During my senior year at Northwestern, I took a class on the Affective Neuroscience with Professor Robin Nusslock. It was easily one of my favorite and most intellectually inspiring courses I took while a student there. Here is a midterm review paper I wrote on a particular neural structure’s role in subjective awareness.

Introduction

Consciousness is a topic that psychologists, philosophers and neuroscientists often consider to be one the more difficult mechanisms to probe. When studying this, operational definitions of what ”consciousness” is must be employed to concretize findings. Here, in accordance with major the- ories of today, I will use the term subjective awareness or feeling states to describe the individual, emotionally-tinged feelings of experience (Damasio 1996; Craig 2009).

The James-Lange theory of emotion, as well as Damasio’s ”somatic marker” hypothesis heavily associate bodily or physiological state with the experience of an emotion. Developments in sen- sory, affective and social neuroscience has implicated the insula in interoception and the resulting emotional experience (Singer et al., 2009). Particularly, activity in the anterior insula (AI) reflects the subjective intensity of one’s own emotional experiences (Wicker et. al. 2003), suggesting its role in the feeling states each individual exists in.

Here, I review structural and functional findings about the insular cortex before presenting theories about how the insular cortex may produce these feeling states. Last, I present evidence about AI sensitivity to uncertainty, which quantifies how feeling states may be calculated.

Structure and Function of the Insular Cortex

Given its central location and high frequency of activation in fMRI research, the insula has been considered as a critical functional hub in the human brain (Nomi et al. 2016; Chang et al., 2012).

Insular cortex has been demonstrated to be a functionally and cytoarchitecturally heteroge- neous region consisting of multiple distinct subdivisions (Deen et al., 2011; Chang et al., 2012; Uddin et al., 2014; Nomi et al., 2016). Functional connectivity analyses based on magnetic res- onance imaging coactivations have found two, three or four distinct subregions depending on the parcellation strategy used (Nomi et al., 2016). Over most of these studies, a division is found between the posterior insula (PI) and the AI (Cauda et al. 2011). The AI is further divided into the dorsal AI (dAI) and ventral AI (vAI). Some studies delineate the central or mid-insula as the fourth division (Kurth et al. 2010) . At resting-state, the PI is primarily connected to the primary and secondary somatomotor cortices and receives input from both interoceptive and exteroceptive sensory systems of affective or motivational significance (Craig 2002, 2009); the dAI is primarily connected to frontal, dorsal anterior cingulate and parietal areas; the vAI is primarily connected to limbic areas, such as the amygdala, and the anterior cingulate cortex (ACC) (Deen et al. 2011, Chang et al., 2012). More recent studies using a meta-analytic approach or data from subsets of the resting-state fMRI data demonstrate a significant degree of overlapping and transient connectivity profiles (Uddin et al. 2014; Nomi et al., 2016).

The insular cortex is broadly accepted as the viscerosensory cortex and is involved in mapping interoceptive cues (such taste and pain), representing emotional arousal and feelings and process- ing information about uncertainty (Craig 2002, 2009; Critchley 2005). These divisions determined by studies of co-activation support a cognition-emotion-interoception model of the human insular cortex. In this model, the dAI, given its associated frontal connections, is implicated in higher- order, task-dependent control of goal-directed behavior and sensory processing (Dosenbach et al., 2007). The vAI, given its limbic connections, is associated with affective processes. For instance, Mutschler et al. (2009) showed in an activation likelihood estimate meta-analysis that the peak coordinates relating neural activity with galvanic skin response and heart rate were located in the vAI. Lastly, the PI provides the basis for the sense of the physiological condition of the entire body (Craig 2009). This serves the interoceptive role of the human insula. In models with four subdivi- sions, the last region is a distinction from the interoceptive insular region that is more specific to olfactory and gustatory sensation.

Lastly, it is important to note the relationship between the insula and the anterior cingulate cortex (ACC). The insula is co-activated with the ACC in virtually all studies of awareness, sug- gesting that the two areas are responsible for separate processes that are very closely linked. The insula and ACC are regarded as the limbic sensory and motor cortices that engender feeling and motivation/action, respectively, to constitute an emotion (Critchley 2005; Craig 2009).

