self e.g. gazing at ones face in a mirror; (ii) high-level, cognitive and narrative aspects of the self e.g. looking at autobiographical words evoking personal memories and traits (Ciaunica et al 2019).
Conversely, reduced awareness of interoceptive information makes it easier to suppress, harder to learn from and manage and opposes developing empathy. A poor recognition of interoceptive signals promotes genuine self-denial and can reduce processing of positive stimuli resulting in a consistent flat line of emotions contributing to e.g. depression (Brewer et al 2021).
Interoceptive sensibility: our self-belief in our awareness and ability to be internally self-focused, engaged and responsive to interoceptive processes. “To what extent do you believe you focus on and detect internal bodily sensations?” A low level of confidence in interoceptive sensibility can lead to feeling a low level of control, for example “I get confused as to whether or not I am hungry” and “I don’t know what’s going on inside me” (Khoury et al 2018).
Interoceptive accuracy is central to the construct of interoception underpinning interoceptive awareness and sensitivity (Baltazar et al 2021). How accurate we truly are in sensing interoceptive information (accuracy) is related to how we represent, think about and appraise this information (awareness) but not our perceived self-beliefs about our ability to so (sensibility) (Garfinkel et al 2015).
Khoury et al (2018) broadened the dimensions of interoception to include:
Interoceptive mode of attention: describes the contrast between a direct, experiential non-judgemental awareness of body sensations and a reflective, labelling or ruminating on interoceptive signals. A direct, experiential non-judgemental awareness of body sensations represents a core aspect of many mindfulness and related mind-body interventions.
Interoceptive attention quality: tendency to pay attention versus ignore body sensations.
Interoceptive attitude: between appraising body sensations as helpful (trusting attitude) or menacing (catastrophising attitude). This dictates feelings about the consequences (including somatic consequences) and cognitive (dys)control in relation to interoceptive signals.
These three qualities determines self-efficacy as it relates to an individuals’ confidence in their ability to focus on a sensation, to sustain or control the mode of attention, and to attain an anticipated outcome from the experience i.e. minimise uncertainty as to ascertain feelings of confidence and control; refer to ‘interoceptive sensibility”. For instance, a catastrophising interoceptive attitude can lead to increased fear and a more ruminative interoceptive mode of attention; this can foster a high level of interoceptive attention tendency that leaves an individual with low levels of interoceptive attention self-efficacy and the belief that he or she is no longer in control (Khoury et al 2018).
This gap between perception and reality is called the ‘trait interoceptive prediction error’ (TIPE). It is a variant of interoceptive awareness, and is the gap between what’s really going on i.e. the raw sensory input (accuracy) and our perceived self-belief in our abilities to sense what’s going on (sensibility) (Murphy et al 2019). In order for us to be aware of when we got it wrong, when our perceptions predicted that ‘x’ was going to happen, but this guess was wrong, and instead our interoceptive input told us that ‘y’ actually happened we need to have good interoceptive accuracy (Baltazar et al 2021). Added importance is given to sensory information when it is deemed incomplete, missing or unreliable; therefore how much weight we give to learning from these prediction errors, how surprised we are to be wrong, depends on how we cling on to our prior beliefs as being ‘true’ i.e. if we cling on to something with absolute clarity then any contradicting sensory information (prediction error) will have far more potent effects (Fradkina et al 2020).
Mental Representation and Interoceptive Information
Our mental representation are formed from our prior knowledge and interpretation of past experiences. It gives structure to interoceptive information by mapping out how we should respond to this constant 24 hours bombardment of relatively nondescript interoceptive bodily signals. This respsonse is dictated by how we interpret and predict what the interoceptive information means, and, anticipates what will happen as a consequence of our actions and in turn, how we should respond to them consequences. Therefore, by being highly personal and subjective in its interpretation and predictions it reflects our unique self-awareness, our subjective experience of being present in the here and now (Ciaunica et al 2019) (refer ‘interoception and sense of self’).
The prior experiences that form our mental representations not only tries to predict what the cause of the actual, or anticipated, interoceptive input is and what will happen as a result of future action, but also shapes our subjective experience, our self-awareness, more so than the current sensory input (Ciaunica et al 2019); for example if we predict, based on prior experience, an anticipated or actual stimulus will be painful it will be experienced as being more painful than if we had been less anxious and just experienced the stimulus for what it was.
Therefore, our perception is a first hand experience achieved by how much weight we shift to either our prior expectations and predictions (top-down priors), or, to any anticipated, or actual, raw untainted incoming sensory information (bottom-up information). This shifting of weight defines to what degree we tip the scales in favour of either (i) our top-down prior expectations resulting in us manipulating bottom-up sensory information through a familiar predictable lens so we can avoid change and perceive our virulently maintained biased preferences, or, (ii) in favour of raw, untainted bottom-up sensory information so it can confirm top-down prior predictions, or, update them allowing us to experience novelty as new expectations and models of ourselves and our environment are formed. However, whether our perception are shaped by the degree to which we tilt the scales to favour maintaining the comforting predictability of our status quo, or to what degree we favour exposing ourselves to novelty and unpredictability, our prior top-down expectations give a point of reference so that we can measure in what way the sensory information has successfully, or unsuccessfully, met out prior expectations and predictions. This gives a personal structure to the anticipated, or actual bottom-up sensory information to make it relevant and personal to ourselves, e.g. it defines the unique characteristic ways in which we define an individual delivering a joke funny or offensive, and is why top-down priors about the characteristic expectations of a stimuli are deployed prior to our perceptions (Reeder et al 2024), e.g. our top-down narrative is already in place so that when we hear a joke we can readily perceive it as being funny or offensive, unless, our top-down narratives are incompatible with the nature of the joke whereby the sensory information is perceived literally and we are at a cross roads trying to work out if the joke is funny or offensive.
Akin to a map, mental representations are crucial to making the navigational decisions within the terrain we are coursing. It does this by: (i) forming a structural representation of our inner and outer world to guide actions. In order to achieve a clear, specified goal of navigation, a tube map doesn’t overwhelm an individual with every detail of the underground (e.g. depth of the station, proportionate distance between the stations), but presents just enough information, so long as we locate ourselves accurately on it, to plan a future journey whilst anticipating future events and obstacles; (ii) representations should be challenged by seeing if what they predict comes true does. Looking up from the map verifiers navigational decisions so corrections can be made. We can be led astray by an outdated map, or we can ‘bury our head in the sand’ and not look up from our map to see where we truly are; (iii) direct interactions with a mental representation can be detached by observing oneself from the outside as to reflect on, and guide cognitive actions. One could sit at home with the map planning a journey, making future predictions by learning the layout and formulating metaphorical autobiographical concepts using simulations and imagery (Gladziejewski 2015).
