Gestural coordination of Brazilian Portugese nasal vowels in CV syllables: A real-time MRI study

Alexsandro R. Meireles, Louis Goldstein, Reed Blaylock, Shrikanth S. NarayananICPhS 2015

Introduction

In linguistics, the Coupling Hypothesis of Syllable Structure (CHSS) suggests that the gestures making up a CV (Consonant-Vowel) syllable—such as the closing of the lips and the movement of the tongue—are triggered simultaneously (in-phase). However, nasal vowels (like the [pã] in the Portuguese word panto) present a mechanical challenge. If a speaker lowers the velum (soft palate) at the exact same time they close their lips for a stop consonant like [p], the air escapes through the nose, turning the [p] into an [m]. This study uses real-time MRI to investigate how speakers coordinate these conflicting gestures and whether the "in-phase" rule still holds for nasalized syllables.

Methods

The researchers employed real-time Magnetic Resonance Imaging (rt-MRI) to capture the vocal tract in motion at 33.18 frames per second. Two native speakers of Brazilian Portuguese provided the data, reading minimal pairs contrasting oral and nasal vowels in four syllable structures: CV (e.g., pato [pa]), CVN (e.g., panto [pã]), NV (e.g., mato [ma]), and NVN (e.g., manto [mã]).

Real-time MRI frame showing the Regions of Interest (ROIs) used for tracking articulatory gestures.

Using a MATLAB-based tool (MVIEW_RT), the team defined Regions of Interest (ROIs) for the lips, tongue tip, tongue body, and velum. They tracked pixel intensity changes in these areas to determine exactly when a gesture (like velum lowering or lip opening) began and ended. The dataset consisted of 12 target words embedded in carrier sentences, repeated 10 times each.

Key Results and Insights

The Velum "Wait" State

The study found that for almost all syllable types, the onset of the consonant and the onset of the vowel's lingual (tongue) gesture happened nearly simultaneously, supporting the CHSS. However, in CVN syllables (like [pã]), the velum does not drop at the start of the consonant. Instead, the velum lowering gesture is precisely timed to the release of the consonant. This specialized timing prevents the stop from sounding like a nasal consonant.

Quantitative Timing Analysis

Statistical analysis of the gestural intervals confirmed the distinct timing patterns required for nasal vowels.

Mean timing measures (in ms) across different syllable structures for two subjects.
  • ConsVons (Consonant onset to Vowel onset): Measured at 22–37 ms for Subject 1 and 15–34 ms for Subject 2. These low values indicate near-synchrony between the consonant and the vowel's tongue movement.
  • CrelNons (Consonant release to Velum onset): In CVN syllables, the interval between the consonant release and the velum dropping was only 34–44 ms, confirming that the "nose-opening" gesture is synced to the release of the stop.
  • VonsNons (Vowel onset to Velum onset): In CVN syllables, this interval was large (107–146 ms), proving that the vowel's tongue position and its nasality do not start at the same time in this specific context.

Conclusion

The fact that the velum gesture in [pã] is timed to the release rather than the onset makes it a "complex" coordination. This helps explain why nasal vowels are acquired relatively late in childhood—they require breaking the default "start everything at once" rule of simple syllables. While this high-resolution study provides a clear proof of concept, the researchers note that future work with larger sample sizes is needed to explore the effects of consonant voicing and broader prosodic emphasis on these tight coordination patterns.

Summary updated

BibTeX
@inproceedings{meireles2015gestural,
  author = {Alexsandro R. Meireles and Louis Goldstein and Reed Blaylock and Shrikanth S. Narayanan},
  title = {{Gestural coordination of Brazilian Portugese nasal vowels in CV syllables: A real-time MRI study}},
  booktitle = {ICPhS},
  year = {2015},
}
Figure