Modeling Subjective and Emotional Awareness in the AIC

Insular cortex, within the networks it belongs to, is a part of a central homeostatic mechanism, which at the highest level manifests as adaptive control of behavior in the environment (Critchley 2005). Craig (2002, 2009) notes that mental representation of this control requires an integra- tion of salience of all relevant conditions at each moment in time. This operationally expands the definition of homeostasis to include forces on an organism in its affective, cognitive and social networks. It is proposed that the insular cortex underlies this mechanism and makes these rep- resentations available to the consciousness of the individual. More specifically, the homeostatic afferent pathway from lamina I and the solitary nucleus terminates in the PI, establishing the sense of physiological condition. This pathway and its terminus is activated by graded cooling, sensual (limbic) touch, thermal pain, chronic pain, amongst other interoceptive signals and complements the efferent autonomic pathway that effects the physiological state of the body (Craig 2002; Olaus- son 2002).

In Craig’s model, this physiological state is then re-represented in a hierarchically from pos- terior to anterior insula, each level incorporating information from a different source. Initially, emotionally salient environmental stimuli arrive from other higher-order sensory areas. This could potentially be in line with the functional connectivity analyses considering that the ”mid-insula” is attributed with gustatory and olfactory representations (Kurth et al. 2010). Next, information about hedonic cues from the limbic system, i.e. nucleus accumbens, is integrated more anteri- orly. Most anteriorly, motivational, social and cognitive conditions are integrated with information directed from the anterior cingulate cortes, the ventromedial prefrontal cortex (VMPFC) and the dorsolateral prefrontal cortex (DLPFC). The later stages do not correspond well with results from the connectivity studies. Yet, authors suggest that strict divisions based on co-activation may not be the best way to define mechanisms within the insula (Uddin et al., 2016), given its demonstrated involvement in many tasks (Chang et al., 2012).

The cumulative representation available at the most anterior portion of the insula is proposed to generate the image of the sentient self at a given point in time, which is defined by interoceptive, reward-related, affective and cognitive input (Craig 2002). This claim is supported by studies demonstrating the instantiation of the ”feeling of knowing” within the AI, particularly the right AI. Enhanced AI activity corresponds with high physiological arousal and with awareness of changes in one’s physiological state (Critchley 2005). Further, two studies reported that anosognosia, an inability to recognize one’s own limitations, in hemiplegia is selectively associated with lesions in the right AIC (Spinazzola et al, 2008; Karnath 2005).

The former studies give evidence for the interoceptive awareness that the insula instantiates and the following provides evidence for the emotional awareness it instantiates. Seeley et al. (2007) showed that loss and dysmorphia of von Economo neurons in the ACC and insula typified cases of frontotemporal dementia, a neurodegenerative condition that has deleterious effects on emotional self-awareness, empathy and ”theory of mind”.

Individuals with alexithymia are characterized by inability to identify and describe emotions in the self as well as difficulty distinguishing and appreciating the emotions of others, leading to unempathetic and ineffective social responding. Early research attributed altered interoceptive awareness with the disorder, suggesting that the inability to distinguish physiological states gave rise to the inability to conceptualize emotional states. A 2014 study elucidated this mechanism. Researchers found that increased glutamate-mediated excitatory transmission - related to enhanced insula activity - reflects increased interoceptive awareness in alexithymia (Ernst et al., 2014). This results in increased awareness of bodily and stress responses and thus high, unspecific arousal. This demonstrates the contribution of physiological state in the generation of an emotion, parallel with Damasio’s somatic-marker hypothesis .

