Table 1.
Predictions of the MMN results based on different models
Fig 1.
The voice spectrographs and uncorrected fundamental frequency contours (white line) for the three lexical tones: T1 (high level tone), T2 (high rising tone), and T3 (low dipping tone).
(Figure created using PRAAT software: PRAAT Ver. 4.6.01, University of Amsterdam, The Netherland.)
Fig 2.
Grand-average ERPs time-locked to the deviant, standard, control stimuli in the different oddball conditions.
The stimuli in A include T3 in T1/T3 (standard/deviant), T3 (100%), and T3 in T3/T1. The stimuli in B include T1 in T3/T1, T1 (100%), and T1 in T1/T3. The stimuli in C include T3 in T2/T3, T3 (100%), and T3 in T3/T2. The stimuli in D include T2 in T3/T2, T2 (100%), and T2 in T2/T3.
Fig 3.
The MMN-a (difference waveforms, namely, ‘the stimuli presented as deviant in the oddball condition’-minus-‘the same stimuli presented in the 100% control condition’) in the four oddball conditions and their topographies.
Fig 4.
The MMN-b (difference waveforms, namely, ‘the stimuli presented as deviant in one oddball condition’-minus-‘the same stimuli presented as standard in the reversed oddball condition’) in the four oddball conditions and their topographies.
Fig 5.
The results of the cluster-based permutation test to identify MMN (permutation test-1).
The dots and the asterisks in the topography indicate the 43 frontal-central electrodes that were included in this permutation test. The asterisks indicate the electrodes over that the MMN reach significance in this permutation test. Although the cluster-based permutation tests were conducted in the step of 2 ms, the figures were shown in 20 ms step due to the limited space. Within every 20 ms, we got the similar pattern of results.
Fig 6.
The results of the cluster-based permutation test to compare the MMNs in the different conditions.
A: the permutation test over the 43 frontal-central electrodes. B: the permutation test over the 19 left electrodes. The dots and the asterisks in the topography indicate the electrodes that were included in the permutation test. The asterisks indicate the electrodes over that the difference between T2/T3 and T3/T2 condition reach significance. The only significant effect we observed was that the MMN in the T2/T3 condition was larger than that in the reversed T3/T2 condition. Although the cluster-based permutation tests were conducted in the step of 2 ms, the figures were shown in 20 ms step due to the limited space. Within every 20 ms, we got the similar pattern of results.