Without this map, or our representation, we would have no construct to our reasoning, we would try to interpret and make predictions about where we thought we were, and where we actually are through blind luck, chaotically scrambling along different paths to reach our destination (Gladziejewski 2015). Also our representations, shaped from our personal interactions with the world and the embodied experiences that accompany it, reflect our own unique experiences and knowledge; being able to predict, based on our representations, what interoceptive, and exteroceptive sensations means, and what the unique sensory consequences will be as a result of our response to that information, is a unique experience that is known only to ourselves and forms a ‘sense of self’ (refer ‘interocpetion and sense of self’). A perceived defect in the ability of our representations to be able to navigate the environment, predicting how we operate and how our actions will impact the environment, can lead to a feeling of insecurity, uncertainty and unpredictability that can lead to an excessive reliance on gathering sensory information to regain certainty and control (Farfan et al 2020 & Teng et al 2016) (refer ‘generative (predictive) models of predictive coding; generative models and obsessive compulsive disorder’)
Once this internally construed mental representation from the cortical hierarchy interprets interoceptive information by inferring (or predicting), the most likely cause (posterior) of these incoming sensory signals (Gladziejewski 2015) it drives down these top-down predictions to resolve, or suppress, any mismatch in conflicting bottom-up interoceptive information that contradicts this prediction (prediction errors) (Seth & Tsakiris 2018) and instigate the appropriate motor and autonomic responses based on their predicted outcome. For example, (subconsciously): bottom-up interoceptive data: low blood sugar; mental representation of interoceptive data: “I feel hungry, it’s probably because my blood sugar is low”; autonomic response based on mental representation: “I predict if I elicit x physiological response I will no longer feel hungry”.
But, any discrepancy between what our mental representations, from the cortical hierarchy, drives down about what it predicts to happen (“x physiological response should elevate blood sugar as not to feel hungry anymore”), and the ascending data that signals what really is happening (“but I’m still hungry”) leads to a prediction error (Joshi et al 2021). When these prediction errors aren’t suppressed or can’t be resolved they are deemed newsworthy and gets flagged up as a ‘wake-up call’ (consciously “I’m hungry I must eat something”); this triggers a motor response (i.e. eating something) that predicts a sensory ‘endgame’ (i.e. not feeling hungry). Due to the sheer quantity of interoceptive signals, and prediction errors, over a 24 hour day our brain has to select what ones it attends to but in order for this process to occur in a healthy cycle (i.e. eat proportionate to appetite) there has to be good level of interoceptive awareness (Baltazar et al 2021).
Good interoceptive awareness requires healthy mental representations that are open minded when interpreting interoceptive information and inquisitive to any prediction errors. Therefore, by being mindful, non-judgmental and open-minded we can accept prediction errors seeing them as wake up calls that challenge and update our mental representations.
However, if worry and rumination have formed rigid, inflexible mental representations that are intolerant to being challenged by prediction errors then these prediction errors must be suppressed or actively changed in order for what we ‘feel’ to be constrained within safe, predicted, boundaries (Bohlen et al 2021), this lowers our interoceptive accuracy and constructs a self-fulfilling prophesy. Therefore, lowering our interoceptive accuracy by suppressing or actively changing our interoceptive signals to conveniently cancel out prediction errors creates a delusional reality that comforts us by making us believe that we know where our interoceptive signals are coming from, that we can predict them as a consequence of an action and that we can process them in their entirety, this strategy attempts regain a sense of control and predictability and foster a sense of self (Liesner & Kunde 2021). When an individual becomes consciously preoccupied with this, they imprison themselves in a self-sustaining loop constantly anticipating what this worry and ruminative style of thinking predicts (Di Lernia et al 2016).
In antipathy to the notion ‘all models are wrong but some are useful’ these inflexible mental representations, reinforced by worry and rumination, centres around an individual’s overly strong expectation that their model is correct and that their expectations and predictions will manifest themselves (e.g. the expectation of pain and disability). There is also a failure to adjust these unyielding high expectations to different contexts that represent the fluidity of the real world (Paulus et al 2019). This is why, despite its negative impact and punishing emotions, this process of obtaining certainty, (as opposed to avoiding uncertainty), about future events can actually be valued as an effective problem solving tool (Teng et al 2016) so that individuals, with, for example, Generalised Anxiety Disorder, whose mental representations shaped from anticipation and worry places so higher value on hypervigilant threat appraisal and negative thoughts about an up-coming event (White et al 2016) that they can believe that worry can be helpful in resolving their problems (Hirsch et al 2019). This re-enforces dysfunctional beliefs and leads to a loss of self confidence whilst reducing sensitivity to updating mental representations by learning from any contradicting, positive, bottom-up information that may presents itself. By engaging in this self-fulfilling prophecy, with blinkered autonomic hyperarousal, these pessimistic highly anxious mental representations construct a negative-bias towards future thoughts that blocks out pleasantly surprising, less threatening, contradicting information (White et al 2016).
Therefore, valuing negative over positive stimuli that impairs effective decision making and perpetuates the worry cycle (Teng et al 2016) is reflected by an overly valued, overly expectant mental representation that is inflexible to being updated by contradicting interoceptive information so it fails to correct prediction errors. Unable to adapt to interoceptive information these dysfunctional mental representations have to either (i) close the prediction error by ‘flipping’ interoceptive information to a more palatable reality or (ii) the individual has to repeatedly experience prediction errors that are manifested through recurrent failings in the external world (i.e. falling short of an objective measure) and in their internal world (i.e. failing to establish homeostatic balance in the nervous system).
This results in unpleasant and uncomfortable sensations triggering anxiety, worry, threat appraisal and inappropriate approach or avoidance behaviours (Paulus & Stein 2006), self-related cognitions, dysfunctional learning about bodily states and increased allostatic load leading to stress and mental illness. In severe cases when the prediction error is so pervasive the error can only be quelled by complete avoidance of all perceived triggers (e.g. agoraphobia) (Paulus et al 2019). This can create an urge to perform compulsions in an attempt to reduce anxiety and cement inaccurate associations. Forming experiences and knowledge through this lens constructs mental representations that contorts interoceptive (and exteroceptive) input and lowers interoceptive accuracy so that reality is blurred and individuals become unable to differentiate between, for example, surprise and fear; cardiac or gastric activity and disgust; increased heart rate and anger; temperature or fear and increased heart rate and blood pressure (Brewer et al 2021); harmless sensations and pain, attribute benign bodily sensations to pathology (Bohlen et al 2021) and have difficulties recognising cues related to feeling hungry or full contributing to eating disorders (Brewer et al 2021).
