A Somatic Ontology for Computational Therapeutics: Formalizing Body-Based Experience for Digital Mental Health

Digital mental health platforms increasingly incorporate body-based therapeutic approaches, yet no standardized ontology exists for representing somatic experience computationally. This paper introduces the Cathexis Somatic Ontology (CSO), a formal vocabulary and relational schema for encoding body-based therapeutic data. The CSO defines entity classes for anatomical zones (29 zones with clinical significance rationale), sensation qualities (20 phenomenological descriptors), neurobiological organization patterns (16 territories derived from predictive processing, primary affect systems, and somatic marker theory), therapeutic change mechanisms (3 memory-system-based pathways), and intervention architectures (7 mechanism types). Entities are organized within a Markov blanket framework comprising four components (internal states, external states, sensory states, and active states), providing a principled computational structure derived from Friston's free energy principle. The ontology is grounded in seven major research programs: constructed emotion theory (Barrett), predictive processing (Friston), affective neuroscience (Panksepp), somatic markers (Damasio), neuropsychoanalysis (Solms), the entangled brain framework (Pessoa), and memory systems (Kandel, Nader, Schiller, Ecker). These seven programs are integrated within the Human Experience System (HES; DeGarbo, 2026c), a transdisciplinary model of embodied human experience that provides the overarching theoretical foundation for the ontology. The HES describes a person as a complex adaptive system of ten interacting layers experienced through the body, organized by prediction, developed through critical periods, and changed through embodied prediction error. It advances eight foundational claims — including a developmental account describing how the layers emerge sequentially and how the conditions of development determine the adult system's predictive architecture — and generates ten testable empirical predictions. Pessoa's work on combinatorial brain organization provides the architectural foundation for how the CSO represents distributed neural activation, resolving a gap in previous versions between theoretical commitments (no single region produces emotion) and visualization architecture (which must render this distributed reality). Cross-mappings to ICD-11, SNOMED CT, and HL7 FHIR are provided. Version 2.0 introduced a trifecta visualization architecture: three complementary rendering surfaces (BRAIN, BODY, WORLD) that correspond to the Markov blanket's internal, sensory, and external state components. Version 2.1 grounds the ontology in the Human Experience System (HES; DeGarbo, 2026c), reports the publication of machine-readable representations (CXSO v1.1) enabling programmatic consumption by clinical systems, research platforms, and AI agents, updates the ontology's HES references to reflect the complete model — including the developmental account, eight foundational claims, ten testable predictions, and psychoneuroimmunological integration — and identifies vocabulary and data model extensions for CXSO v2.0.