Neurobiological Foundations of Cognition and Progress Toward Artificial General Intelligence—Volume II

The last several decades have produced transformative advances at the core of systems neuroscience and machine intelligence: (i) identifying how distributed neuronal circuits give rise to intelligent behavior, and (ii) designing machine systems that can reproduce selected cognitive functions. Progress in both domains traces back to foundational ideas formulated in the last century: neuronal assemblies as functional units of brain computation, the Turing machine as a general model of computation, and the transistor as the basic hardware element. Since then, systems neuroscience has been propelled by technologies that provide unprecedented access to brain dynamics across scales (e.g., MRI and related approaches) and by theories that aim to capture organizing principles of brain function (e.g., Active Inference, Neuronal Group Selection, Integrated Information, among others). In parallel, advances in machine intelligence have been driven by algorithms that emulate specific cognitive functions—particularly neural-network methods for learning—and by increasingly efficient hardware for implementing them.Recent breakthroughs in language and multimodal processing—enabled by large-scale neural network models—have sharpened long-standing questions central to systems neuroscience: What are the circuit-level and network-level mechanisms that support cognition and understanding? Can engineered systems meaningfully benefit from neurobiological principles of computation, control, and learning? Conversely, can modern AI successes help generate testable hypotheses about brain function, or do they remain largely orthogonal to the mechanisms underlying natural cognition and language? These debates include strong claims of divergence—e.g., that key aspects of understanding resist computational simulation (Penrose), or that the successes of large language models shed little light on the biological foundations of language (Chomsky). This topic revisits these questions with the explicit goal of moving from broad comparisons to systems-level, mechanistic convergence.A central motivation comes from a systems neuroscience perspective on constraints: human intelligence emerged under biophysical limits on brain size, wiring, and metabolic energy. This raises a timely question for 2026: can we identify the principles by which neural systems expand behavioral and cognitive repertoires without proportionally increasing energy use or neural resources, and translate these principles into new computational and experimental frameworks? On this view, today’s AI trend toward scaling compute and power may diverge from evolution’s strategy—highlighting the need for theories of efficient, robust, and adaptive intelligence grounded in biological systems.Accordingly, the potential benefits of a new synthesis are dual: (i) new insights into the systems-level neural mechanisms of intelligence, and (ii) biologically grounded approaches for building intelligent behavior on low-to-moderate capacity platforms, informed by how brains achieve efficiency, generalization, and resilience.We invite contributions that address the state of the art in mapping core intelligence functions—attention, awareness, perception, learning, understanding, prediction, explanation, consciousness—to neuronal and circuit processes, with a focus on:a) Assessing the relative contribution of these functions to intelligent performance at the network and systems level.b) Translating these functions into principles of system design, including hypotheses that can be tested experimentally and models that connect mechanism to behavior.This topic seeks to advance a systems neuroscience account of how the brain achieves expanding cognitive complexity under tight constraints—and what those constraints, mechanisms, and architectures imply for the next generation of machine intelligence and human–machine collaboration.This Research Topic is part of the Neurobiological Foundations of Cognition and Progress Toward Artificial General Intelligence series, first volume here. Topic editors: Robert Kozma, Yan Mark Yufik, James Kenneth Kroger Source: https://www.frontiersin.org/research-topics/86938/neurobiological-foundations-of-cognition-and-progress-toward-artificial-general-intelligence-volume-ii