In pain catastrophising, how we process ascending interoceptive input from pain, what we choose to accept or suppress, how we feel about it, how we anticipate our needs in response to pain, how we prepare to satisfy them needs before they arise by activating the autonomic nervous system represents our perceived unique experience, our ownership, our interoceptive representation of pain. Whether our mental representations of ourselves, our condition, our pain and how it effects our life parallels what it should based on the biological nature of the condition is of course another matter. The same can be true about our relationship with interoceptive sensations of hunger or satiety and food, our relationship with increase heart rate and stress, etc.
Interoception and sense of self
Some perceptual effects of motor activities are “public”; motor control of the eye enables you to perceive more or less, the same sensory event as anyone else looking at the same thing. However, interoceptive stimuli is more private, it is unique to only me, no one can else can experience and rationalise my interoceptive stimuli; it involves how I alone can control and feel sensory, interoceptive, stimuli that arise from my body and is apparent only to me; this engenders a sense of ownership that ‘my body belongs to me’ ensuring stability and unity, anchoring oneself in an ever changing, public, external environment (Tsakiris 2017). This sense of how I inhabit my body and my interpretations of the world in belonging to a practical, purposive place with possibilities that I am entwined and embodied with (“that is so me”) is different from seeing things in a purely object-type way. This embodiment of ourselves and ourselves in the world is why it is the interaction of our interoceptive and exteroceptive senses that has a substantial impact on shaping our sense of self (Ciaunica et al 2019).
Our self consciousness coordinates and integrates divergent information from:
Sense of agency: a robust, slowly learnt (Ishikawa et al 2021) experience of control, that based on my context, my knowledge and beliefs that form my judgements and intentions that I am the author, and can predict the consequences of, my own actions. It is therefore dependent on top-down predictions being confirmed by bottom-up information derived from either motor actions or social or environmental cues (Braun et al 2018).
Sense of ownership: a quickly learnt, adaptable (Ishikawa et al 2021) feeling, largely dependent on integrating bottom-up information such as visual or somatosensory inputs (Braun et al 2018), as proprioceptive signals stem from ones own body as an innate self-reference of where ‘my’ body is in space, an integration of proprioceptive signals can also promote a sense of body ownership (Liesner & Kunde 2021). This allows us to relate to, and understand, the first-hand narrative that integrates diverse experiences into a single awareness (Dorhary et a 2021) creating a sense of belonging, being grounded (Wang et al 2021) in the familiarity of being ‘me’. This awareness represents the brain’s attempt to make sense of reality, integrating tightly bound bottom-up information, to create a caricature that represents an embodied ownership over an external object, our body parts, personality, qualities, values and the characteristic ways we engage with ourselves and others. This constructed caricature also gives ownership over an illusional continuous identity by forming a conveniently cracked lens through which the ‘present me’ can see a persistent identity with the ‘past me’ and ‘future me’ (Dorhary et a 2021).
This integration by the self-consciousness makes the brain believe these events have occurred in the same time-line and have a causal relationship in order to minimise uncertainty (Ishikawa et al 2021) so we can intertwine ourselves with the experience of how we completed an intended action within our environment (agency) and embodied this experience (ownership) to develop the ‘minimal self’ (Forch & Hamker 2021, Riva 2008).
The minimal self is the most fundamental embodied experience an individual has of belonging to their own body, their sense of self, the intrinsic unity between their self-consciousness and the world to which they merge with and become rooted in; it forms the most basic experience of being a subject of a given experience. It is not a grandiose third person awareness of oneself as an object, or a conscious evaluation that labels ones experiences. It is a first person perspective that contains only the minimum to experience the core of ‘what-it’s-like-for-me’ feeling. This feeling defines how the individual, as an unseen point of origin, intends to broadly orientate themselves towards the world, framing how they interact with their environment so this internal feeling of individual identity can externally manifest itself; how this manifestation materialises enables the individual to become self-acquainted so they can foster predictions about the meaning of being ‘me’ (Klar & Northoff 2021), and integrate sensory information and mental representations to reduce prediction errors about the sensory consequences of volitional actions (Ishikawa et al 2021) and the likely causes of sensory signals (Ciaunica & Crucianelli 2019).
Interpreting and trying to infer what these afferent nerve signals mean to us as individuals is unavoidably burdened with prediction and uncertainty. For example, we can interoceptively sense, and objectively measure, an increase heart rate but how do we interpret what that means? Is it determination? Excitement? A sense of challenge? Fear or threat? Biologically a challenge response is characterised by increased cardiac output and a decrease total peripheral resistance, and in contrast, a threat response is characterised by no change or a small increase in cardiac output and no change or an increase in total peripheral resistance (Uphill et al 2019); however, our brain, receiving this interoceptive information as nerve signals, not as descriptive biological text, has to interpret, predict and infer what the cause of this interoceptive information is if it is to define what it means to us, and then elicit the appropriate motor or autonomic responses based on our own prior experiences. This is why we rely on our own unique interpretation of autobiographical metaphors to identify how we feel “I got butterflies in my stomach” or “I feel a bit off” as opposed to relating to how we feel in purely detached biological terms.
These bottom-up afferent interoceptive signals should ascend the nervous system informing each level, or hierarchy, as to what’s happening in the body. In this ‘raw’ untainted form interoceptive signals are accepted, without prejudice or bias, so they can be appraised in a less sensitive and more open-minded inquisitive nature. This flexible mindset allows any bottom-up sensory data, that conflicts with what we predicted it was going to be, to challenge and mould our perceptions. This bringing together of (i) our representations, our perceptions of what we have learnt from prior experience to be the ‘truth’ and constructs our predictions that anticipate what the bottom-up interoceptive data is going to be, and (ii) what the raw, untainted, interoceptive data actually is, leads to a greater understanding and ownership of ones own body. This ‘bringing together’ process, closing the discrepancy between what we predicted to feel and what is truly felt, by changing our mental representations to fir the interoceptive data, is called ‘reducing the prediction error’.
