A study of velopharyngeal area using real-time magnetic resonance imaging with simultaneous speech recordings.

摘要

Along with rapid technical advancements, magnetic resonance imaging (MRI) has become popular in speech research providing information on vocal tract geometry and specific muscle configurations. MRI speech research, however, traditionally has suffered from lengthy data acquisition time and intense noise generated during MRI scans, and these challenges have been addressed in the current study by improving the effective image sampling rate and conducting simultaneous speech recordings during MRI scans. The main objective of the study was to examine the relationship between acoustic and physiologic aspects of the velopharyngeal mechanism by combining fast real-time MRI with simultaneous speech recordings and subsequent acoustic analyses.;Ten normal adult individuals participated in the study, five males and five females. All MRI-related procedures were performed using a 3-Tesla MRI scanner at the Biomedical Imaging Center, University of Illinois at Urbana-Champaign. Gender differences of the midsagittal palatal structures and levator veli palatini muscle configurations were examined based on the three-dimensional anatomic MRI data during a resting state. Males exhibited longer hard palate, wider distance between the points of the levator veli palatini muscle origin, and longer levator veli palatini muscle compared to females. Midsagittal dynamic MR images were reconstructed at an effective sampling rate of 30 images per second. A set of acoustic parameters of nasalization including the peak amplitude and bandwidth of the first resonant frequency (F1) at the rate of 30 sets per second was obtained from the speech data simultaneously collected during dynamic MRI scans. With the dynamic midsagittal MRI and acoustic data, the acoustic-physiologic relationships of the velopharyngeal mechanism in relation to tongue movements were investigated. Results indicated that each of the acoustic parameters of nasalization represented as the peak amplitude and bandwidth of the first resonant frequency (F1) was reliably accounted for by a group of multiple predictors including gender, vowel context, and six velar- and tongue-related physiologic variables. It was also found that the acoustic nasalization and midsagittal velopharyngeal port opening gestures were temporally aligned, while both were significantly influenced by the different vowel context. Greater degree of velopharyngeal port opening with lengthened period was observed with the low vowel context to achieve nasalization. The current study linked acoustic features of nasalization with velar- and tongue-related physiologic variables to account for the velopharyngeal mechanism using real speech, offering a better understanding of the normal human speech mechanism.