| Invited Symposia | Symposia

Note: Symposia have not yet been scheduled.

Attention, Gaze, and Speech Processing During Learning Across the Lab-to-Classroom Continuum

Chairs: Yimeng Wang¹, Lisa D. Sanders¹; ¹University of Massachusetts Amherst
Presenters: Xiaorui Xue, Julie M. Schneider, Elizabeth Jiwon Im, Lisa D. Sanders

Cognitive neuroscience research has specified the ways in which perceptual processing, auditory and visual attention, and engagement influence learning outcomes. However, most of this research has been conducted under carefully controlled conditions that fail to capture the complexities of multi-learner environments, which include social interactions and multiple sources of distraction. Advances in portable EEG, eye-tracking systems, and machine learning enable the study of these cognitive processes in more diverse, dynamic, and naturalistic settings. This symposium integrates research across online and in-person learning, visual and auditory complexity, and more and less constrained social interactions in young children through young adults. Starting with learner-driven processes, the symposium examines engagement across two central learning contexts in contemporary education, online learning and in-person classrooms, to index relationships between engagement and educational outcomes. To further elucidate these attentional processes and their underlying neural mechanisms, talks will describe research about how classroom acoustics affect neural speech tracking and language learning, how children and adolescents selectively sample and engage with visual information across home and classroom environments, and how engagement with live lectures can be decoded from multi-modal data to predict learning. Together, this symposium highlights the value of ecologically valid approaches for understanding attention and engagement in real-world learning contexts. The broader goal of this line of research is to determine how the basic cognitive neuroscience principles relating perception, attention, and engagement to learning outcomes can be leveraged to support real learners in a variety of learning environments.

Presentations

Following the Wandering Mind: Multimodal Approaches to Attention and Learning

Xiaorui Xue¹, Vishal Easwar¹, Jason Geller¹, Na’ama Y. Av-Shalom¹, Dax Ovid², Ido Davidesco¹; ¹Boston College, ²University of Georgia

Mind-wandering (MW) occurs when attention shifts away from an ongoing task toward task-unrelated thoughts. Although numerous studies demonstrate that MW is common during learning, it remains unclear whether its consequences depend on how often it occurs, where attention is directed, and what learners think about when their minds wander, particularly when applied to real-world learning settings. Across two studies, we combined eye tracking and thought probes to examine attention and learning in both online and in-person learning. In the classroom study (N = 28 undergraduate students), self-report thought probes were used to detect MW events while eye tracking glasses tracked students’ gaze. During lectures, students were either allowed or not allowed to use personal devices (e.g., laptops). The use of personal devices was found to alter where students looked during MW events without changing how frequently they mind-wandered, revealing differences in visual attention that self-report measures alone did not capture. In the online study, participants (N = 112) viewed a recorded lecture that was interspersed with either guided breaks, where students were prompted to think about a lecture-related question, or unguided breaks, where students were left to mind-wander. We found that only when the experimental manipulation affected the content of internally directed thought, significant learning gains were detected, suggesting that the presence of a break alone does not improve performance. Together, these findings demonstrate the need to triangulate multiple research approaches to uncover the nature and consequences of MW in naturalistic learning environments.

Tracking Speech Through Noise: Neural and Behavioral Consequences of Classroom Acoustics in Children

Julie M. Schneider¹, Jacob P. Momsen², Meli Ayala³, Zoe Mao¹, Jennie K. Grammer¹; ¹University of California, Los Angeles, ²Yale School of Medicine, ³University of Delaware