Closing the prediction error doesn’t just occur by changing our mental representations to match the interoceptive data, we can also, selectively change the interoceptive data to match our mental representations, whilst, placing an excessive value on sensory data that reinforces the values that define our mental representations; this allows us to perceive the world (and self) not as it is, but as it is useful to do so (Seth & Tskaris 2018). Allowing our prejudices and biases, especially those cemented through worry and rumination, to rigidly contextualises the prior experiences that form our mental representations skewers, to varying degrees, our perspectives and self-awareness of our interoceptive input to lower interoceptive awareness. Therefore, it is the strength of our prejudices and biases, in what direction we turn our attention and where the blindspots in our imagination are, that dictates our need to maintain a sense of control and certainty in light of challenging or contradicting interoceptive data (Lim et al 2020). Actively engaging in this self-fulling prophesy of changing interoceptive data (active inference) reduces the prediction error by forming a convenient justification for the values of the prior experiences that form our ‘truth’, our mental representations (perceptual inference). *: refer to ‘generative models of predictive coding’.
Prior precision is the term that refers to the clarity or ambiguity of the top-down prior expectations that form the mental representations by which we measure, or fail to measure bottom-up sensory information against. These prior expectations should be precise or imprecise, strong or weak, and flexible or inflexible depending on the situation we find ourselves in. For example, you should have precise prior expectations for a friend’s face so you can recognise them. However, if you know your friend has shaved their beard your prior expectations of what they will look like should become less precise and more flexible so that you can measure any visual information against these ‘looser’ expectations of their appearance. Alternatively, when detecting a stimulus that does not change identity, but, can appear at different contrasts having a precise, inflexible, and narrow prior expectations for that stimulus can be beneficial, for example, looking at a friends face from an angle that makes them unrecognisable (change in contrast), means you might miss them unless you stay steadfast to the characteristic features that you have modelled as being indicative of them (precise, inflexible narrow priors resistant to being updated by a change in contrast). Therefore, the mental representations of ourselves and the world become maladaptive when they become unresponsive to sensory information that highlights the anomalies, that should, in any given context, update our prior beliefs and expectations (Reeder et al 2024).
Commonly, however a combination of these processes occurs to reduce the prediction error, some contradicting interoceptive data tweaks our mental representations (increasing interoceptive awareness), whilst other contradicting interoceptive data gets suppressed or explained away (decreasing interoceptive awareness). Reducing the prediction error so we feel how our self-consciousness provides a sense of predictability and ownership, characterised by performing an action, correctly predicting the outcome of that action and embodying the experience so we feel intertwined with it (Riva 2018) allows us to feel fully immersed and present (Riva 2018) and in touch with ourselves and the world (Ciaunica et al 2021). This fosters a sense of control (Lim et al 2020) that maximises the evidence for self existence (Ciaunica et al 2019) as the prediction error diminishes by the interoceptive data confirming our mental representations, and, symbiotically, our mental representations confirming the interoceptive data so they become intertwined leaving no doubt that we can shape our own narrative. This is why, when a prediction error does occur, characterised by something unexpected happening, individuals feel amiss or disconnected from their bodies (Ciaunica et al 2021), e.g. strolling along without a care in the world, correctly predicting where your feet will land, and then unexpectedly stumbling.
Underpinning all this is how our perceptual experiences, that are dependent on our interactions with others and the environment (Ciaunica et al 2021), defines our logic in constructing tests that measures the consequences of our actions that should reinforce adaptive, and challenge maladaptive thought processes (Lim et al 2020) so that we interact with the world in a way to find out more about ourselves. This determines in what direction we grow and evolve,"stultus est sicut stultus facit”, being equally as applicable to exteroceptive and proprioceptive signals,
For instance, when biasing pain perception catastrophisers become easily overwhelmed by uncertainty, anxiously anticipating, and being hypersensitive to the threat of pain. Through worry and rumination they dwell on repetitive, fearful, negative, unhelpful thoughts making it logical to pre-empt the certainty of catastrophic consequences by fostering a hypervigilent inflexible mental representation of their pain, their condition and level of disability; these representations are not prepared to bend to any information that challenges them (prediction errors) in an attempt to regain control and forge a path to safety; blinkered to the rewards associated with finding more open minded, challenging, novel solutions and preferring safer, more predictable, familiar, behaviours that appeal to preferential biases and ideals (Brewer et al 2021) results in an ‘unmoderated, predetermined hypersensitive-to-warning-signs’ mental representation of pain that predicts an unyielding imaginary level of threat (Lim et al 2020) or pain during a particular movement; “just the thought of touching it sends pain through me”. A logical appraisal after the event can be just as inflexible by failing to fully appreciate the maladaptive nature of their response and in turn reinforce learned associations that underpin habitual actions (Lim et al 2020).
Therefore, our mental representation informs how we choose a motor activity based on what we predict we will feel, and process, as a consequence of that motor activity. This interpretation of events is unique to us and reaffirms an awareness of ownership when the predicted sensory consequences of a motor activity matches the actual sensory consequences (Tsakiris 2017). This predictability makes these actions feel more self-generated and intended (Blakemore et al 2000), as they are highly anticipated so the individual is ready for them, being less distracted and focused on the sensory consequences of the upcoming event (van Kemenade et al 2016). For instance, how we lower our heart rate, predict what we will ‘feel’ as a consequence of actively lowering our heart rate, and actually ‘feel’ what we predicted we would fosters a greater sense of unity, stability and ownership, than someone else objectively measuring our heart rate and telling us we’re relaxed because our heart rate just happens to be low.
Generative (predictive) models of predictive coding
Generative models: perceptual inference and active inference
We are constantly generating mental representations to predict future states of the external world and our internal world. To do this our brain must construct a ‘generative model’ to explain what gives rise to sensory inputs, so when given some sensory inputs, there cause can be inferred. In this process, each level of the processing hierarchy receives bottom-up sensory input from the level below and top-down predicted guesses from the level above; the bottom-up sensory inputs are then contrasted or “matched” against what our top-down predictions anticipated these inputs would be; if there is a mismatch between what these top-down learnt predictions anticipated, or guessed, what the information was going to be, and what the bottom-up input is saying the information actually is, then a ‘prediction error’ occurs; closing this error gap, through synaptic plasticity, to get better at predicting the future can be achieved by generative models (perceptual inference and active inference).
These two models constantly interact in the following way: perceptual inference: prior experience > mental representations > top-down predictions guess what bottom-up signals are going to be as a consequence of a movement being performed —> active inference: engages actions to instigate this movement, during the process of which, feedback (bottom-up signals), confirms or denies (prediction error) our perception-based predictions —> perceptual inference: this learning process updates our predictions by contextualising this sensory information in order to illuminate where our blindspots are. The aim is to select the appropriate action that confirms our perceptual-based predictions and minimise the prediction error.
For this process to occur smoothly we need good interoceptive accuracy (Baltazar et al 2021) and a symbiotic relationship between how well a motor action was performed, active inference, i.e. did the action produce the desired consequences, and the updating of our expectations from this, our perceptual inference, “if the actions did (or didn’t) produced the desired response I anticipate it will (or won’t) do again”.
