Observation of cortical state-based learning in infants in a functional near-infrared spectroscopy paradigm.

Significance: Learning can be context-dependent, with better outcomes under some circumstances than others. Adult functional magnetic resonance imaging studies have shown that learning outcomes vary as a function of participants' brain states-patterns of intrinsic neural activity-prior to the learning task. Whether this is also the case in young infants is currently unknown. We report the first functional near-infrared spectroscopy (fNIRS) study that shows prior brain state-dependent learning in a language task in 6.5-month-old infants. Babies whose functional connectivity was lower in the right hemisphere, but not in the left, during a 2-min period prior to the task learned better a grammatical regularity in an artificial grammar learning task. Aim: Adult neuroimaging studies have shown that variability in brain states immediately before specific learning tasks is correlated with variability in learning outcomes. Whether the developing infant brain also shows similar state-based learning is currently unknown. Approach: We have explored whether 6.5-month-old infants' ability to learn artificial grammar was related to their brain state during a 2-min baseline period of rest prior to the grammar task. We have asked if functional connectivity, a global metric of the cortical brain state, as measured by fNIRS, is correlated with learning a non-adjacent regularity in the artificial grammar task. Results: We have found that the overall level of functional connectivity in the 2-min period immediately prior to the learning experience is negatively correlated with the fNIRS measure of learning in the right hemisphere but not in the left. Conclusions: We show for the first time that the cortical state of an infant immediately prior to a learning experience determines how well that infant learns and that this can account for some of the variability in learning outcomes.