Sunday, March 09, 2014

Alejandra Sel - Predictive Codes of Interoception, Emotion, and the Self - A Commentary on Anil K. Seth


This is an interesting discussion between Anil K Seth and Alejandra Sel on Seth's recent paper in Trends in Cognitive Science (pdf). Here is a brief summary of Seth's position, as understood by Sel:
Seth's proposal that sensory processing involves predictions is nothing new. What is new in Seth's model is that perception of internal body signals (interoception), paralleling the perception of external signals, relies on top-down predictions of the causes of the sensory input, rather than being a passive, bottom-up process.
While Sel agrees in principle with Seth's position, there remain four points [assumptions] that Sel feels the need to address before launching any studies to validate Seth's model.
1. [E]motions are defined as affective states relying on interactions between top-down interoceptive predictions and bottom-up interoceptive prediction errors.

2. Seth's model refers to the anterior insular cortex [AIC] as the key structure that generates, compares, and updates interoceptive predictions. Empirical evidence has shown that AIC houses a secondary associative area where interoceptive, exteroceptive, and motivational signals converge (Seth and Critchley, 2013).

3. [A]lthough a free-energy model of self has been proposed (Apps and Tsakiris, in press), as yet there is no evidence to suggest that self-processing follows the principles of predictive coding [PC], [as implied by Seth's model].

4. An individual's attention to the body can be significantly enhanced by the practice of Mindfulness (Farb et al., 2013), which also has the effect of enhancing both cortical responses of interoceptive attention and self-reported interoceptive awareness (Mehling et al., 2013). Within Seth's model this might increase the accuracy of interoceptive inference, emotions, and self-awareness.
The original article in only 9 pages including references, and the commentary below is brief. It's cool to see ideas proposed and addressed in an open forum.

Full Citation: 
Sel A. (2014, Mar 4). Predictive codes of interoception, emotion, and the self. Frontiers in Psychology: Cognitive Science; 5:189. doi: 10.3389/fpsyg.2014.00189

Glossary (from Seth)

  • Active inference: an extension of PC (and part of the free energy principle), which says that agents can suppress prediction errors by performing actions to bring about sensory states in line with predictions. 
  • Augmented reality: a technique in which virtual images can be combined with real-world real-time visual input to create hybrid perceptual scenes that are usually presented to a subject via a head-mounted display. 
  • Appraisal theories of emotion: a long-standing tradition, dating back to James (but not Lange), according to which emotions depend on cognitive interpretations of physiological changes. 
  • Emotion: an affective state with psychological, experiential, behavioral, and visceral components. Emotional awareness refers to conscious awareness of an emotional state. 
  • Experience of body ownership (EBO): the experience of certain parts of the world as belonging to one’s body. EBO can be distinguished into that related to body parts (e.g., a hand) and a global sense of identification with a whole body. 
  • Free energy principle: a generalization of PC according to which organisms minimize an upper bound on the entropy of sensory signals (the free energy). Under specific assumptions, free energy translates to prediction error. 
  • Generative model: a probabilistic model that links (hidden) causes and data, usually specified in terms of likelihoods (of observing some data given their causes) and priors (on these causes). Generative models can be used to generate inputs in the absence of external stimulation. 
  • Interoception: the sense of the internal physiological condition of the body. 
  • Interoceptive sensitivity: a characterological trait that reflects individual sensitivity to interoceptive signals, usually operationalized via heartbeat detection tasks. 
  • Predictive coding (PC): a data processing strategy whereby signals are represented by generative models. PC is typically implemented by functional architectures in which top-down signals convey predictions and bottom-up signals convey prediction errors. 
  • Rubber hand illusion (RHI): a classic experiment in which the experience of body ownership is manipulated via perceptual correlations such that a fake (i.e., rubber) hand is experienced as part of a subject’s body. 
  • Selfhood: the experience of being a distinct, holistic entity, capable of global self-control and attention, possessing a body and a location in space and time [64]. Selfhood operates on multiple levels – from basic physiological representations to metacognitive and narrative aspects.
  • Subjective feeling states: consciously experienced emotional states that underlie emotional awareness. 
  • Von Economo neurons (VENs): long-range projection neurons found selectively in hominid primates and certain other species. VENs are found preferentially in the AIC and ACC. 
* * * * *

    Predictive codes of interoception, emotion, and the self

    Alejandra Sel
    • Department of Psychology, Royal Holloway University of London, Egham, Surrey, UK
    A commentary on:
    Interoceptive inference, emotion, and the embodied self, by Seth, A. K. (2013). Trends Cogn. Sci. 17, 565–573. doi: 10.1016/j.tics.2013.09.007

    Interoception is the ability to perceive and integrate physiological signals from within the body. It is closely related to the autonomic system and is a key component in the generation of affective states and abstract representations of the self (Critchley et al., 2004; Ainley and Tsakiris, 2013). Seth proposes a predictive coding (PC) model of interoception that involves a free-energy based explanation of emotion awareness and selfhood. In this model, emotions, and in turn the sense of self, rely on predictions of the causes of interoceptive signals. Within this framework, the interoceptive system minimizes free-energy, or the discrepancy between predictions and interoceptive signals. Free-energy can be minimized either by updating predictions about the causes of the sensory signals (perceptual updating), or by acting to change autonomic states such that bodily states are more predictable (active inference).