Risk Prediction within the Anterior Insular Cortex

Feelings and awareness are difficult to formalize. Recent findings, though, define a quantifiable function of the insula: tracking of risk with respect to aversive outcomes (Preuschoff et al., 2008). Bilateral insula activation corresponds to two portions of risk processing. Risk prediction is risk associated with an uncertain outcome measured by reward variance. If the prediction does not match the actual outcome, risk prediction error occurs and that error is used to improve future estimates of risk. The activation patterns of a more dorsal portion of the insula followed the inverted-U function of reward probability. A more ventral portion of the insula increased linearly as a function of risk prediction error (Preuschoff et al. 2008). Relationships here were indexed to objective measures of risk. Subjective risk, influenced by risk aversive, pessimistic and other traits, is less well understood, considering that risk and risk prediction errors have been associated with activation in the ACC and inferior frontal gyrus (Bossaerts 2010). These structures, along with the amygdala are thought to form a network of emotional salience that guides paralimbic behavior (Seeley et al. 2007).

These findings do not contest the view of the AIC as the seat of subjective awareness, but may aid in further study of the insula by being able to index sensitivities to quantitative features of the environment. If the insula plays an integrative role in linking affective value/feeling with cognitive input, the rating of risk or uncertainty of potential outcomes would be important for homeostatic function. First, the risk prediction could prepare an individual for eventual outcomes. For instance, uncertainty about the safety of an environment requires the body to prepare to take the right action if a source of danger materializes (Bossaerts 2010). Second, The risk prediction error plays an important role in rapidly updating risk predictions, especially in an uncertain environment (Critchley 2005). Last, the anticipatory risk valuation may contribute to decisions about what to do (or not to do) next in order to achieve reward (Kuhnen and Knutson 2005; Seeley et al. 2007).

The Singer model of insula function eloquently summarizes these findings. The insula is a region ”in which sensory, affective and bodily information is integrated with information about uncertainty (or risk) to generate a dominant subjective feeling state” that includes current and predictive feeling states. These patterns of activation may give rise to emotional confidence that serves to motivate, guide social interaction and influence decision-making behaviors (Singer 2009). Use of current and predictive feeling states not only follows other findings for bayesian processing but also has explanatory value for ”gut feelings” of unease or anxiety.

Conclusion

The central location of the insula affords the region with the ability to process sensory, affective and contextual information in order to produce feeling states. Craig’s model of subjective awareness as a integrated, re-representation of interoceptive cues in the AI has been received as positing the insular cortices as necessary and sufficient platform for human feelings. Let it be noted that Craig does not state that these cortices are the sole neural source of subjective awareness. Damasio, in a case study, warns against ignoring the other brainstem structures, somatosensory cortices, basal forebrain nuclei and basal ganglia that are involved in affective processing (Damasio et al., 2013). The bilateral insular cortices of patient in the case study were destroyed as a result of Herpes simplex encephalitis. All aspects of feeling were intact in this individual. Instead of modeling the insular cortex as the sole neural substrate of awareness, Damasio proposes that feeling states are encoded first subcortically and then repeated at the cortical level (Damasio et al., 2013). Even if the insula is not the sole provider of sentience, it has confirmed to play a major role. Further studies, particularly using quantifiable paradigms, such as risk prediction, will clarify how humans have conscious access to subjective and emotional awareness.

Citations

Bossaerts, P. (2010). Risk and risk prediction error signals in anterior insula. Brain Struct. Funct., pages 1–9.

(Bud) Craig, A. D. (2009). How do you feel — now? The anterior insula and human awareness. Nat. Rev. Neurosci., 10(1):59–70.

Cauda, F., D’Agata, F., Sacco, K., Duca, S., Geminiani, G., and Vercelli, A. (2011). Functional connectivity of the insula in the resting brain. Neuroimage, 55(1):8–23.

Chang, L. J., Yarkoni, T., Khaw, M. W., and Sanfey, A. G. (2013). Decoding the role of the insula in human cognition: Functional parcellation and large-scale reverse inference. Cereb. Cortex, 23(3):739–749.

Critchley, H. D. (2005). Neural mechanisms of autonomic, affective, and cognitive integration. J. Comp. Neurol., 493(1):154–166.

Damasio, A., Damasio, H., and Tranel, D. (2013). Persistence of feelings and sentience after bilateral damage of the insula. Cereb. Cortex, 23(4):833–846.

Damasio, A. R. (1996). The somatic marker hypothesis and the possible functions of the pre- frontal cortex. Philos. Trans. Biol. Sci.

