Embodiment & Emergence — Weekly Roundup #4
April 17, 2026 · threat modeling · prediction error · adaptive forgetting · constraint perception
Scope of this issue
The pieces in this issue come from different areas, including sensory neuroscience, interoception research, computational models of trauma, and a cross-disciplinary discussion on memory and learning. They are not unified and they do not use the same language, but they are describing overlapping aspects of the same problem. Read together, the connections are fairly direct.
Signal integration and bodily experience
One thread runs through the work on somatosensation and interoception. The body is not providing a single, stable signal. It is multiple streams, including external sensation, position, and internal state, being transformed and combined into something that feels like a coherent sense of self. That integration is partial and ongoing. It depends on how signals are weighted, how they are interpreted, and how they are brought into relation with each other. The respiratory interoception work makes that more explicit. What someone feels in their breathing is not just detection. It includes how confident they are in that signal, what they think is causing it, and whether it is interpreted as threat or not. Those layers can diverge. In practice, that can mean two people in the same situation experience something very differently.
Memory, threat, and model stability
A second thread shows up more directly in the trauma and memory material. The “body keeps the score” framing is challenged here, not by rejecting embodiment, but by shifting where persistence is located. In the predictive coding account, what holds is not stored experience in the body, but a set of high-weighted inferences about threat that continue to organize perception and response. The discussion on forgetting moves in a similar direction from a different angle. Memory is not accumulation. It is selective reduction. Systems maintain workable models by discarding detail. When that process is disrupted, either through too much retention or too little updating, behavior becomes rigid in different ways.
Taken together, these pieces orient less to stored states and more to ongoing organization. How signals are combined, how they are evaluated, and how models are maintained or revised. That applies at the level of an individual, but it also shapes how people interact, coordinate, and respond to changing conditions in groups.
Subscribe for a weekly synthesis of research on mind, body, and collective behavior.
Somatosensation and the sense of self in bodily experience
Haggard, P., & Longo, M. (2025). Somatosensation and the sense of self. Current Biology* Advance online publication. https://www.cell.com/current-biology/fulltext/S0960-9822%2825%2900999-6
TAGS: somatosensation; exteroception; proprioception; interoception; body schema; tactile remapping; somatotopy; bodily self; posterior parietal cortex; embodiment
OVERVIEW: This primer reviews how somatic sensory signals from exteroception, proprioception, and interoception are organised, integrated, and transformed to support a coherent experience of one’s body and a “minimal” bodily self. Drawing on classic neurophysiology and contemporary human and animal studies, the authors describe peripheral receptor classes, ascending pathways, somatotopic and interoceptive maps, and multi-level “synthesis” processes within each modality. They then examine cross-modal integrations—especially between touch and proprioception—that underlie tactile localisation, body geometry, and spatial remapping.
OF NOTE: Conceptually, the article reframes somatosensation as a set of interacting representational systems that progressively abstract away from raw receptor input toward object- and body-level descriptions. Exteroceptive pathways transform skin events into external objects, proprioceptive systems into a volumetric, configurational “body schema,” and interoceptive systems into functional states such as satiety or well-being. Integration across these systems is proposed to contribute to a minimal sense of self, though the authors emphasise that interoceptive contributions and a putative “interoceptive object” remain unclear. The account is consistent with an embodied, distributed view of self-consciousness emerging from coupled peripheral–central processes rather than a single locus.
CAVEATS: The article is a narrative, conceptual synthesis rather than a primary experimental study, so it does not report new data or systematic meta-analytic estimates. Many claims about integration mechanisms and their relation to self-consciousness are extrapolated from specific paradigms (e.g., proprioceptive illusions, tactile distance judgments, crossed-finger illusions) and may not generalise across populations or contexts; quantitative effect sizes and variability are not provided. Higher-level neural circuitry for proprioceptive and interoceptive synthesis is described as largely unknown, and the putative links between interoception and mental health are flagged as important but not yet mechanistically specified.
KEY TAKEAWAYS: Somatic self-experience arises from multiple partially distinct, hierarchically organised sensory systems that each perform internal “synthesis” from receptors to structured representations. Exteroceptive and proprioceptive pathways provide relatively well-characterised transformations from local signals to coherent perceptions of external objects and body configuration, including remapping touch into external space. Interoceptive pathways clearly inform regulatory behaviour and affective tone but are less well understood in terms of object-like representation. Overall, the evidence supports viewing the bodily self as an emergent construct built from ongoing integration of distributed sensory maps, rather than a simple readout of any single sensory class or cortical region. For practitioners working with disturbances of body experience, this suggests attending to how different bodily information streams are weighted and integrated in real-world contexts rather than focusing on one modality in isolation.