Perceptual inference.
The perceptual process is an interaction between the brain’s model of what is to be expected, e.g. the expected chain of events, and its comparison to the actual sensory evidence (Paulus et al 2019). If interoceptive information conflicts with what we anticipated, what we always knew to be true, and can’t get explained away, it gets openly flagged up as a prediction error and transferred up the cortical hierarchy. Therefore our perception is not what we sense, but a compromise between the top-down expectations from our mental representations beliefs around what we should be sensing and the bottom-up sensation of what actually is experienced (Paulus et al 2019).
An optimal perceptual system ensures the least amount of energy will be spent in updating the prior models we have of ourselves and our environment in anticipation of, or during, a flood of sensory information. To accomplish this our top-down prior expectations about what the bottom-up sensory information is anticipated to be, or actually is, is deployed prior to our perception of it. This allows us to have in place our own unique top-down model already in situ and firing so that we can measure bottom-up sensory information against. This means sensory information that was predicted by our prior models, e.g. a typical heel strike whilst walking, can be kept in lower levels of consciousness, and sensory information that was unpredicted by our prior models e.g. unexpectedly missing a step, can be registered in higher levels of consciousness. In an optimal system when we experience any contradictions to our top-down prior expectations we should ease up on the precision and clarity of our prior expectations and tilt the scales to increase confidence in sensory information, e.g. unexpectedly missing a step means our prior model of the terrain was unreliable and we should rely on more sensory information derived from looking ahead for further obstacles, listening attentively to people pointing out further obstacles and proprioceptive feedback whilst walking. Conversely, less sensory precision, e.g. a vague noise, should tilt the scales to favour one’s confidence in precise top-down prior expectations to fill in the blanks, e.g. using prior knowledge to work out what the source of the noise could be. This natural inclination to conserve energy by minimising any unexpected, or unpredicted sensory information (prediction error), either during sensory stimulation, or, in anticipation of sensory stimulation, by either adaptively updating our prior expectations, or maladaptively by actively manipulating the sensory information (active inference), is inherent to the interpretative nature of top-down mechanisms of perception (Reeder et al 2024).
This efficient, conscious awareness of any deviations to what we predicted, and openly accept any conflicting sensations to what we anticipated enables us to rapidly make sense of sensory inputs as to see how precise our models of the world are we can adapt them, through, for example, contemplative techniques such as equanimity, curiosity, or acceptance (Khoury et al 2018). This allows us to generate the most accurate model of the world and help guide the most adaptive behaviour in nature’s inherently uncertain environments (Paulus et al 2019). By closing these prediction errors in this way uses a flexible mindset that allows our mental representations to be moulded by bottom-up interoceptive data so that mental representations and interoceptive data gelling into one. Therefore, confidence in raw, untainted, interoceptive information creates a ‘fluid interoceptive landscape’ so we can be flexible in our predictions and flexible in processing our experiences and expectations (Bohlen et al 2021) allowing us to grow and learn from our previous experiences (Brown & Brune 2012) using mindfulness principles (Khoury et al 2018).
Even though in order to build this cognitive flexibility we can interact with sensory information in multiple ways to define how we experience the richness of reality, this perceptual experience must be true and not delusional (Reeder et al 2024). For a ‘true’ perceptual experience we need to be able to successfully differentiate what is perceptually real (externally originated memories) and what is imagined, day-dreamed or symbolically represented in our mental imagery (internally originated memories) (Lavallé et al 2023). Being able to discern ‘fact from fiction’ allows us to identify the origin of the knowledge, attitudes and beliefs (Garrison et al 2017) by which we metacognitively encode, consolidate, and retrieve sensory information against, to help us learn how to predictable act in accordance to a stimulus. False memories, recalling false internally generated imaginary events as actually having happened, is influenced by prior knowledge, mental state, emotions, and context influence resulting in (i) flawed encoding of sensory information that constructs memories and integrates new information with old memories; (ii) an excessive familiarity and generalisation from the overall meaning derived from the general gist of events. This forms fuzzy representations of events without the clear context that would be derived verbatim from a more detailed representation; (iii) inaccurate reconstruction of memories during memory retrieval and flawed retrieval monitoring that remembers these false memories with great sensory detail (Lentoor 2023).
Therefore, memories don’t come with a label citing their source which has to be implied from cues relating to the overall nature and meaning of events (perceptual details), the direct facts verbatim (contextual detail) and the complex mental processes in encoding and recalling the information (cognitive operations) (Lavallé et al 2023). So trying to memorise a list ‘sparrow, hawk, robin’ (externally originated stimulus), involves encoding this sensory information verbatim and, in clear parallel, encoding the gist associated with the meaning behind these words e.g. they’re all birds (Lentoor 2023). However, when there’s not a distinct parallel, between these two memory packages, i.e. (i) an externally generated stimulus, the bird names on the list, and (ii) the internally generated meaning and associations derived from this list, the word ‘bird’, then these two separate memory packages merge into one so that, incorrectly, it is recalled that the word ‘bird’ was on the list. This faulty integration of what should have been separate and distinct memory packages as to blur fact from fiction (Lauzon et al 2022) in our perceptions is more potent if there’s a strong self-referential bias, e.g. in cases of a bird phobia! Even subtle errors in information gathering from being unable to tell if a memory was perceptually real or imagined, or, by the characteristic nature of the lens through which you appraise yourself and the event, can significantly alter how memories are reconstructed and retrieved (Lentoor 2023).
This reduced ability to monitor reality is utilised in mind control where through manipulation you are mistakenly made to believe you are the progenitor of your own thoughts, and is also characteristic of the delusions and/or hallucinations associated with psychosis. Delusions and hallucinations arise from individuals mistakenly imagining top-down internally generated information as having originated from real bottom-up information derived from the external world. A lack of trust in their ability to interpret the bottom-up sensory information (Reeder et al 2024) that would otherwise generate more ‘true’ externally generated memories (Lavallé et al 2023) leads to a decrease confidence in this sensory information and a tendency for the individual to contract within themselves and become over-reliant on the top-down priors (Reeder et al 2024) that generates internally derived delusional, but more imaginative, memories (Lavallé et al 2023). As these internal high-level priors are resistant to being updated from any mistrusted external sensory information there’s a stronger focus on maladaptive top-down beliefs over bottom-up sensory evidence (Reeder et al 2024) and in the absence of a unitary experience of self‐agency from the ability to reliably predict the outcome of self‐generated actions (Lavallé et al 2023) a need for active inference to live out self-fulfilling prophecies.