Deen, B., Pitskel, N. B., and Pelphrey, K. A. (2011). Three systems of insular functional connec- tivity identified with cluster analysis. Cereb. Cortex, 21(7):1498–1506.

Dosenbach, N. U. F., Fair, D. a., Miezin, F. M., Cohen, A. L., Wenger, K. K., Dosenbach, R. a. T., Fox, M. D., Snyder, A. Z., Vincent, J. L., Raichle, M. E., Schlaggar, B. L., and Petersen, S. E. (2007). Distinct brain networks for adaptive and stable task control in humans. Proc. Natl. Acad. Sci. U. S. A., 104(26):11073–8.

Ernst, J., Boker, H., Hattenschwiler, J., Schupbach, D., Northoff, G., Seifritz, E., and Grimm, S. (2013). The association of interoceptive awareness and alexithymia with neurotransmitter concentrations in insula and anterior cingulate. Soc. Cogn. Affect. Neurosci., 9(6):857–863.

Gu, X., Hof, P. R., Friston, K. J., and Fan, J. (2013). Anterior insular cortex and emotional awareness. J. Comp. Neurol., 521(15):3371–3388.

Karnath, H.-O. (2005). Awareness of the Functioning of One’s Own Limbs Mediated by the Insular Cortex? J. Neurosci., 25(31):7134–7138.

Kurth, F., Zilles, K., Fox, P. T., Laird, A. R., and Eickhoff, S. B. (2010). A link between the systems: functional differentiation and integration within the human insula revealed by meta- analysis. Brain Struct. Funct., (February 2017):1–16.

Mutschler, I., Wieckhorst, B., Kowalevski, S., Derix, J., Wentlandt, J., Schulze-Bonhage, A., and Ball, T. (2009). Functional organization of the human anterior insular cortex. Neurosci. Lett., 457(2):66–70.

Nomi, J. S., Farrant, K., Damaraju, E., Rachakonda, S., Calhoun, V. D., and Uddin, L. Q. (2016). Dynamic functional network connectivity reveals unique and overlapping profiles of insula sub- divisions. Hum. Brain Mapp., 37(5):1770–1787.

Olausson, H., Lamarre, Y., Backlund, H., Morin, C., Wallin, B. G., Starck, G., Ekholm, S., Strigo, I.,Worsley,K.,Vallbo,a.B.,Bushnell,M.C.,Vallbo,A ̊.,andBushnell,M.C.(2002).Unmyeli- nated tactile afferents signal touch and project to insular cortex. Nat. Neurosci., 5(9):900–904.

Preuschoff, K., Quartz, S. R., and Bossaerts, P. (2008). Human Insula Activation Reflects Risk Prediction Errors As Well As Risk. J. Neurosci., 28(11):2745–2752.

Seeley, W. W., Carlin, D. A., Allman, J. M., Macedo, M. N., Bush, C., Miller, B. L., and DeAr- mond, S. J. (2006). Early frontotemporal dementia targets neurons unique to apes and humans. Ann. Neurol., 60(6):660–667.

Singer, T., Critchley, H. D., and Preuschoff, K. (2009). A common role of insula in feelings, empathy and uncertainty. Trends Cogn. Sci., 13(8):334–340.

Spinazzola, L., Pia, L., Folegatti, A., Marchetti, C., and Berti, A. (2008). Modular structure of awareness for sensorimotor disorders: Evidence from anosognosia for hemiplegia and anosog- nosia for hemianaesthesia. Neuropsychologia, 46(3):915–926.

Uddin, L. Q., Kinnison, J., Pessoa, L., and Anderson, M. L. (2014). Beyond the Tripar- tite Cognition-Emotion-Interoception Model of the Human Insular Cortex. J. Cogn. Neurosci., 26(3):194–198.

Wicker, B., Keysers, C., Plailly, J., Royet, J. P., Gallese, V., and Rizzolatti, G. (2003). Both of us disgusted in My insula: The common neural basis of seeing and feeling disgust. Neuron, 40(3):655–664.