    The free-energy principle is currently in vogue in neuroscience. We are no longer strangers to the idea that perception is an active iterative process between abstract representations (predictions) and sensory feedback (prediction errors) (Clark, 2013). The basic idea of PC in the cognitive sciences began with the notion of neural energy (Helmholtz, 1860) and it has been present since in the form of theoretical proposals and empirical findings, especially in the visual domain (Lee and Mumford, 2003). Therefore Seth's proposal that sensory processing involves predictions is nothing new. What is new in Seth's model is that perception of internal body signals (interoception), paralleling the perception of external signals, relies on top-down predictions of the causes of the sensory input, rather than being a passive, bottom-up process.

    Is then Seth's interoceptive inference model an interesting proposal to explain emotion awareness and selfhood? My opinion is yes and that it is worth investigating. However, there are some aspects to consider before designing studies to empirically test Seth's model.

    Seth's model builds on three main assumptions. First, emotions are defined as affective states relying on interactions between top-down interoceptive predictions and bottom-up interoceptive prediction errors. Following the principles of PC, there is a constant attempt to minimize the discrepancy between the predicted and the actual sensory events, either through updating perceptual expectations or through active inference (Friston et al., 2010). As Seth nicely explains, active inference in interoception occurs when predictions are transcribed into reference points that trigger autonomic homeostatic regulation, occurring when the weight of the error is low and attention to errors is attenuated (Gu et al., 2013).

    Fortunately, advances on biomedical tools allow us to experimentally monitor the body's physiological signals. Although, some methodological challenges still remain when investigating interoception. This general issue may also impact on PC studies of interoception. However, applying PC to interoception, as proposed in Seth's model, may allow us to overcome these challenges. The main argument of PC is that all sensory systems are linked by working under identical code schemes (Friston and Kiebel, 2009). Therefore, Seth's PC model allows us to apply knowledge from visual and other domains to investigate brain and behavioral mechanisms of interoception. Neuroimaging studies have demonstrated direct evidence of PC in visual brain areas (Egner et al., 2010; Wyart et al., 2012). Likewise, Seth's anatomical predictions (i.e., anterior insular cortex -AIC) can be tested by using multivoxel pattern analysis approaches, in combination with orthogonal experimental designs where the stimulus presentation probability is held constant in all conditions (Egner et al., 2010).

    The second assumption in Seth's model refers to the AIC as the key structure that generates, compares, and updates interoceptive predictions. Empirical evidence has shown that AIC houses a secondary associative area where interoceptive, exteroceptive, and motivational signals converge (Seth and Critchley, 2013). An important principle of PC explains that the surprisal generated in one unimodal system can be explained away by inferences in other system via high-order neural areas (Apps and Tsakiris, in press). Considering the multimodal nature of the AIC, one could suggest that the errors in the interoceptive signal can be explained by exteroceptive inferences (or vice versa) and that the interoceptive generative models are only a part of the way the system explains errors. Whether the AIC exclusively codes the surprisal evoked by interoceptive signals or, alternatively, if the AIC is involved in top-down general predictions directed to a more specialized interoceptive circuit, still remain open questions.

    The third crucial aspect of Seth's model is the concept of selfhood. Seth has employed the idea that selfhood is formed by the integration of predictive interoceptive and exteroceptive signals (Tajadura-Jimenez and Tsakiris, in press). Individual differences in the accuracy of interoceptive awareness influence integration of interoceptive and exteroceptive information, as shown by studies in body illusions (Tsakiris et al., 2011). Individuals with low accuracy show more susceptibility to body illusions, which Seth interprets as lower precision-weighting of interoceptive prediction errors. However, although a free-energy model of self has been proposed (Apps and Tsakiris, in press), as yet there is no evidence to suggest that self-processing follows the principles of PC.

    Another crucial factor that may influence interoceptive awareness, and therefore self-awareness, is attention. In PC, attention is considered to be a mechanism that optimizes the precision of prediction errors during hierarchical inference (Feldman and Friston, 2010). For example, studies in vision have demonstrated that attention enhances the neural specificity for expected vs. unexpected stimuli in visual cortex (Jiang et al., 2013). Similarly, directing attention toward internal body signals might increase the precision of interoceptive prediction errors and therefore improve interoceptive awareness. An individual's attention to the body can be significantly enhanced by the practice of Mindfulness (Farb et al., 2013), which also has the effect of enhancing both cortical responses of interoceptive attention and self-reported interoceptive awareness (Mehling et al., 2013). Within Seth's model this might increase the accuracy of interoceptive inference, emotions, and self-awareness.

