This study investigates the effect of tone inventories on brain activities underlying pitch without focal attention. We find that the electrophysiological responses to across-category stimuli are larger than those to within-category stimuli when the pitch contours are superimposed on nonspeech stimuli; however, there is no electrophysiological response difference associated with category status in speech stimuli. Moreover, this category effect in nonspeech stimuli is stronger for Cantonese speakers. Results of previous and present studies lead us to conclude that brain activities to the same native lexical tone contrasts are modulated by speakers’ language experiences not only in active phonological processing but also in automatic feature detection without focal attention. In contrast to the condition with focal attention, where phonological processing is stronger for speech stimuli, the feature detection (pitch contours in this study) without focal attention as shaped by language background is superior in relatively regular stimuli, that is, the nonspeech stimuli. The results suggest that Cantonese listeners outperform Mandarin listeners in automatic detection of pitch features because of the denser Cantonese tone system.
Pitch perception is very important for tone languages, which utilize pitch patterns to distinguish lexical meanings. For example, in Mandarin, a tone language, the same segmental syllable /ma/ means “mother” when produced with a high level pitch contour but means “hemp” when produced with a high rising pitch contour [
Before the pitch signal is transmitted to the cortex level, the frequency following responses (FFR) of pitch at the brainstem have been shown to be sensitive to language-relevant aspects of pitch contours but not specific to speech [
The phonological processing of pitch contours was investigated in the study of [
Although the study compared different conditions involving the category status and context type, it did not compare different levels of attention. Therefore, it is not clear whether the observed group difference stems from the explicit category information (lexical items distinguished by pitch contours) only, which often requires focal attention to do online judgment, or also from automatic feature detection which can be done even without focal attention. In the oddball paradigm, the brain must form a representation of the repeated auditory stimulus before the occurrence of a deviant stimulus, regardless of attention [
In the present experiment, we examine native Mandarin and Cantonese subjects’ electrophysiological responses to the same set of speech and nonspeech tonal stimuli as those used in [
Fifteen native Mandarin speakers (7 F; age: 22.7
This study includes two sets of stimuli, speech syllable /i/ and nonspeech complex tone. Each set included three stimuli drawn from a continuum of eleven stimuli, that is, a within-category deviant (stimulus number 1), a standard (number 4), and an across-category deviant (number 7).
The eleven speech stimuli, each of duration 500 ms, were synthesized from the Tone 1 syllable /i/ uttered by a native Mandarin speaker with the pitch contours manipulated as illustrated in Figure
Pitch contours of stimuli, number 1 within-category deviant, number 4 standard, and number 7 across-category deviant (a). ERPs for standard (averaged over two language groups and two types of context), deviants (averaged over two language groups, two types of context, and two types of deviants), and the difference wave between deviants and standard on Fz (b). Topographic distribution of early mismatch component (250–350 ms) and late mismatch component (500–700 ms) (c).
The stimuli (Figure
After the EEG recordings, a behavioural same/different
Each comparison unit was comprised of all trials in four types of comparisons (AB, BA, AA, and BB). Discrimination response (
The EEG recordings were re-referenced offline against average-mastoid, and 0.5–30 Hz band-pass was filtered. ERPs were 900 ms in duration with a 100 ms prestimulus baseline obtained from each condition and each subject. Trials with ocular artifacts were excluded from averaging. Mismatch components (MC) were obtained by subtracting the ERP of the standard from that of each type of deviant. Two negative components were determined from the maximal negativity of the grand-averaged difference waves across all experimental conditions (see Figure
Three-way mixed design repeated-measures analysis of variance (MANOVA) was carried out on the behavioural and electrophysiological responses. Two within-subject factors were
The
There was only a significant main effect of
Discrimination response
Cantonese
Mandarin
The MANOVA did not reveal any significant effect of mean amplitude (Figure
Mismatch component (MC) averaged over eight electrodes from Cantonese and Mandarin participants. The mean amplitude of early MC (eMC) (a), the peak latency of early MC (b), the mean amplitude of late MC (c), and the peak latency of late MC (lMC) (d).
There was not any significant effect on peak latency (Figure
The posttest behavioural task shows that discrimination of the across-category pair is easier than that of the within-category pair, regardless the type of context or language background. This result verifies that the selection of across-category deviant (number 7) and within-category deviant (number 1) relative to the standard stimulus (number 4) in the electrophysiological recording is appropriate.
The early MC showed an interaction effect of peak latency between category status and context type. It may reflect that across-category and within-category deviants are detected with different levels of difficulties (as reviewed in [
At the late time window, across-category deviant elicits a larger MC than within-category deviant in nonspeech context only. This suggests that the category effect for tones may be present even without focal attention, although at a much later time than the classic MMN [
The late MC has been reported to reflect the summation of MMN generators and memory trace formation on gestalt bases [
The post hoc tests reveal that the category effect from nonspeech context is only present in the electrophysiological responses of Cantonese speakers. No effect reaches significance in the responses of Mandarin speakers. This result is consistent with the hypothesis proposed in the previous study [
There is no evidence for the category effect in the speech context without focal attention in the present study. In contrast, with focal attention, category effect is stronger in the speech context than in the nonspeech context [
The authors declare that there is no conflict of interests regarding the publication of this paper.
The authors thank Manson Fong and Lin Zhou for their help in collecting the experimental data. The experiment was conducted in the Chinese University of Hong Kong and the data analysis was conducted in the Shenzhen Institute of Information Technology, Shenzhen, China. This work is supported in part by Grants from National Natural Science Foundation of China (NSFC: 11074267, 61135003, and 61170283), Research Grant Council of Hong Kong (GRF: 455911), National High-Technology Research and Development Program (863 Program) of China under Grant 2013AA01A212, Ministry of Education in the New Century Excellent Talents Support Program under Grant NCET-12-0649, and Shenzhen Technology Plan under Grants JCYJ20120613102750045 and JCYJ20120615100614529.