npj Science of Learning最新文献

The use of artificial intelligence in education (AIEd) has grown exponentially in the last decade, particularly intelligent tutoring systems (ITSs). Despite the increased use of ITSs and their promise to improve learning, their real educational value remains unclear. This systematic review aims to identify the effects of ITSs on K-12 students' learning and performance and which experimental designs are currently used to evaluate them. The 28 studies analyzed in this systematic review included a total of 4597 students (N = 4597) and used quasi-experimental designs with varying intervention durations. Overall, our findings suggest that the effects of ITSs on learning and performance in K-12 education are generally positive but are found to be mitigated when compared to non-intelligent tutoring systems. However, additional research with longer interventions and increased sample sizes with greater diversity is warranted. Additionally, the ethical implications of using AI for teaching should be investigated.

Eye movement patterns and consistency during face recognition are both associated with recognition performance. We examined whether they reflect different mechanisms through EEG decoding. Eighty-four participants performed an old-new face recognition task with eye movement pattern and consistency quantified using eye movement analysis with hidden Markov models (EMHMM). Temporal dynamics of neural representation quality for face recognition were assessed through decoding old vs new faces using a support vector machine classifier. Results showed that a more eye-focused pattern was associated with higher decoding accuracy in the high-alpha band, reflecting better neural representation quality. In contrast, higher eye movement consistency was associated with shorter latency of peak decoding accuracy in the high-alpha band, which suggested more efficient neural representation development, in addition to higher ERP decoding accuracy. Thus, eye movement patterns are associated with neural representation effectiveness, whereas eye movement consistency reflects neural representation development efficiency, unraveling different aspects of cognitive processes.

Learning behavioural responses and adapting them based on feedback is crucial from a young age, continuing to develop into young adulthood. This study examines the development trajectory and contributing factors from childhood to adulthood using a reversal learning paradigm. We tested 202 participants aged 10 to 22 in an online study, where they learned and reversed stimulus-outcome associations in a new blocked design paradigm and were assessed for working memory capacity. Results showed that reversal learning performance improved with age, particularly for 10- to 14-year-olds. Flexible responses to negative feedback correlated with better reversal learning. Additionally, pubertal development and working memory were positively associated with reversal learning. These findings align with previous research, highlighting flexible feedback responses as a key factor in reversal learning. As the overall rate of flexible reactions did not change with age, it could support reversal learning independent of age, potentially changing its role during development.

Contextual shifts are crucial for episodic memory, setting event boundaries during event segmentation. While lab research provides insights, it often lacks the complexity of real-world experiences. We addressed this gap by examining perceptual and conceptual boundaries using virtual reality (VR). Participants acted as salespeople, interacting with customers in a VR environment. Spatial boundaries separated visually distinct booths, while conceptual boundaries were defined by customer requests. Memory was assessed through a recency discrimination task. Results indicated boundary crossings impaired sequence memory, consistent with previous findings. Crucially, conceptual boundaries, but not spatial boundaries, significantly influenced the accuracy of sequence memory, suggesting that top-down processes dominate bottom-up perceptual processes in naturalistic event segmentation. Confidence in correct responses indicated that perceived memory quality was highest when participants stayed within and did not cross both spatially and conceptually defined events. Findings highlight VR's effectiveness for studying hierarchical contextual influences in interactive episodic memory tasks.

The International Science and Evidence Based Education Assessment examined whether current education systems develop each person's full potential (aligned with the UN Declaration of Human Rights) and contribute to Sustainable Development Goal 4. Embracing a multidisciplinary approach, nearly 300 scientists from 45 countries conducted the assessment, calling for a shift in education's focus from economic growth to fostering human flourishing. Key findings included (a) the need for an integrative approach to learning, (b) moving beyond meritocracy and exploring potentiality as a better measure of student learning potential, and (c) using technology judiciously for scalable, equitable, and personalised learning. This paper seeks to highlight themes that were foundational to the assessment but not fully discussed within it. It advocates a global, transdisciplinary research agenda to close evidence gaps and inform policy to consider the complexity of the educational system and the need to think beyond existing conventions.

Learning to perform multiple tasks robustly is a crucial facet of human intelligence, yet its mechanisms remain elusive. Here, we formulated four hypotheses concerning task interactions and investigated them by analyzing training sequence effects through a continual learning framework. Forty-nine subjects learned seven tasks sequentially, each of the seven groups following a distinct sequence. Results showed that subjects learning a task later in a sequence exhibited poorer performance in six tasks (Contrast, Vernier, Face, Motion, Auditory, and N-back tasks, except for the Shape task) compared to those who learned this task earlier. Interestingly, sequence position had minimal impact on forgetting. A complementary dual-task experiment corroborated these findings. Through detailed analyses of session and block learning curves, we revealed task-specific anterograde interference, but no retrograde interference. These findings support the integrated reweighting theory and shed light on the meta-plasticity mechanism governing how human brain balances plasticity and stability.

Breaking down complex words into smaller meaningful units (e.g., unhappy = un- + happy), known as morphemes, is vital for skilled reading as it allows readers to rapidly compute word meanings. There is agreement that children rely on reading experience to acquire morphological knowledge in English; however, the nature of this experience has remained unclear. We quantify the morphological information in 1200 popular children's books and offer the first concrete description of how readers may learn affix morphemes through real-life text input. Our account considers the realities of morpheme presentation in natural language, such as the low number of distinct words in which affixes appear and the fact that their identification often requires specialised linguistic knowledge. This theory further accounts for the challenge posed by spelling patterns that may lead to incorrect morphological parsing. We conclude by exploring the implications of our findings for instructional programmes in morphology.

This systematic review explores the burgeoning intersection of Artificial Intelligence (AI) applications and self-regulated learning (SRL) in higher education. Aiming to synthesize empirical studies, we employed a qualitative approach to scrutinize AI's role in supporting SRL processes. Through a meticulous selection process adhering to PRISMA guidelines, we identified 14 distinct studies that leveraged AI applications, including chatbots, adaptive feedback systems, serious games, and e-textbooks, to support student autonomy. Our findings reveal a nuanced landscape where AI demonstrates potential in facilitating SRL's forethought, performance, and reflection phases, yet also highlights whether the agency is human-centered or AI-centered leading to variations in the SRL model. This review underscores the imperative for balanced AI integration, ensuring technological advantages are harnessed without undermining student self-efficacy. The implications suggest a future where AI is a thoughtfully woven thread in the SRL fabric of higher education, calling for further research to optimize this synergy.