Active inference.
Active inference involves performing actions based on our expected observations, e.g. checking to make sure something is where we thought we put it, in order to produce sensory input, e.g. visual confirmation, to test and resolve any differences between our predicted and actual observations, e.g. it’s where I thought it was; an important point to note is that the motor output that produces these actions, that we gain sensory information from, is not fired from a point of ‘neutrality’, or reality, but is fired from the expectations and predictions that characterise our mental representations i.e. we actively performed a motor action to confirm something was where we expected it to be in order to resolve any discrepancies between expectations and actual observations. In order to gain or learn from the sensory input from these motor actions they have to produce the desired consequences, e.g. to see if something is where we thought it was we’re better off using our eyes than our ears, and we have know how certain or uncertain the consequences of these actions are (Fradkina et al 2020). As our motor and autonomic nervous system are constantly firing, constantly predicting, expecting, what the sensory feedback will be from these actions, it maintains the system in an expected state (Gladziejewski 2015).
Although homeostasis entails maintaining physiologically essential internal variables (e.g. glucose level, blood pressure) within tight ranges all the time allostasis deals with the regulation of bodily states through change; therefore active inference, by performing actions to change incoming sensory data can ‘head off’ any undesirable changes (Ciaunica et al 2019). When these predicted sensory consequences of active inference don’t come to fruition, e.g. ‘as a consequence of my actions I though x would happen but y happened instead’, active inference fails to head off any undesirable changes. This highlights an error in this system (prediction error) that should refer back to perceptual inference in order to adjust the individual’s mindset, so that models or representations used to predict the sensory consequences of actions promote flexible learning (Smith et al 2020).
However, being unprepared to close the prediction error by using a flexible mindset to change our top-down predictions (high interoceptive accuracy) the prediction error can only be closed by changing the bottom-up input (low interoceptive accuracy). To do this active inference can actively drive activity so sensory data can be suppressed, reappraised, distracted from (Khoury et al 2018) or motor activity can actively change it. Being confident in the belief of ones predictions, ‘the things we always knew to be true’, can form dysfunctional beliefs and expectations that become immune to any exteroceptive or interoceptive information that challenges them. Therefore, to draw, what would liked to be, the ‘right’ conclusions, involves actively changing either exteroceptive or interoceptive sensations to meet prior expectations (Bohlen et al 2021) to feverishly keep the individual within expected bounds. Conveniently ‘flipping’ interoceptive information in this way, so ‘what we always predicted can come true’, excites autonomic reflexes that creates visceral feelings; these feelings fulfil the needs of the anticipated predictions as to avoid surprise and experience an emotional or motivational moment in time; this experienced moment will project forwards to influence what is felt in the future in an elegantly orchestrated self-fulfilling prophecy; for example actively drawing attention (e.g. auditory or visual) to information that supports and maximises the expectations that appeal to biases (perceptual inference) and therefore fails to illuminate where the individual’s blindspots are.
This can be seen in visual illusions, e.g. the ponzo illusion, where strong priors are based on learned stimulus regularities, i.e. if there are two objects of equal length one close to you and one far away, the one farther away, should appear smaller. This belief, that is true in a three-dimensional world, is so strong that it transferred to a two dimensional world, i.e. a drawing on a piece of paper. Therefore, when two lines of equal length are drawn on to a piece of paper so that one line looks closer than the other, the sensory information that can see the two lines are of equal length is actively over-riden by the belief that the line that seems more distant is bigger than the line that seems closest (refer ‘figure one’) (Reeder et al 2024). The alternative to using active inference to ‘head off’ contradicting sensory information is to experience repeated prediction errors in the form of failure in the external world (i.e. falling short of an objective measure) or in the internal world (i.e. failing to regulate homeostasis in our nervous system) that can, given a certain mindset, go from being informative to creating mental health problems (Paulus et al 2019).
Figure one: Example of active inference. The Ponzo Illusion.
Generative models and Obsessive Compulsive Disorder
In Obsessive Compulsive Disorder there is a high level of confidence in expecting a diverse chain of damaging events (e.g. because of … you might not have checked it properly, if you don’t check it properly highly precise catastrophic consequences are predicted to happen). What makes these highly precise predictions about what ‘is’ going to happen so catastrophic is that the individual feels out of their depth and intimidated in an environment with potential danger around every corner. ‘Being out of their depth’ centres around fear and anxiety that any learning from prior experiences that makes guesses as to what motor action will produce what sensory information, their memory, (e.g. checking once should foster a certainty that the task has been safely completed) is deemed irrelevant; it is believed these ill-informed predictions will leave the individual unable to navigate this threatening environment (Fradkina et al 2020) and there is an inability to form new maps that predict how actions can successfully adapt to the uncertainty and complexity of the terrain (Sharp et al 2021). Therefore, feeling ill-equipped and out of their depth, unable to predict how their actions will perform in, and how to adapt their actions to successfully interact with, the environment, gives a sense of abandonment so certainty can only be obtained by collecting high value sensory information from recent observations (e.g. further checking); this in turn makes the individual hyper-responsive to sensory information, that is finely combed for any imperfections, giving excessive weight to any unpredictabilities and prediction errors (Fradkina et al 2020).
However, no matter how well motor actions are performed to gain highly valued, but volatile, sensory input, the sensory stimuli struggles to achieve its goal in experiencing the sufficient approval and certainty required to reduce the prediction error and quench the anxiety of these catastrophic predictions (e.g. checking has to elicit the ‘perfect’ sensory stimuli); this incites further motor action (e.g. further checking) and also validates jumping between an array of different actions (e.g. I predict checking in this way, as well as that way, should provide certainty) in order to gather more high valued self-confirming sensory information.
