Ten right-handed Persian-speaking children with normal visual and auditory acuity, aged from 6 to 7 years, participated in the study. Recording of electroencephalography (EEG) was done by 64 A g/AgCl electrodes. A new auditory paradigm was used with three deviant stimuli (/gam/,/jam/, and/tam/) which differed in the first consonant from a repeated standard word (/dam/). A total of 1500 stimuli, 750 standards and 750 deviants (250 each deviant), were presented by Cogent 2000 running in MATLAB software using two loud speakers.
MMN peaked over the fronto-central sites at around 380–424 ms after the onset of the stimulus. The comparison of the MMN amplitudes elicited by three deviants revealed a significant “initial phoneme type” effect in Fz and Cz (p < 0.05). This negativity was found to be larger for manner and voicing deviants compared with that of the place of articulation (p < 0.001) on midline scalps. The comparison of the MMN latencies revealed no significant main effect of all variables (p > 0.05). Also, the results revealed that only the MMN amplitude for the/gam/deviant correlated with the percentage of correct responses (R = −0.86, p < 0.01).
The current study showed that words can elicit MMN responses in ∼200 ms after the onset of changes. We can objectively evaluate children's neural speech sound discrimination using the developed paradigm in a natural word context. This paradigm can be useful objectively for investigating distinctive features of sounds and phonological discrimination development in normal children.