The rapid diffusion of generative artificial intelligence (GenAI) is reorganizing the ecology in which cognition, social understanding, agency, authorship, and emotional regulation develop. This critical integrative review treats GenAI as a developmentally consequential interactive artifact: responsive, adaptive, linguistically fluent, multimodal, and increasingly embedded in learning, play, and communication. Drawing on sociocultural, constructivist, ecological, extended-cognition, AI-literacy, and hybrid-learning perspectives, we introduce the Hybrid Mind Framework. It defines human-AI coupling as recurrent, contingent coordination and distinguishes immediate AI-supported performance from durable changes in unaided competence, competence during coupling, trust, support-seeking, or authorship. Four domains are considered: cognitive hybridization, social-epistemic calibration, affective scaffolding and support-seeking, and agency/authorship. The review integrates emerging GenAI evidence with research on earlier responsive technologies and differentiates early childhood, middle childhood, early adolescence, and later adolescence. Existing evidence indicates possible short-term affordances and risks, but long-term developmental effects remain largely untested. Rather than treating GenAI as either a pedagogical miracle or a developmental threat, we propose that AI be designed and governed as part of children's developmental ecology. Testable propositions and age-calibrated implications distinguish observed findings from theory-derived and precautionary claims.
Early mathematics experiences at home and in school lay the groundwork for children's later academic success. Spontaneous focus on mathematical features (SFOM)—an individual's tendency to focus on mathematical features, such as numerosity and symbolic numbers, without explicit prompting—may be a predictor of children's mathematics performance. This meta-analysis synthesizes findings from 47 studies, encompassing 333 effect sizes, to examine the correlation between SFOM and mathematics achievement. Results revealed a significant moderate correlation (r = 0.29, 95% CI [0.24, 0.34], p < .001) between SFOM and mathematics achievement. To explain variability across studies, we explored potential moderators, including SFOM types and tasks, math assessment characteristics, study designs, and participant characteristics. Findings indicate that the SFOM-math correlation is stronger when SFOM is measured using finding, imitation tasks compared to matching tasks. Aggregated measures containing multiple tasks produced stronger correlations than single-task measures (e.g., imitation, matching, model, picture, or selection tasks). Longitudinal studies reported stronger correlations than concurrent studies, highlighting the predictive value of SFOM over time. Additionally, the relation between SFOM and mathematics achievement is weaker in post-secondary participants compared to younger cohorts, including preschool, elementary, and secondary students.
Hands are among the earliest and most socially meaningful body parts infants see, feel and use to act on the world. From birth, they provide a continuous source of visual and multisensory information, supporting exploration and emotion sharing. Yet the developmental origins of visual hand recognition, and particularly recognition of one's own hand, remain largely unknown.Here, we synthesize behavioral, neural, and developmental evidence to argue that hands hold a privileged status in early body representation. Dense multisensory exposure, continuous visual access, and early social use create a self-reinforcing loop that tunes infants to the form, movement, and meaning of hands.We propose a developmental framework in which early predispositions for body processing interact with visual and multisensory experience. These elements keep affecting one another forming a cycle that may support early hand-specific tuning, driven by the privileged role hands play in infants' perceptual, motor and social word. Testing this hypothesis offers a unique entry point into the origins of body ownership and agency, and provides new insights into the origins of bodily self-awareness and early social cognition.
Neural sensitivity to rhythm emerges prenatally, yet observable movement synchronization to an external beat occurs later and more gradually over development. Integrating findings from developmental psychology, neuroscience, music cognition, and social interaction research, we propose a triadic framework suggesting that infants' rhythmic synchronization emerges from the interaction of three components: rhythm perception, motor development, and social scaffolding. Based on this framework, early synchronization abilities may be present earlier than commonly assumed, but their expression depends on interactions between perceptual capacities, developing motor abilities and social scaffolding, providing predictive cues and reward. We further argue that the detection of early rhythmic behavior may depend on methodological and analytical approaches as infant rhythmic behavior is variable, context-dependent, and expressed intermittently. Methodological and analytical improvements are outlined to better capture the dynamic development of rhythmic coordination in infancy. Understanding variability in early rhythmic behavior will provide insight into developmental trajectories and be informative for identifying early markers of potential deficits in rhythm processing, which have been described in neurodevelopmental disorders.
The role early vocabulary plays in later language development and academic success has been widely acknowledged, and common developmental trajectories are known. Nevertheless, how general cognitive processes support individual trajectories of vocabulary development is not well understood. We explored this issue via a scoping review covering literatures on the relations between nine general cognitive processes (visual processing, visual/spatial attention, attention shifting, sustained attention, visual working memory, novelty detection, object recognition, statistical learning, and declarative memory) and vocabulary development. Processes were selected based on their instantiation in a successful neural-process model of word learning. Auditory processing and discrimination was also reviewed to inform future theory and model development. Systematic database searches were conducted for each process. Papers returned from three databases were screened, reviewed, and the relevant information for each process extracted. While the final number of included papers was limited, we found some evidence of links between nine of the processes and vocabulary development. Evidence of processing differences in children at risk for language impairment was also found for three processes. We discuss the revealed relations and gaps in the literature. We also provide a critical appraisal of the literature and implications for both theory and future work, highlighting the need for integrative longitudinal work and the usefulness of formal models in examining how multiple processes interact to support development.
