Phonemic versus allophonic status modulates early brain responses to language sounds: an MEG/ERF study

Research output: Contribution to conferencePosterResearchpeer-review

232 Downloads (Pure)


Objective: An early component of the auditory event-related potential (ERP), the mismatch negativity (MMN/MMNm), has been shown to be sensitive to native versus non-native language sounds (Brandmeyer et al., 2012; Kazanina et al., 2006; Näätänen et al., 1997); i.e. sensitive to phonemic versus allophonic sound contrasts. So far this has only been attested between languages. In the present study we wished to investigate this effect within the same language: Does the same sound contrast that is phonemic in one environment, but allophonic in another, elicit different MMNm responses in native listeners?
For this purpose we employed the /t/-/d/-contrast in Danish (heavily aspirated versus unvoiced; only realized syllable-initially, not in syllable coda (Grønnum 2005)). Thus, [tæ] and [dæ] are separate words in Danish (‘take’ and ‘then’, respectively), whereas [æt] and [æd] are not (both meaning ‘that’).
This allowed us to manipulate the phonemic/allophonic status of exactly the same sound contrast (/t/-/d/) by presenting it in different immediate phonetic contexts (preceding a vowel (CV) versus following a vowel (VC)), in order to investigate the auditory event-related fields of native Danish listeners to a sound contrast that is both phonemic and allophonic within Danish.
Methods: Relevant syllables were recorded by a male native Danish speaker. The stimuli were then created by cross-splicing the sounds so that the same vowel [æ] was used for all syllables, and the same [t] and [d] were used for the relevant syllables. The final stimuli were validated by two native Danish listeners, ignorant to the design of the experiment. MEG was recorded from 17 native Danish listeners (5 females) while watching a silent movie and listening to the auditory stimuli. [tæ] and [æt] acted as standards, and [dæ] and [æd] thus as deviants, respectively. Data were preprocessed using Elekta’s MaxFilter software and SPM8, all statistical analyses were conducted in sensor-space using SPM8.
Results: Focusing on the 150-300 ms time period after stimulus onset (typical MMNm time range for language sound contrasts), only the phonemic [tæ]-[dæ]-contrast showed significant effects (FWE-corrected at the cluster-level). Comparing the differences of the two contrasts, the phonemic [tæ]-[dæ]-contrast showed a significantly larger difference around 250 ms after stimulus onset than the allophonic [æt]-[æd]-contrast (FWE-corrected at the cluster-level).
Conclusion: By manipulating the immediate phonetic context, we demonstrate that native Danish listeners’ early brain responses to exactly the same language sound contrast are modulated by its phonemic versus allophonic status.
Original languageEnglish
Publication date10 Nov 2013
Number of pages1
Publication statusPublished - 10 Nov 2013
EventNeuroscience 2013 - SFN annual meeting - San Diego, United States
Duration: 9 Nov 201313 Nov 2013


ConferenceNeuroscience 2013 - SFN annual meeting
Country/TerritoryUnited States
CitySan Diego

Cite this