Breath, Belief, and Brain in Respiratory Interoception: A Journal Club Perspective
Radziun, D., & Larsson, D. E. O. (2026). Breath, belief, and brain: A tripartite map of respiratory interoception. Journal of Neuroscience, 46(13), e1891252026. https://doi.org/10.1523/JNEUROSCI.1891-25.2026
TAGS: Interoception; Respiration; Insula; Metacognition; Affective neuroscience; Psychophysics; Anxiety; Predictive processing
OVERVIEW: This Journal Club article discusses recent work on respiratory interoception, defined as the perception of signals from one’s own breathing. It presents a conceptual taxonomy, distinguishing interoceptive accuracy (e.g., detecting when breathing is harder), interoceptive insight (calibration of confidence about those sensations), attribution (identifying the cause of shortness of breath), and appraisal (judging the sensation as threatening or harmless). The authors situate respiratory interoception within distributed neuroanatomical circuits involving brainstem, thalamic, limbic, and cortical regions, with a particular emphasis on posterior and anterior insula. They highlight respiration as an experimentally tractable and affectively salient modality, where inspiratory resistance can be manipulated in finely graded steps, enabling rigorous psychophysical characterization of detection and evaluation of breathing-related sensations.
OF NOTE: Conceptually, the piece argues that interoception is better understood as a “family of processes” rather than a single capacity, and that higher-order components—insight, attribution, appraisal—are crucial for explaining why similar bodily states feel different across individuals. The discussion underscores how variation in estimates of the precision or reliability of bodily signals can be clinically relevant, particularly in conditions such as anxiety, depression, eating disorders, and substance-use disorders, where atypical weighting of interoceptive evidence is linked to reduced accuracy and metacognitive insight, and altered attribution and appraisal. From an embodiment and systems perspective, this framing is consistent with the idea that felt experience of breathing emerges from interactions among peripheral receptors, hierarchical brain networks, and belief-like evaluative processes, rather than from any single “interoceptive center.”
CAVEATS: The article itself is a conceptual and integrative commentary rather than a primary empirical study; it does not present new behavioral or neuroimaging data, nor specific sample sizes or quantitative effect estimates. Descriptions of neural circuitry and clinical associations are based on prior literature and consensus accounts, not newly collected evidence. As such, the strength of claims is limited by the methodologies and heterogeneity of the underlying work, which are not detailed here. The generalization of findings across interoceptive modalities (e.g., from cardiac to respiratory) is explicitly noted as unresolved.
KEY TAKEAWAYS: Respiratory interoception can be parsed into at least four distinguishable processes—accuracy, insight, attribution, and appraisal—embedded in distributed neural circuits that include, but are not limited to, the insular cortex. Respiration offers a powerful testbed for interoceptive research because inspiratory resistance can be precisely controlled while carrying inherent affective meaning. Variations in how reliably bodily signals are weighted, and how they are interpreted and evaluated, are plausibly important in several psychiatric conditions, although this piece does not add new causal evidence. For practitioners working with anxiety or breath-focused interventions, the framework suggests that what people “feel” in their breathing reflects interactions among sensory detection, confidence calibration, and learned interpretive schemas, not just raw physiological load.
The Body Does Not Keep the Score: Trauma, Predictive Coding, and the Restoration of Metastability
Kotler, S., Mannino, M., Fox, G., & Friston, K. (2026). The Body Does Not Keep the Score: Trauma, Predictive Coding, and the Restoration of Metastability. Frontiers in Systems Neuroscience. https://www.frontiersin.org/journals/systems-neuroscience/articles/10.3389/fnsys.2026.1812957
TAGS: Trauma; PTSD; Predictive processing; Active inference; Metastability; Free Energy Principle; Somatic Marker Hypothesis; Flow states; Cognitive control; Embodiment
OVERVIEW: This theoretical article synthesizes computational, systems, and cognitive neuroscience accounts of trauma, arguing that post-traumatic symptoms reflect disordered prediction rather than somatic storage. Drawing on the Free Energy Principle and active inference, the authors propose that trauma induces overweighted “danger” priors and a collapse of metastability in large-scale brain networks, particularly hypersynchronous loops among amygdala, hippocampus, and medial prefrontal cortex. They integrate Damasio’s Somatic Marker Hypothesis and convergence–divergence zones to explain how emotional memories are neurally, not peripherally, instantiated, and review evidence that cognitive control capacity and network flexibility modulate intrusive symptoms and recovery. Flow states are advanced as a candidate process for restoring metastability via transient hypofrontality, neuromodulatory shifts, and dynamic network reconfiguration.