This whole process reflects a failure in the use of learning from prior experiences (perceptual inference) to reappraise and re-contextualise sensory information (active inference) to reduce the prediction error, and make even trivial deviations from predictions feel more catastrophic than what they are; with a lack of self confidence in the individual’s ability to use their prior experiences, their memory, to predict what the sensory effects of their motor actions means, and what change they will produce in their environment as a consequence of them actions, means they can’t instigate goal directed forward thinking with any level of certainty. This creates feelings of incompletion and a lack of direction leaving the individual caught in a loop, unable to move on. This uncertainty in using their adaptive goal-orientated forward thinking results in indecisiveness and volatile behaviour, and so logically, there develops a reliance on self-confirming sensory stimulation. Being reliant on self-confirming sensory stimulation can take the form of simpler habitual behaviour (Fradkina et al 2020) that in response to a stimulus reduces intolerance to uncertainty with a comforting ritualistic, stereotyped behaviour that gathers familiar and highly predictable information in a stable way. As a general rule simple models, so long as they provide enough certainty and accuracy, will always be favoured by the brain as they have a very low complexity cost. This means their ritualistic ‘that’s good enough’ motor actions, or behaviour, they prescribe e.g. overgeneralised, lazy stereotypes, lacks the more flexible thinking that is adaptive towards a desired goal; however in day-to-day use these simple models allows for an overall broader behavioural repertoire, e.g. ritualistic behaviour in driving uses minimal attention freeing up capacity for more complex models, e.g. looking out for hazards on the road (FitzGerald et al 2014). This sensory information gathering is seen in other anxiety disorders where individuals can even invest in actions with uncertain short-term consequences in order to gain informative input that can provide long-term certainty (Teng et al 2016).
Treatment for impaired Interoeption
Impaired interoception can be from the physiological signal itself, perception (consciously or subconsciously) of the signal, or identification (labelling/recognition) of the signal (Brewer et al 2021).
Increasing interoceptive sensitivity to positive and negative emotions can increase self-regulation and learning (Brewer et al 2021), promote pain tolerance, contextualise worry about pain sensations, foster a healthy emotional awareness and place more trust and attention to bodily information (Joshi et al 2021). This can be seen in optimistic individuals who translates goals into behaviours through reinforced learning by being sensitive to, and being able to appraise any inconsistencies between their expectations and what turns out to be the reality. Having good interoceptive accuracy, enables them to be sensitive to, and learn from these prediction errors so they can move on from any previously anticipated worries by experiencing the reality and creating positive-rewarding expected values for future thoughts; this can be seen even when the reality confirms any previously anticipated worries but is appraised positively as an essential learning experience that is crucial for creating positive expectations for the future. In contrast, less optimistic individuals, with a hypervigilant threat appraisal system that places a higher value on anticipating and worrying about an up-coming event are less sensitive to learning from any contradicting, positive, bottom-up information that occurs during the actual event; this fails to challenge the frustrations from their dysfunctional patterns and update their pessimistic mental representations (White et al 2016).
This is why, when superimposing external symbolic interoceptive information to compensate for interoceptive dysfunctional patterns the following principles must be adhered to: (i) ‘a seed can not grow in stone’ the individual must be malleable to external input to determine their own interoceptive information; (ii) ‘drip feed in manageable instalments’ this external input must feed information into the subordinate interoceptive level that can be consistently integrated in the interoceptive matrix, without excessively violating the internal consistency and causing substantial prediction errors. This enables the interoceptive system to adapt without triggering a cascade of stressful autonomic responses by oscillating between activation and deactivation to improve the balance of the autonomic system, re-framing the interoceptive representations (Di Lernia et al 2016).
Mindfulness.
Mindfulness practice aims to promote emotional regulation and a more open and inquisitive mindset. This means that conclusions don’t have to be drawn from mental representations formed from worry and rumination. This enables the individual to be truly self-confident in embracing the complexity of a constantly shifting external world (Joshi et al 2021).
This positive framework, entertaining curiosity towards interoceptive, exteroceptive and proprioceptive sensations, promotes the use of bodily information as a useful resource. By adapting internal representations to varying contexts, in order to re-appraise and re-interpret interoceptive information, it enables the individual to identify interoceptive information and effectively intervene (Brewer et al 2021). This can be encouraged by using vivid mental imagery with its associated somatic and emotional evaluation to encourage concrete logical thinking, that can otherwise be suppressed, by worry, characterised by more generalised verbal thinking (Teng et al 2016) that effortlessly skips from one negative topic to another, impeding problem solving and trapping the individual in a worry cycle (Hirsch et al 2019).
Therefore, using mindfulness to encourage a more flexible mindset adapts and changes to bottom-up information in order to learn how to reduce the discrepancy between perceived and true interoceptive information. This is opposed to trying and reduce the discrepancy between perceived and true interoceptive information by suppressing bottom-up information to meet a rigid inflexible mindset. Shifting interoceptive information in this way from being threatening, promoting fear-based reactions, to being informative promotes flexible adaptivity (Joshi et al 2021) and learning (Brewer et al 2021) by cementing positive stimuli and reappraising negative stimuli as a positive learning experience.
Attentional Bias Modification Training.
When worry and rumination cumulates in high levels of threat and uncertainty the individual becomes highly motivated to pre-empt and avoid errors by focusing on negative, threat-related information. This heightened sensitivity to errors can be reduced by Attentional Bias Modification Training that attempts to reduce the tendency of individuals to selectively allocate attention to negative, threat-related information (Klawohna et al 2020).
Osteopathy.
Osteopathy can effect interoceptive sensitivity by (Bohlen et al 2021):
Deep touch and osteopathic mobilisation significantly increases interoceptive accuracy in patients with chronic low back pain.
Myofascial release techniques increase interoceptive sensitivity, but not significantly.
Afferent C–tactile fibers are stimulated during affective, low force, dynamic touch possibly playing a role in restoring impaired interoceptive function, body awareness, homeostatic regulation, allostasis, emotion, and affective disorders. The primarily function of C-tactile afferents is to provide information about the homeostatic and emotional effects of touch, rather than the properties of what or who is being touched. Activation of these C-tactile receptors on the skin then may specifically relate to the positive consequences of interpersonal touch, such as reducing feelings of social exclusion, soothing pain and communicating social support (Ciaunica et al 2021).
Social affective touch can start in foetal development when the foetus, entirely covered in fine hair (lanugo hairs), moves in the amniotic fluid to directly stimulate these C-tactile afferents; this activates the hypothalamus and insular cortex that promotes an anti-stress effect via the release of oxytocin and stimulates the foetal growth. Also, foetuses spend a significant amount of time in tactile exploration frequently touching certain body areas, such as the lips, cheeks, ears, and parietal bone, creating a self-stimulatory pattern, which enhances innervation. Importantly, when the foetus touches the forehead, innervation increases, and the boundary migrates allowing the foetus to move on to touch a new innervated boundary until the whole body is fully innervated. Additionally, it has been shown that maternal touch of her own abdomen increases arm, head, and mouthing movements in the foetus having more impact than maternal voice in the foetus's movements. Later in development, well before humans are able to recognise themselves in a mirror to perceive themselves from a visuospatial perspective, they experience themselves and their surroundings via the proximal senses where touch represents a fundamental step in the development of both self- and other-awareness, as well as self-other distinction (Ciaunica et al 2021).
