ONE Function for the Anterior Cingulate Cortex and General AI: Consistency Curation

Abstract

Control is essential in all complex systems serving any purpose, and cognitive control is claimed to fill a pivotal role for the cognition of natural agents. The Anterior Cingulate Cortex (ACC) is a neural structure, widely believed to be involved in cognitive control. In the human brain, the ACC undoubtedly takes a center position and features several peculiarities. It has been implicated in a wide range of tasks ranging from sensing pain and autonomic regulation to high-level executive functions like error monitoring and abstract control. A sketch how this seemingly irreconcilable plurality can be understood as manifestations of only one general process working in diverse contexts on different content is presented and also how this constitutes evidence for a specific cognitive architecture. Going beyond extant proposals for ACC function, the here presented one recommends an embedding into a very general overarching model of cognition. Discrepancy monitoring highlights deviations between predictions derived from memory content and current activations. A process doing this has recently been proposed as central to the 'Ouroboros Model' under the title of 'consumption analysis'. It is claimed that ONE functional account centered on consistency checking and, based thereupon, the overall minimization of discrepancies ("consistency curation and cultivation”) can parsimoniously explain the plentitude of reported observations for the ACC, thus shedding new light on the neural computations, which form the foundations of cognition and, finally, consciousness. A specific "natural" reconciliation of various current conceptualizations and theoretical models is offered, which points out promising directions for future modelling efforts and attempts to harness the findings generally for artificial intelligence.