Modulation of the mismatch negativity (MMN) to vowel duration changes in native speakers of Finnish and German as a result of language experience

https://doi.org/10.1016/j.ijpsycho.2007.10.012Get rights and content

Abstract

While crucial for phoneme distinctions in the Finnish language, mere vowel duration is of lower relevance as a phonetically distinctive cue in the German language. To investigate the pre-attentive processing of vowel duration between these two languages, the mismatch negativity (MMN), a component of the auditory event-related potential (ERP) that is an index of automatic auditory change detection, was measured in Finnish and German native speakers for vowel duration changes embedded in the pseudoword sasa. Vowel duration changes thereby were presented as a shortening or a lengthening of either the first- or second-syllable vowel. An additional non-speech condition measured the MMN to duration and frequency changes in tones.

In both language groups, diminished MMN amplitudes for the shortening of vowel duration in the word-final syllable suggested a generally more difficult discrimination of vowel duration in a word-final position. Further, shorter MMN latencies for the Finns than the Germans for vowel duration as well as tone duration deviants suggested a generally higher sensitivity to duration contrasts in the Finnish language group. No latency difference between the groups was found for tone frequency processing. Moreover, the Finns, but not the Germans, showed a leftward shift of the MMN scalp distribution for changes in vowel duration, whereas the MMN topography was highly similar between both groups in the tone condition. An enhanced phonetic processing of vowel duration changes and possibly an enhanced processing of sound duration in general is thus indicated for the Finns as a result of their extensive linguistic experience with phonetically distinctive vowel duration contrasts in the native language.

Introduction

Cross-linguistic studies demonstrated a substantial influence of language experience on the perception and the processing of speech sounds. Listeners usually show increased discrimination sensitivity to those acoustic cues in speech sounds that signal phoneme distinctions in the native language, i.e. that distinguish the language-specific smallest units of speech that can change the meaning of a word. On the other hand, speech-sound differences which are not relevant for phoneme distinctions in the native language, are typically difficult to discriminate (e.g., Iverson et al., 2003, Kuhl et al., 1992, Zhang et al., 2005; see also Trubetzkoy, 1939/1969). This effect can already be observed at the end of the first year of life (Kuhl et al., 1992, Cheour et al., 1998).

An electrophysiological measure that reflects differences in sound discrimination sensitivity is the mismatch negativity (MMN), which is a component of the auditory event-related potential (ERP) and which indexes auditory change detection. The MMN (Näätänen et al., 1978) is elicited by an irregular sound (the deviant) occurring in an otherwise regular sequence of so-called standard sounds (for review see e.g., Schröger, 1998). This negative deflection is observed in the ERP to the deviant sound compared to the ERP to the standard sound and is best visible in the deviant-minus-standard difference wave. The MMN displays a frontocentral scalp distribution and a latency between 100 and 250 ms after deviance onset. Main neural generators of the MMN have been localized bilaterally in the auditory cortex with additional bilateral generators in the frontal lobe (e.g., Alho, 1995, Rinne et al., 2000). Whereas especially the frontal generator usually shows a right-hemispheric predominance when non-speech sounds are presented (e.g., Levänen et al., 1996, Paavilainen et al., 1991), speech stimuli often elicit a more symmetrical or even left-hemispherically lateralized MMN or magnetic MMNm (e.g., Alho et al., 1998, Rinne et al., 1999, Takegata et al., 2004, Tervaniemi et al., 2000), associated with enhanced activity of generators in the left-hemispheric auditory cortex (Näätänen et al., 1997, Pulvermüller et al., 2001).

Importantly, the MMN is elicited irrespectively of voluntary sound discrimination and even for an auditory stimulation presented outside the focus of voluntary attention. It thus indexes auditory change detection at an automatic, pre-attentive level. Yet, the parameters of the MMN closely correlate with the behavioral discrimination performance measured in additional active discrimination tasks: higher discrimination sensitivity here has been found to be associated with larger MMN amplitudes and/or shorter MMN latencies (e.g., Kujala et al., 2001, Näätänen et al., 1993, Novitski et al., 2004, Tiitinen et al., 1994; for a review see also Näätänen, 2001). Additionally, enhanced MMN amplitudes and shorter MMN latencies have been found as a consequence of short-term auditory discrimination training for non-speech sounds (e.g., Menning et al., 2000, Näätänen et al., 1993) as well as speech sounds (e.g., Kraus et al., 1995, Menning et al., 2002). For the processing of speech-sound changes, enhanced amplitudes and shorter latencies of the electric and magnetic MMN response were frequently found for native, in contrast to non-native phoneme contrasts, reflecting the effects of native-language-specific phoneme perception (e.g., Cheour et al., 1998, Dehaene-Lambertz, 1997, Näätänen et al., 1997, Sharma and Dorman, 2000, Winkler et al., 1999; see Näätänen 2001 for review).