References
Di Lernia D, Serino S, Cipresso P & Riva G (2016). Ghosts in the Machine. Interoceptive Modeling for Chronic Pain Treatment
Bohlen L, Shaw R, Cerritelli F & Esteves J (2021). Osteopathy and Mental Health: An Embodied, Predictive, and Interoceptive Framework
Paulus M, Feinstein J & Khalsa1 S (2019). An Active Inference Approach to Interoceptive Psychopathology
Seth A & Friston K 2 (2016). Active interoceptive inference and the emotional brain
Unalac O, Caner O, Göktuğ E, Frederike A, Petzschnera H, Klaas Y & Stephanab E (2021). Inference on homeostatic belief precision
Seth A & Tsakiris M (2018). Being a Beast Machine: The Somatic Basis of Selfhood
Brown E & Brüne M (2012). The role of prediction in social neuroscience
Lim M, O'Grady C, Cane D, Goyal A, Lynch M, Beyea S & Javeria Hashmi J (2020). Threat Prediction from Schemas as a Source of Bias in Pain Perception
Babić R, Babić M, Rastović P, Ćurlin M, Šimić J, Mandić K, Katica Pavlović K (2020). Resilience in Health and Illness
Bohlen L, Shaw R, Cerritelli F & Esteves J (2021). Osteopathy and Mental Health: An Embodied, Predictive, and Interoceptive Framework
Liesner M & Kunde W (2021). Environment-Related and Body-Related Components of the Minimal Self
Haar, S., van Assel, C.M., Faisal, A.A., 2020. Motor learning in real-world pool billiards.
Massen C and Prinz W (2008). Movements, actions and tool-use actions: an ideomotor approach to imitation
Murphy J, Catmur C,Bird G (2019). Classifying individual differences in interoception: Implications for the measurement of interoceptive awareness
Garfinkel S, Seth A, Barrett A, Suzuki K, Critchley H (2015). Knowing your own heart: distinguishing interoceptive accuracy from interoceptive awareness
Tsakiris N (2017). The multisensory basis of the self: From body to identity to others
Brewer R, Murphy J, Bird G (2021). Atypical interoception as a common risk factor for psychopathology: A review
Gładziejewski P (2015). Predictive coding and representationalism
Joshi V, Graziani P & Del-Monte J (2021). The Role of Interoceptive Attention and Appraisal in Interoceptive Regulation
Paulus M & Stein M (2006). An insular view of anxiety
Blakemore S, Smith J, Steel R, Johnstone C, Frith C (2000). The perception of self-produced sensory stimuli in patients with auditory hallucinations and passivity experiences: evidence for a breakdown in self-monitoring
van Kemenade B, Arikan B, Kircher T & Straube B (2016). Predicting the sensory consequences of one’s own action: First evidence for multisensory facilitation
Smith R, Schwartenbeck P, Parr T & Friston K (2020). An Active Inference Approach to Modeling Structure Learning: Concept Learning as an Example Case
BALTAZAR M, GREZES J, GEOFFRAY M, PICQ J & CONTY L (2021). NEURAL CORRELATES OF INTEROCEPTIVE ACCURACY: beyond cardioception
White S, Geraci M, Lewis E, Leshin J, Teng C, Averbeck B, Meffert H, Ernst M, Blair J, Grillon C, Blair K (2016). Prediction Error Representation in Individuals With Generalized Anxiety Disorder During Passive Avoidance
Klawohna J, Hajcak G, Amird N, Kathmanna N, Riesele A (2020). Application of attentional bias modification training to modulate hyperactive error-monitoring in OCD
Teng C, Otero M, Geraci M, Blair R, Pine D, Grillon C, and Blair K (2016). Abnormal decision-making in generalized anxiety disorder: Aversion of risk or stimulus-reinforcement impairment?
Hirsch C, Beale S, Grey N and Liness S (2019). Approaching Cognitive Behavior Therapy For Generalized Anxiety Disorder From A Cognitive Process Perspective
Uphill M, Rossato C, Swain J, O’Driscoll J (2019). Challenge and Threat: A Critical Review of the Literature and an Alternative Conceptualization
Ciaunica A, Roepstorff A, Fotopoulou A,4Petreca B (2021). Whatever Next and Close to My Self—The Transparent Senses and the “Second Skin”: Implications for the Case of Depersonalization
Ishikawa R, Ayabe-Kanamura S, Izawa J (2021). The role of motor memory dynamics in structuring bodily self-consciousness
Klar P. Northoff G (2021). When the World Breaks Down: A 3-Stage Existential Model of Nihilism in Schizophrenia
Ciaunica A & Crucianelli L (2019). Minimal Self-Awareness: from Within A Developmental Perspective
Forch V, Hamker F (2021). Building and Understanding the Minimal Self
Riva G (2008). Enacting interactivity: The role of presence
Riva G (2018). The neuroscience of body memory: From the self through the space to the others
Paulus M, Feinstein J, Khalsa S (2019). An Active Inference Approach to Interoceptive Psychopathology
Fradkina I, Parrf T, Adamsbcde R, Roiserb J (2020). Searching for an anchor in an unpredictable world: A computational model of obsessive compulsive disorder
Sharp P, Dolan R, Eldar E (2021). Disrupted state transition learning as a computational marker of compulsivity
FitzGerald T, Dolan R, Friston K (2014). Model averaging, optimal inference, and habit formation
Khoury N, Lutz J, Schuman-Olivier Z (2018). Interoception in Psychiatric Disorders: A Review of Randomized Controlled Trials with Interoception-based Interventions
Braun N, Debener S, Spychala N, Bongartz E, Sörös P, Müller H, Philipsen A (2018). The Senses of Agency and Ownership: A Review
Wang W, Pierce J, Li D, Wang G, Li J, Niu X (2021). Eliciting Psychological Ownership of Object by Marking Organizational Name: The Role of Belongingness
Reeder RR, Sala G, van Leeuwen TM. (2024). A novel model of divergent predictive perception.
Lentoor A (2023). Cognitive and neural mechanisms underlying false memories: misinformation, distortion or erroneous configuration?
Garrison JR, Bond R, Gibbard E, Johnson MK, Simons JS. (2017). Monitoring what is real: The effects of modality and action on accuracy and type of reality monitoring error.
Lavallé L, Brunelin J, Jardri R, Haesebaert F, Mondino M. (2023). The neural signature of reality-monitoring: A meta-analysis of functional neuroimaging studies
Lauzon SA, Abraham AE, Curcin K, Butler BE, Stevenson RA. (2022). The relationship between multisensory associative learning and multisensory integration.