Plasticity in EEG oscillations associated with auditory verbal learning

Abstract

On a frequent basis, humans need to vocally learn and remember a list of unrelated items. Advances in cognitive neuroscience have begun to identify the brain regions involved. However, the mechanisms by which those regions interact during learning remain elusive. There is growing support for the proposition that the oscillations within and among these regions provide a substrate for their interaction. This proposition is investigated in the present study by evaluating changes in brain oscillations during verbal learning. Previous research in this domain has provided only limited clues about the influence of ecologically significant factors such as repetition and mnemonics on learning performance and brain dynamics. The present study evaluates independent components analysis, power spectral analysis, and coherence of 32-channel electroencephalogram recorded while subjects learned a list of unrelated nouns. The learning task included repetition and either conventional spoken learning or learning with a musical mnemonic. The results show that as subjects make the transition from repetition to learning, their alpha frequency band activity undergoes a state transition from synchronized to desynchronized oscillations over right posterior cortex. A similar state transition is observed when learning includes a musical mnemonic, but its topographic distribution in the right hemisphere is reversed and relative desynchrony occurs over right prefrontal cortex. The results suggest that verbal learning, even in the context of repetition, is associated with modulation of brain oscillations and that an anatomically distinct network is recruited when learning includes a musical mnemonic. The study has implications for the basic cognitive neuroscience of learning, clinical rehabilitative applications using learning mnemonics, and the architectures of biologically-plausible machine learning algorithms.