German and Finnish, the two here investigated languages, differ in the extent in which they apply vowel duration changes to signal phonetic distinctions. In so-called quantity languages, such as Finnish, Japanese or Thai, “long” and “short” vowels (and/or consonants) are distinguished as different phonemes based on the relative physical durations of the speech sounds in a sequence. In Finnish, for example, vowel duration contrasts (so-called “quantity oppositions”) are used to distinguish words (e.g., /tuli/ [short] = ‘fire’ vs. /tu

li/ [long] = ‘wind’) as well as to mark grammatical inflections and cases (e.g., /talon/ [short] = ‘of the house’ vs. /talo
n/ [long] = ‘into the house’). Non-quantitative languages as e.g., Russian, on the other hand, do not differentiate between different vowel quantities. Although German conventionally differentiates between long and short vowel phonemes, it can neither be regarded as a genuine quantity nor a pure non-quantity language. German quantity oppositions are concurrently accompanied by large spectral differences for most vowels1, and it is not satisfactorily clarified which cues, durational or spectral, German listeners actually use to distinguish between both instances (cf. Becker, 1998, Ramers, 1988). Furthermore, compared to Finnish, German shows a rather restricted use of these vowel oppositions: these occur predominantly in stressed syllables (and thus not in the typically unstressed, word-final positions) and do not indicate grammatical functions. On the other hand, vowel duration in German is an important cue for word stress (see Dogil and Williams, 1999). The role of vowel duration as a phonologically distinctive cue thus is more restricted and controversial in German (cf. Becker, 1998), whereas it is highly relevant for phonological distinctions in Finnish (see e.g., Lehtonen, 1970, Suomi and Ylitalo, 2004, Suomi, 2005).

Recent studies showed that linguistic experience with vowel duration as a phonetically distinctive cue is reflected in the pre-attentive MMN response: Native speakers of Finnish (quantity language), compared to native speakers of Russian (non-quantity language), were found to show an enhanced MMN to deviants in vowel duration (Nenonen et al., 2003, Nenonen et al., 2005, Ylinen et al., 2006). This difference was apparent even though the investigated children (aged 10 to 14 years) were fluent second-language (L2) users of the quantity language Finnish. Together with studies applying fMRI or NIRS to investigate the processing of vowel duration (e.g., Gandour et al., 2002, Minagawa-Kawai et al., 2005), these MMN results suggest that vowel duration processing is crucially influenced by the linguistic experience with vowel duration in the phonological system of the native language. Furthermore, they indicate an influence of the linguistic status of the native language as a quantity or non-quantity language already at the pre-attentive level of speech-sound processing.

The intermediate status of the German language regarding the use of distinctive vowel duration differences therefore stimulates the question to which extent vowel duration processing, especially on the pre-attentive level, might differ when native speakers of German are contrasted to speakers of a genuine quantity language such as Finnish. Previously, Menning et al. (2002) found larger magnetic MMN responses with shorter latencies for Japanese subjects when comparing the processing of duration contrasts embedded in Japanese words between native speakers of German and Japanese. However, since the stimuli were meaningful words in Japanese but not German, this might have at least partly been due to lexical processing in the Japanese subjects, which has been found to foster the MMN response (e.g., Pulvermüller et al., 2001). In a recent study, Tervaniemi et al. (2006) found enhanced MMN responses for Finns in contrast to Germans for deviants in tone duration, but not for deviants in tone frequency. The authors concluded that the different long-term experience with vowel duration as a linguistic cue in Finnish and German has been transferred to the processing of sound duration on a non-speech level. Since the processing of duration, however, may differ between speech and non-speech sounds (e.g., Nenonen et al., 2003), it is fundamentally interesting to compare duration processing between the two language groups by using phonetic stimuli.

The present study measured the electroencephalographic MMN response elicited by changes in vowel duration in adult native speakers of Finnish and German. Vowel duration changes were presented in the context of a naturally sounding pseudoword, the pseudoword sasa. On the one hand, the choice of pseudoword stimuli avoided the contribution of possible effects of lexical processing. On the other hand, it allowed a direct comparison of vowel duration processing in the first- and second-syllable position of a word within as well as between the language groups. In an additional tone condition, tone duration and tone frequency changes were presented similarly as in Tervaniemi et al. (2006), to assess language-specific differences in the duration processing of non-speech sounds (duration deviant), as well as possible non-language related differences between the groups (frequency deviant).

