The DIME Architecture: A New Unified Theory of the Brain // Detect – Integrate – Mark – Execute // Representations, Sequences, Value, and Consciousness in Neural Systems

In the absence of a unified cognitive architecture, modern neuroscience offers fragmented accounts of perception, memory, prediction, emotion, and consciousness. Contemporary theories—Predictive Coding, engram-based models, hierarchical cortical networks, recurrent–dynamical models, and sensorimotor architectures in robotics—successfully explain specific neural processes, yet they fail to provide a coherent structure capable of integrating these processes into a common mode of operation. This monograph introduces an entirely new architecture, termed DIME (Detect–Integrate–Mark–Execute), which unifies fundamental neuronal mechanisms within an operational model designed to describe how the brain perceives, stores, evaluates, and sequences information, thereby generating adaptive behavior and conscious experience. The DIME model is built around four core components: the engram (the functional unit of representation and processing), the execution thread (the internal flow of neuronal sequencing), neuromodulatory markers (a universal mechanism for value, attention, and selection), and the hyperengram (the narrative–global network supporting context, memory integration, and consciousness). The theory introduces a unified formalization of match/mismatch mechanisms, elucidates the role of local and global recurrence in the emergence of stable representations, and proposes a natural integration of hippocampal systems, higher-order cortical networks, and affective systems. DIME provides a framework in which perception, episodic and semantic memory, attention, emotion, prediction, and motor control are described as instances of the same underlying mechanism. Furthermore, the architecture enables an operational account of consciousness, defined here as an emergent process of the hyperengram, sustained by the synchronization of large-scale networks (DMN, Salience, Executive Control) and by marker-based competition. Beyond its biological grounding, DIME is explicitly designed to be fully implementable in computational terms. The monograph includes complete algorithmic pseudocode, simulation examples, systematic comparisons with Predictive Coding and contemporary artificial models, as well as applications in cognitive robotics and artificial intelligence. Through this dual orientation—biological and algorithmic—DIME positions itself as a general framework for the development of artificial cognitive systems, the integration of neuroscientific data, and the re-examination of human brain function.