Operationalizing Full-Body Intelligence: A Multimodal Experimental Framework for Interoceptive, Autonomic, Neural, and Predictive Dynamics

Full-body intelligence (FBI) has been proposed as a falsifiable organism-level construct arising from dynamic interactions among interoceptive, autonomic, sensorimotor, neural, neuroendocrine, and environmental processes. The central empirical problem is now measurement: established physiological and cognitive measures do not become validated measures of full-body intelligence merely by being assembled into a battery. This paper specifies a prospective multimodal framework for testing whether candidate dimensions form reproducible, incrementally predictive patterns under controlled perturbation. The framework prioritizes four core candidate domains: multidimensional interoceptive performance and metacognitive calibration; autonomic and cardiorespiratory regulation; flexible interoceptive–exteroceptive precision weighting; and recovery and adaptive reorganization after perturbation. Heartbeat-evoked potentials and cortical–autonomic coupling are treated as candidate neural measures requiring stringent artifact control and reliability assessment. Anticipatory physiological activity, examined separately in Preprint 3, is retained only as an optional exploratory module and is excluded from the primary construct score unless independently confirmed. The proposed research program combines baseline phenotyping, standardized perturbations, repeated measurement, multimodal physiology, preregistration, multilevel modeling, latent-variable analysis, cross-validation, measurement-invariance testing, and independent multi-site replication. The principal validation question is not whether individual component measures show expected effects, but whether a higher-order full-body intelligence model demonstrates reliability, discriminant validity, incremental prediction, and replicability beyond its component processes. The framework is explicitly falsifiable. Failure to identify a reliable and replicable higher-order construct is treated as an informative outcome that would favor a multidimensional profile or component-process account over a unitary construct. This reflects the manuscript's central methodological distinction: established physiological measurements are not automatically validated measurements of full-body intelligence.