Attention in laboratory tasks versus naturalistic learning contexts: Heterogeneity in measurement and mechanisms.

An important though less studied element in neuroscience of classroom learning, is translation of laboratory-derived metrics to quantify learning in naturalistic settings. In the current study we sought to test if a common neural metric of visual attention, alpha-range (8-12 Hz) oscillations, correlates across laboratory versus naturalistic learning activities. Thus, we test if laboratory-derived measures capture the same mechanistic sources when deployed in more naturalistic environments. We acquired electroencephalography signals in eighty 6-10-year-old children while they participated in naturalistic learning activities and during a laboratory spatial working memory task. We found that alpha power modulations during the activities did not correlate with modulations during stimulus processing in the laboratory task, despite (i) significant correlations across baseline periods of each context, (ii) internal consistency across contexts in the relationship of baseline and modulation effects, and (iii) significant association with behavioral indicators of attention. The results point to measurement heterogeneity across contexts, consistent with different mechanisms contributing to these metrics during learning activities versus laboratory task. The results have implications for experimental design in the domain of education neuroscience and argue for consideration of multiple processes - such as multisensory processing, arousal, motivation, and motor activity - when interpreting students' observable behavior.