Language is central to human experience, but we know very little regarding its neural substrate during early development because existing techniques for observing brain activity in infants have low anatomical or temporal resolution, an indirect relationship to neural activity, require more behavioral maturity, or deposit ionizing radiation. In contrast, magnetoencephalography (MEG) is silent and well tolerated by infants. MEG directly and instantaneously measures intraneuronal currents underlying cognitive information processing. We propose to record whole-head 306 channel MEG during word processing by infants. Good anatomical resolution will be obtained by constraining the sources to lie in the cortex of the same individual, reconstructed from MRI performed while the infant is asleep. Our initial focus is on ~14 month old infants, who typically understand a few tens of words, the minimum needed for testing. We will then extend these methods to ~18 month old infants, whose burgeoning expressive vocabularies permit correlation of brain responses with individual differences in language development. In the basic task, infants listen to words they understand, and to noise matched for amplitude, contour and spectral density to each word. In preliminary data, this comparison reliably evokes MEG activity at ~400ms localized to classical language regions. In latency and localization, this activity resembles the N400m evoked by words in adults. Confirmation of this homology will be sought by testing if the infant N400m shares key cognitive characteristics with the adult N400m, specifically sensitivity to repetition and semantic priming. Semantic priming will be examined by presenting a picture and then an auditory word which either matches the picture or not. Preliminary data indicates that the infant N400m, like the adult N400m, is depressed by semantic congruence. The proposed studies will establish a basic pattern of neural activity evoked by words in infants, dissociate it from sensory responses, and demonstrate its relationship to semantic understanding. This would help reveal the neural circuits and synaptic mechanisms used to acquire language understanding, and allow their measurement in infants at risk for language disorders.
Disorders of language acquisition are common and can severely limit psychosocial development. This research will develop a method to safely measure the brain activity underlying infant word-processing, to be used to study the neural basis of language development in both normal and atypically developing infants. Future applications of these methods may include the assessment of children at-risk for delayed language.