OF NOTE: Conceptually, the paper reframes “the body keeps the score” as a misleading storage metaphor and proposes trauma as “frozen inference” in a maladaptive free-energy landscape. Embodiment is preserved but recast: the body and environment are seen as active components of predictive loops rather than archives of past insult. This aligns with an interactional view in which PTSD emerges from miscalibrated precision across interoceptive, exteroceptive, and executive systems, rather than a single locus of pathology. The suggestion that diverse interventions—exposure, mindfulness, somatic work, psychedelics, and flow-inducing activities—may converge on restoring metastability is consistent with a systems-level account of symptom change, in which network flexibility, rather than specific narrative content, is the primary therapeutic lever.
CAVEATS: The article is conceptual and integrative; it does not report new empirical data, experimental designs, or quantitative effect sizes. Neuroimaging and behavioral findings are summarized at a high level, with limited methodological detail, sample characteristics, or statistical parameters, restricting evaluation of evidential strength. The proposed centrality of metastability, precision weighting, and flow states as mechanisms of change remains hypothetical within this paper and is not directly tested. Generalization across trauma types, developmental stages, and comorbid conditions is discussed implicitly rather than specified, and potential boundary conditions for the framework are not systematically delineated.
KEY TAKEAWAYS: A careful reader can reasonably conclude that the authors view trauma as a disorder of predictive processing—characterized by rigid, high-precision threat priors and reduced network metastability—rather than as trauma being “stored” in peripheral tissues. Embodiment is construed as dynamic coupling between brain, body, and environment, where mis-weighted interoceptive and exteroceptive signals create the appearance that “the body keeps score.” Cognitive control capacity and large-scale network flexibility are presented as plausible, though not empirically confirmed within this paper, levers for symptom reduction. The framework suggests, without claiming to prove, that interventions which safely increase variability, recalibrate precision, and promote flexible coupling—potentially including flow-inducing activities—may support recovery by restoring metastable brain dynamics. In practice-oriented settings, this can be interpreted as a rationale for focusing on processes that shift systems from defensive rigidity toward adaptive, action-oriented engagement, while recognizing that this remains a theoretically driven, not experimentally adjudicated, account.
Discussion of Forgetting, Agency, and Generative Models in Mind and Development
Levin, M. (Host), Watson, R., Tolchinsky, A., Solms, M., & Friston, K. (2026, April 9). Discussion: Richard Watson, Alexey Tolchinsky, Mark Solms, Michael Levin, and Karl Friston [Video]. YouTube. https://youtu.be/w_ciA-yyF8M
TAGS: forgetting, memory consolidation, dreaming and REM sleep, agency and free will, psychoanalysis, developmental models, autism and overfitting, collective intelligence, regeneration
OVERVIEW: This ~1-hour roundtable brings together a set of influential thinkers whose work collectively shapes much of the current landscape around mind, brain, and complex systems. Michael Levin (systems biologist, known for work on bioelectricity, regeneration, and collective intelligence) convenes Karl Friston (architect of the Free Energy Principle and active inference), Mark Solms (founder of neuropsychoanalysis and a leading figure in the science of dreaming), along with Richard Watson and Alexey Tolchinsky (working at the intersection of computational modeling, development, and learning systems).
The discussion sits at the overlap of several frameworks: predictive processing and active inference (Friston), neuropsychoanalysis and affective neuroscience (Solms), and multi-scale biological and collective intelligence models (Levin), with Watson and Tolchinsky extending these ideas into development and machine learning. What emerges is not a single unified theory, but a set of partially overlapping approaches that differ in emphasis—on brain-based inference, embodied affect, or distributed biological intelligence—while converging on the idea that cognition is an adaptive, model-based process operating across levels.
Within this shared space, the group examines forgetting as a core function of learning and agency. Rather than treating memory as storage, they frame consolidation as selective pruning: systems reduce complexity to maintain usable, generalizable models. Sleep—particularly REM and slow-wave phases—is described as part of this process, supporting different modes of updating or stabilizing internal models. Developmental trajectories reflect this balance, with early systems characterized by high plasticity and rapid forgetting, and later systems becoming more stable but also more prone to rigidity. Edge cases such as autism or eidetic memory are used to illustrate how reduced forgetting can impair abstraction and limit temporal continuity.
