Biological neural systems are typically described as networks of neurons that transmit and process signals across distributed circuits. Within this framework, information flow is assumed to be sufficient to account for the emergence of functional brain states. However, signal propagation alone does not explain why only a small subset of dynamically accessible neural activities becomes stabilized as coherent, executable states, nor why neural dynamics remain robust despite the combinatorial explosion of physically possible activity patterns. This indicates an ontological gap that signal processing alone cannot bridge.
Here we propose a complementary framework in which brain function is understood as the interaction between a signal layer, which generates a high dimensional space of potential neural configurations, and an admissibility layer, which enforces the structural constraints that determine which of these configurations can be realized. We introduce the concept of a Global Admissibility Field (GAF), a spatially extended constraint geometry that specifies the realizability conditions for distributed neural activity. Neural activity is thus represented as a space of potentials constrained by an admissibility manifold, whose intersection defines the set of realizable brain states.
Recent experimental observations reveal that astrocytes form long range, plastic connectivity patterns through gap junctions, dynamically reorganizing in response to environmental and sensory perturbations. These networks do not transmit information in the neuronal sense; instead, they operate as a global constraint system that modulates the conditions under which neural activity becomes functionally realizable. Within this framework, neuronal activity represents potential configurations, while astrocytic networks define the admissibility boundaries that determine which configurations can persist.
We further propose that astrocytic connectivity implements a form of geometric exclusion, suppressing incompatible patterns of activity and thereby increasing the resolution with which the system distinguishes admissible from inadmissible states. Learning and plasticity are reinterpreted not as the creation of new pathways alone, but as the refinement of admissibility geometry, corresponding to a contraction of the admissible manifold and an increase in the exclusion resolution parameter ρ.
Beyond neural systems, decisive evidence for admissibility driven execution arises from recent advances in three dimensional organoid cultivation. In two dimensional culture—an environment where admissibility boundaries collapse into low dimensional degeneracy—iPS derived immune cells stall in immature states despite adequate signaling cues. In contrast, the introduction of a three dimensional artificial thymic organoid (ATO) restores the admissibility geometry required for maturation, enabling the realization of functional CD4⁺ phenotypes. This demonstrates that biological identity is not determined by signal content, but by the three dimensional exclusion geometry that eliminates incompatible developmental trajectories.
Recent molecular scale observations further extend this principle. In plant regeneration, wounding induces the rapid nuclear translocation of the heat stress transcription factor HSFA1, which directly activates reprogramming factors such as WIND1, PLT3, and ZAT6. Within the framework of admissibility geometry, HSFA1 functions as a molecular boundary defining factor that instantiates a local Admissibility Generation Entity (AGE), reconstructing the admissibility manifold and enabling cellular reprogramming. As in the organoid case, the regenerative state emerges not through the addition of new information, but through the elimination of incompatible configurations under a sharpened exclusion geometry.
In parallel, artificial intelligence systems provide a complementary failure mode. Contemporary architectures operate predominantly within a signal generation layer, lacking a geometric exclusion mechanism. As a result, they exhibit characteristic phenomena such as hallucination and incoherence, which can be interpreted as ontological leakage arising from the absence of an admissibility field.
Taken together, these results demonstrate that admissibility driven execution is independently realized across multiple, structurally distinct domains: astrocytic constraint networks in the brain, geometric constraint environments in organoid systems, molecular boundary defining mechanisms in plant regeneration, and their absence in artificial intelligence systems. Despite fundamental differences in substrate, these systems converge on a single structural principle: realization is governed by constraint geometry rather than by signal accumulation.
This framework provides a unified interpretation of neural stabilization, plasticity, representation, and large scale coordination, and establishes a general, strictly scale invariant and domain independent principle of admissibility driven execution governing the realization of functional states across biological and computational architectures.