    Therefore, I agree with Seth's proposal that the brain is a prediction machine that integrates interoceptive and exteroceptive information in a Bayesian way. However, future research is needed to elucidate the internal properties of the interoceptive inference.

    Acknowledgments

    This work was supported by the European Research Council Starting Investigator Grant (ERC-2010-StG-262853). I would like to thank Manos Tsakiris and the reviewer for their insightful comments and Lara Maister and Vivien Ainley for their help with manuscript editing.


    References


    Ainley, V., and Tsakiris, M. (2013). Body conscious? Interoceptive awareness, measured by heartbeat perception, is negatively correlated with self-objectification. PLoS ONE 8:e55568. doi: 10.1371/journal.pone.0055568 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Apps, M., and Tsakiris, M. (in press). The free-energy self: a predictive coding account of self-recognition. Neurosci. Biobehav. Rev. doi: 10.1016/j.neubiorev.2013.01.029 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behav. Brain Sci. 36, 181–204. doi: 10.1017/S0140525X12000477 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Critchley, H. D., Wiens, S., Rotshtein, P., Ohman, A., and Dolan, R. J. (2004). Neural systems supporting interoceptive awareness. Nat. Neurosci. 7, 189–195. doi: 10.1038/nn1176 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Egner, T., Monti, J. M., and Summerfield, C. (2010). Expectation and surprise determine neural population responses in the ventral visual stream. J. Neurosci. 30, 16601–16608. doi: 10.1523/jneurosci.2770-10.2010 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Farb, N. A. S., Segal, Z. V., and Anderson, A. K. (2013). Mindfulness meditation training alters cortical representations of interoceptive attention. Soc. Cogn. Affect. Neurosci. 8, 15–26. doi: 10.1093/scan/nss066 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Feldman, H., and Friston, K. J. (2010). Attention, uncertainty, and free-energy. Front. Hum. Neurosci. 4:215. doi: 10.3389/fnhum.2010.00215 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Friston, K., and Kiebel, S. (2009). Predictive coding under the free-energy principle. Philos. Trans. R. Soc. B Biol. Sci. 364, 1211–1221. doi: 10.1098/rstb.2008.0300 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Friston, K. J., Daunizeau, J., Kilner, J., and Kiebel, S. J. (2010). Action and behavior: a free-energy formulation. Biol. Cybern. 102, 227–260. doi: 10.1007/s00422-010-0364-z Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Gu, X., Hof, P. R., Friston, K. J., and Fan, J. (2013). Anterior insular cortex and emotional awareness. J. Comp. Neurol. 521, 3371–3388. doi: 10.1002/cne.23368 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Helmholtz, L. F. V. (1860). Handbuch der Physiologischen Optik [Handbook of Physiological Pptics]. Leipzig: Voss.

    Jiang, J., Summerfield, C., and Egner, T. (2013). Attention sharpens the distinction between expected and unexpected percepts in the visual brain. J. Neurosci. 33. 18438–18447. doi: 10.1523/jneurosci.3308-13.2013 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Lee, T. S., and Mumford, D. (2003). Hierarchical Bayesian inference in the visual cortex. J. Opt. Soc. Am. A Opt. Image Sci. Vis. 20, 1434–1448. doi: 10.1364/JOSAA.20.001434 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Mehling, W. E., Daubenmier, J., Price, C. J., Acree, M., Bartmess, E., and Stewart, A. L. (2013). Self-reported interoceptive awareness in primary care patients with past or current low back pain. J. Pain Res. 6, 403–418. doi: 10.2147/JPR.S42418 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Seth, A. K., and Critchley, H. D. (2013). Extending predictive processing to the body: emotion as interoceptive inference. Behav. Brain Sci. 36, 227–228. doi: 10.1017/S0140525X12002270 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Tajadura-Jimenez, A., and Tsakiris, M. (in press). Balancing the “Inner” and the “Outer” self: interoceptive sensitivity modulates self-other boundaries. J. Exp. Psychol. Gen. doi: 10.1037/a0033171 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Tsakiris, M., Tajadura-JimĂ©nez, A., and Costantini, M. (2011). Just a heartbeat away from one's body: interoceptive sensitivity predicts malleability of body-representations. Proc. Biol. Sci. 278, 2470–2476. doi: 10.1098/rspb.2010.2547 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text

    Wyart, V., Nobre, A. C., and Summerfield, C. (2012). Dissociable prior influences of signal probability and relevance on visual contrast sensitivity. Proc. Natl. Acad. Sci. U.S.A. 109, 3593–3598. doi: 10.1073/pnas.1120118109 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text
    Post a Comment