Section snippets

Participants

A group of Finnish and a group of German native speakers were tested. Finnish participants were recorded in Helsinki (Cognitive Brain Research Unit, University of Helsinki, Finland); German participants were recorded in Leipzig (Research Group of Cognitive and Biological Psychology, University of Leipzig, Germany) by the same experimenter (first author). For each language group, 13 participants who were matched between the groups in sex and in approximate age were included in the data analysis

Presence of MMN

Negative peaks were observed in the deviant-minus-standard difference waves for both language groups in all deviant conditions (see Fig. 3; see Fig. 2 for examples of standard and deviant ERPs). These negative deflections were statistically significant for all groups and conditions except the Second-Syllable Duration-Decrement condition for the Germans (see Fig. 5). They were characterized by typical MMN latencies (120 to 180 ms after deviance onset) and scalp distributions (a frontocentral

Discussion

We investigated whether Finnish and German native speakers differ in the pre-attentive processing of vowel duration. Whereas this feature is highly phonetically relevant in Finnish, it is less important in the German language. Both groups thus differ in their linguistic experience with vowel duration contrasts. The MMN elicited by decrements and increments of vowel duration in the first or second syllable of the pseudoword sasa, as well as the MMN elicited by duration and frequency deviants in

Conclusion

The present study showed differences in the pre-attentive processing of vowel duration between Finns and Germans as a consequence of the long-term experience with the native language. Generally, the discrimination of vowel duration thereby was more difficult for a decrement of duration in a word-final position. Shorter MMN latencies for vowel duration and tone duration contrasts for the Finns, however, indicated an enhanced sensitivity to duration contrasts for the Finns due to the frequent

Acknowledgements

Parts of these data were acquired during a stay of Ursula Kirmse at the Cognitive Brain Research Unit in Helsinki, Finland. We wish to thank the members of the CBRU for their support in various matters at this time. We further thank our anonymous reviewers for their very helpful comments on earlier versions of this manuscript.

References (56)

  • KujalaT. et al.

    The mismatch negativity as an index of temporal processing in audition

    Clin. Neurophysiol.

    (2001)
  • NäätänenR. et al.

    Early selective-attention effect on evoked-potential reinterpreted

    Acta Psychol. (Amst.)

    (1978)
  • NenonenS. et al.

    Linguistic relevance of duration within the native language determines the accuracy of speech-sound duration processing

    Cogn. Brain Res.

    (2003)
  • NenonenS. et al.

    Speech-sound duration processing in a second language is specific to phonetic categories

    Brain Lang.

    (2005)
  • NovitskiN. et al.

    Frequency discrimination at different frequency levels as indexed by electrophysiological and behavioral measures

    Cogn. Brain Res.

    (2004)
  • PaavilainenP. et al.

    Right-hemisphere dominance of different mismatch negativities

    Electroencephalogr. Clin. Neurophysiol.

    (1991)
  • PeltolaM.S. et al.

    Early exposure to non-native language alters preattentive vowel discrimination

    Neurosci. Lett.

    (2005)
  • PulvermüllerF. et al.

    Memory traces for words as revealed by the mismatch negativity

    NeuroImage

    (2001)
  • RinneT. et al.

    Separate time behaviors of the temporal and frontal mismatch negativity sources

    NeuroImage

    (2000)
  • SuomiK.

    Temporal conspiracies for a tonal end: segmental durations and accentual f0 movement in a quantity language

    J. Phon.

    (2005)
  • SuomiK. et al.

    On durational correlates of word stress in Finnish

    J. Phon.

    (2004)
  • SussmanE. et al.

    Automatic and controlled processing of acoustic and phonetic contrasts

    Hear. Res.

    (2004)
  • WeberC. et al.

    Discrimination of word stress in early infant perception: electrophysiological evidence

    Cogn. Brain Res.

    (2004)
  • WinklerI. et al.

    Pre-attentive detection of vowel contrasts utilizes both phonetic and auditory memory representations

    Cogn. Brain Res.

    (1999)
  • YlinenS. et al.

    Mismatch negativity (MMN) elicited by changes in phoneme length: a cross-linguistic study

    Brain Res.

    (2006)
  • ZhangY. et al.

    Effects of language experience: neural commitment to language-specific auditory patterns

    NeuroImage

    (2005)
  • AlhoK.

    Cerebral generators of mismatch negativity (MMN) and its magnetic counterpart (MMNm) elicited by sound changes

    Ear Hear.

    (1995)
  • BeckerT.

    Das Vokalsystem der deutschen Standardsprache

    (1998)
  • Cited by (53)

    • Listen-and-repeat training improves perception of second language vowel duration: Evidence from mismatch negativity (MMN) and N1 responses and behavioral discrimination

      2020, International Journal of Psychophysiology
      Citation Excerpt :

      Iverson et al. (2003) found that speakers of Japanese have a highly reduced sensitivity to sound frequency changes most relevant for the detection of the English /r/-/l/ contrast, explaining the major difficulties Japanese speakers have with this contrast. Similar results have been found for Finnish speakers, who, as speakers of a quantity language, show a higher sensitivity to phonetic duration differences than Germans (Kirmse et al., 2008); furthermore, Finns, unlike Russians, also seem to have developed phonemic categories for duration differences (Ylinen et al., 2005b). Popular models of second language acquisition, such as the Speech Learning Model (SLM) by Flege (e.g. Flege, 1987) and the Perceptual Assimilation Model (PAM) by Best (e.g. Best, 1995) posit that discrimination of second language phonemes is based on how they fit within the existing native language phoneme categories.

    View all citing articles on Scopus
    View full text