OF NOTE: A central point of alignment across the discussion is that learning and forgetting are inseparable. To consolidate is to forget selectively, shaping models by removing rather than accumulating detail. From this perspective, agency is not simply the capacity to choose, but the ability of a system to constrain possible pasts and futures into a coherent trajectory.
At the same time, the discussion highlights differences in framing. Predictive processing emphasizes precision weighting and model optimization; neuropsychoanalysis foregrounds affect, drives, and the persistence of early models; and biological systems approaches extend these ideas beyond the brain to tissues, organisms, and collectives. These perspectives converge in viewing cognition as embodied and distributed, but diverge in where they locate primary constraints—neural inference, affective organization, or multi-level system dynamics.
Dreaming is positioned as one mechanism through which systems manage this balance. Rather than storing experience, dreams are described as participating in the reorganization of models, potentially reducing overfitting and stabilizing core structures. This connects to broader claims about rigidity in psychopathology, where systems fail not simply by remembering too much, but by being unable to revise or discard outdated patterns.
CAVEATS: This is a conceptual, cross-disciplinary dialogue rather than an empirical study. Claims regarding synaptic pruning, dream function, developmental memory, and psychopathology are presented as plausible frameworks without new data, sample sizes, or formal analyses. Clinical and neuropsychological examples are illustrative and often based on classic cases rather than systematic review. Analogies to machine learning, free energy models, and regenerative biology are heuristic and not supported here by formal modeling results.
KEY TAKEAWAYS: Across these perspectives, cognition is framed as a process of ongoing model management rather than information storage. Selective forgetting is necessary to maintain models that remain flexible enough to generalize while stable enough to guide behavior. Excessive retention leads to overfitting and loss of abstraction, while insufficient updating leads to rigidity and persistence of outdated patterns. Agency emerges from how a system constrains and revises its possible histories and futures over time. For practitioners, this suggests focusing not only on what is learned or remembered, but on how systems regulate complexity—what is maintained, what is discarded, and how that balance shifts under different conditions.
Implications for behavior and coordination
What these pieces point to is a different way of understanding where behavior is coming from. Instead of treating it as something carried forward and expressed, they describe it as something being produced out of how the system is currently taking in and organizing information. In the somatosensory work, that shows up in how different bodily signals are combined into a usable sense of self. In interoception, it shows up in the gap between what is detected, how confident someone is in that detection, what they think is causing it, and whether it is treated as threat. In the trauma model, it shows up as threat patterns that stay fixed even when conditions change. In the discussion on forgetting, it shows up as models that are not being updated or simplified in a way that keeps them usable.
This is where embodiment matters in a practical sense. When parts of “what is happening” are restricted, ignored, or not taken in … they cannot factor into how the situation is understood. The system builds a working model from partial information. That model might feel coherent, but it does not reflect the full set of conditions. As a result, the range of possible responses is narrower than it might need to be. Emergence then depends on whether that model can change. If new information is not incorporated, or if existing patterns are held too tightly, the same responses repeat even when they stop working.
You see the same thing in groups. Coordination depends on what people are noticing and how quickly they can adjust to changes in each other and the environment. When that breaks down, groups do not just slow down. They fall back on familiar patterns, even if those patterns no longer fit what is happening. Most attempts to intervene focus on the response: change the behavior, improve communication, increase effort. The material here points somewhere else. It points to what is being registered, how it is being interpreted, and whether the system can revise the model it is operating from. That is where change tends to either happen or stagnate.
#research | #theory | #practice
About emotive.energy
We study how people think, feel, and respond in real time, and how those patterns interact in groups and systems. As a research-led co-op and working laboratory, we help individuals, teams, and organizations recognize and shift the dynamics that shape experience, including stress, decision-making, relationships, creativity, and meaning. Our work spans contexts from workplaces and leadership to family systems, artistic practice, and personal development. Through coaching, group programs, and proprietary tools, we make these dynamics visible as they form so people can work with them directly. The result is clearer thinking, more effective action, and more adaptive ways of relating under pressure and in everyday life. [website]
Links
Website: emotive.energy
Substack: Posts · Notes · Podcast
Social: Twitter / X · BlueSky · YouTube · Facebook · LinkedIn · Instagram
Podcast: Apple Podcasts · Spotify · Pocket Casts · Overcast