Browsing by Author "Roikonen, Mira"

Rapid learning or fast mapping reflects the human brain’s ability to form new memory traces to novel words during exposure without the need for a long overnight consolidation period before the word can be used in conversation. This ability to acquire new words almost instantaneously may very well reflect how well-tuned the human language systems are to the phonemes of the native language. However, the neural basis of rapid learning has been largely unknown until recent neuroimaging studies. In this study on adult learners (n = 15), I recorded brain’s event-related potentials elicited by three different types of auditory bi-syllabic stimuli (native words, native pseudowords and pseudowords with unfamiliar phonemes, all acoustically closely matched) in a passive EEG-recording session before and after subjects participated in two types of training conditions. In the attend condition, subjects listened to a flow of stimuli while pressing a button when a target stimulus appeared. In the articulation condition subjects repeated out loud the heard stimuli. An auditory memory recognition test was administered after training to allow the comparison of neural learning effects to observable change in behaviour. Larger evoked responses were expected to correlate with better performance in the recognition task. All analyses were time-locked to the onset of the second syllable (critical disambiguation point/recognition point). A two-peak waveform was observed to all stimuli after both conditions, with the earlier peak appearing circa 40 ms and the later peak circa 140 ms after second syllable onset. Unlike in similar previous studies where responses increased as a result of learning, all responses decreased in magnitude. No statistically significant differences between the conditions were observed. This may have been due to either the small sample size, test subject fatigue or suppression effects due to repetition, masking any possible learning effects. For the late peak, native pseudowords evoked significantly stronger responses than native words or non-native pseudowords. Performance in the memory recognition task was good (above chance for all stimuli in both conditions), and as such learning cannot be excluded even though statistically significant differences in the evoked responses were not found. Further research and re-exploration of the data acquired here utilising source modelling might enable to assess in more detail the effect of attentive listening vs. articulation in rapid learning.