Independent research. Shared knowledge. Human insight.
CURIOSITY STARTS HERE

Find your next
research perspective.

Explore the Cortexplore collection. Connect with the evidence behind human behaviour.

42published articles
open to discovery
EXPLORE
THE CORTEXPLORE COLLECTION

Search results

2 articles for “Executive Functions”

Research Articles

The Offloading Window: A Conceptual Framework for AI-Assisted Cognition in Early Childhood: A Conceptual and Theoretical Contribution

Divyaansh Gupta · Komal Jain

Children now have daily use of artificial intelligence (AI) like voice response systems, conversational agents (e.g., chatbots), and Digital Learning Systems that provide an AI assist to support their learning. The use of such technology is expected to offer children personalized chances for educational enhancement and cognitive assistance when used appropriately. However, there is limited research on the effect of technology on the developmental/maturational process of children through cognitive offloading; thus, there is insufficient evidence of one or more elements leading to cognitive development through cognitive offloading by adults (i.e., complete adults) who have already developed their cognitive capacity. This paper integrates the current cognitive psychology, developmental neuroscience, and generative-AI-in-education literature to address the above-stated gap in knowledge. The proposed constructs are hierarchically related: The Developmental Offloading Hypothesis (DOH), a general principle confirming that cognitive offloading results in different kinds of effects when components of competence form as opposed to when they are being applied; The Offloading Window (OW), the contextual condition that defines whether the DOH is developmentally active for any particular child at any point in time based on their individual capacity and pattern of use; and the Scaffolded Cognitive AI Integration Framework (SCAIF), the operational framework for translating the principles of DOH and OW into practical guidelines that differentiate scaffolds from substitutions. Each construct is defined behaviourally, evaluated against alternative theoretical accounts (including the Extended Mind, Distributed Cognition, Vygotskian scaffolding, and Cognitive Load Theory), and expressed as a set of falsifiable propositions. The paper concludes with a roadmap for the empirical work required to test these propositions, including a planned mixed-methods study of parents and early-childhood educators.

White Papers

Cognitive Mirrors: The RCS Triad for Retrieval,Control, and Stewardship in Neuro-Inspired LLMSystems

Lalith Kumar Vemali · Rakesh Gandla

As large language model systems evolve from standalone text generators into retrieval-enabled, tool-using, and multi-session agents, researchers increasingly turn to cognitiveneuroscience for interpretive guidance. This paper argues that such comparisons areuseful only when they are bounded, functionally specified, and empirically falsifiable. Weanalyze retrieval-augmented generation as a separation between parametric knowledgeand external evidence access, and we analyze agent runtimes as control layers responsiblefor planning, routing, tool use, and context management. We distinguish these runtimefunctions from Model Context Protocol, which is an interoperability standard for exposingresources, prompts, and tools rather than a theory of memory or executive control. On thisbasis, we propose the RCS Triad Retrieval, Control, and Stewardship as a compactframework for describing modern LLM systems. Retrieval concerns evidence access andmemory extension; Control concerns orchestration, tool selection, and active context;Stewardship concerns provenance, consent, auditability, risk boundaries, and humanoversight. The paper’s central claim is methodological rather than biological: neuroscienceshould not be used to assert one-to-one equivalence between brains and models, but togenerate disciplined architectural hypotheses. We position the RCS Triad against eightadjacent retrieval and agent frameworks in a comparative table, specify its boundaryconditions through a Convergence-Divergence Matrix across eight cognitive domains,and outline a benchmarkable empirical agenda with four falsifiable predictions. We alsointroduce two original governance concepts architectural negligence and goaltransparency as frameworks for understanding the liability and rights questions thatarise when AI systems confabulate and persist goal-states.