The built environment plays an underappreciated role in shaping children's learning, particularly with regard to classroom acoustics. Although guidelines exist to limit classroom noise, adherence is voluntary, and many children learn in acoustically suboptimal environments. Noisy environments affect children's speech perception, literacy, attention, memory, and brain development, yet prior work has focused on external noise sources rather than the internal classroom acoustics that directly shape everyday learning. This talk examines how classroom acoustics shape the neural mechanisms supporting language comprehension, from syllable-level to sentence processing. Using temporal response function analyses of EEG data from 25 school-aged children (7–12 years), we show that noise reduces neural speech tracking, with downstream consequences for comprehension. During an auditory statistical learning paradigm, children learned novel word patterns significantly worse in noise than in quiet (p = .014), an effect linked to weaker neural tracking of speech at the syllable level (p = .049). During a narrative comprehension task, noise significantly reduced children’s neural tracking of speech (p = .048), resulting in poorer post-listening comprehension (p = .035). Together, these findings suggest that degraded acoustic environments constrain learning by disrupting neural synchronization with speech, from the syllable to the sentence level. These results provide a translational bridge from neuroscience to practice, informing policy and design standards for optimal learning environments. By linking measurable features of classroom design to neural processing and learning outcomes, this work positions classroom acoustics as a modifiable feature of the built environment with direct implications for educational equity.

Do you see what I see? Mobile eye-tracking in natural environments reveals selective social-visual engagement and developing visual-field biases

Elizabeth Jiwon Im¹, Kalanit Grill-Spector¹; ¹Stanford University

As adults, we efficiently recognize faces and read words, yet these abilities have different developmental trajectories: face recognition emerges early, while reading is acquired later through education. Both face and word representations continue to develop neurally and behaviorally into adulthood, with longitudinal data finding that face- and word-selective regions expand by recycling cortical regions previously selective for limbs. Eye-tracking studies in lab settings reveal corresponding developmental differences in viewing behavior of faces and words, suggesting that visual demands and visual diet may shape cortical development. However, how visual diet and fixation patterns change across childhood remains unknown, as existing naturalistic datasets are limited in age range and environment and lack eye-tracking data. For the first time, we use mobile eye-tracking to compare the visual diet and gaze behavior of children (4–6 yrs) and teenagers (14–17 yrs) in home and classroom environments. We find that the prevalence of social-visual information in children’s and teenagers’ visual diet does not always predict how they engage with that information. Moreover, different contexts lead to different visual statistics and gaze behaviors, even within the same age group. We also find that visual-field biases continue to develop from childhood into adolescence. Together, our findings suggest that children and teenagers sample the world differently, with context shaping gaze behavior across development, suggesting that models of visual development should incorporate not only scene statistics but also age- and context-dependent fixation statistics.

Speech Processing and Engagement Decoding in Live Lectures

Lisa D. Sanders¹, Yimeng Wang¹, Sidharth Anupkrishnan¹, VP Nguyen¹; ¹University of Massachusetts Amherst

Decoding student engagement is important for understanding and improving learning effectiveness. Although behavioral and physiological measures have been used to assess engagement, most studies have been conducted under relatively controlled experimental conditions. Furthermore, the role of live instruction in naturalistic learning settings, particularly how students respond to instructors’ spoken language, has not been fully captured or incorporated into engagement-decoding framework. To address this gap, a live instructor delivered real-time lectures to participants while scalp-EEG, mobile headband EEG, and eye-tracking data were recorded simultaneously. Instructor speech features were integrated with headband EEG and gaze data in a multimodal Convolutional Recurrent Neural Network (CRNN) model. The model achieved an MAE of 0.628, balanced within-one accuracy of 91.2%, and QWK of 0.249, with performance comparable to a trimodal configuration using scalp EEG instead of headband EEG. Input-ablation analyses showed that removing gaze produced the largest performance degradation, followed by speech envelope and headband EEG, demonstrating the contribution of speech to engagement decoding. To further characterize neural responses to live speech, we examined neural encoding across multiple levels of speech representations. Greater reported engagement was associated with enhanced neural envelope tracking and stronger encoding of higher-level linguistic features, such as word surprisal, potentially reflecting increased top-down predictive processing during engagement. Overall, these interdisciplinary analyses demonstrate the feasibility of scalable, mobile engagement decoding during live lectures, while highlighting instructor speech as both a valuable source of information for engagement decoding and a window into the underlying neural mechanisms of attentional engagement.