Osi-Learn: A Unique Controlled Serious Game Platform for HCI-Driven Lexico-Semantic Evaluation in Early Childhood Education

This study bridges Human-Computer Interaction (HCI) and educational game design, introducing Osi-Learn, a novel serious game platform designed to measure lexical- semantic interaction in early childhood within controlled environments. The system integrates animated storytelling, mini-games, and automated feedback to assess and enhance children's semantic competencies. Built in Unity using a Model-View-Controller (MVC) architecture, Osi-Learn captures attention, motivation, and lexical-semantic performance metrics. A pilot study with 29 children, including neurotypical and neurodivergent participants and educators, was conducted. Usability rated excellent (System Usability Scale, SUS = 79.6), and most users displayed high motivation and engagement; however, adaptation was needed for children with Autism Spectrum Disorder (ASD) or intellectual disabilities (ID). This study contributes to Human-Computer Interaction (HCI) and educational game design by presenting a controlled game-based assessment platform that provides real- time feedback on linguistic learning. Future work includes integrating adaptive Artificial Intelligence (AI) feedback and expanding evaluation cohorts.