Mechanism for producing a neoglottal fricative [h] in tracheoesophageal speech

Abstract

Over the past 24 years, we have been performing the tracheoesophageal (TE) fistulization for voice restoration following a total laryngectomy. The principle of this technique is to divert the exhaled air through the TE fistula into the hypopharynx, where the thyropharyngeus muscle forms the retropharyngeal prominence on which the neoglottis is located. In TE speech, the electromyographic (EMG) activity of the thyropharyngeus muscle decreases for voiceless positive production. This neoglottic articulatory adjustment plays an important role in opening the neoglottis. It has been previously reported that the glottis is open for the glottal fricative ([h]) sound produced with laryngeal articulatory adjustment and that turbulence through the glottis produces the [h] sound in laryngeal speech. Of all the voiceless consonants in the Japanese language, [h] is the most difficult to produce in TE speech. This suggests that other adjustments in addition to the neoglottic articulatory adjustment are necessary to produce the [h] sound in TE speech. This study was designed to clarify the mechanism by which the neoglottal fricative ([h]) sound is produced. Eight speakers who could pronounce [h] were included in this investigation, a fiber optic examination of the neoglottis, aerodynamic study, and an EMG examination of the thyropharyngeus muscle during the production of words containing voiceless fricative [h] and [s] sounds were performed in each subject. The two groups were classified according to the subject's success or failure in producing words with initial [h] and medial [h] sounds. Out of 80 TE speakers who could not pronounce [h], 4 subjects were selected as the control group. Fibroptic examination revealed a transient neogottal opening during the production of [h] and [s] sounds. The median of the average opening time was 0.26 seconds for initial [h] sounds and 0.19 seconds for medial [h] sounds. No significant difference in opening time was observed for [h] and [s] sounds. The aerodynamic investigation demonstrated a transient increase in supraneoglottal pressure and airflow through the neoglottis and a transient decrease in subneoglottal pressure for [h] production, while no changes were observed in the control group. The EMG study demonstrated that the activity of the thyropharyngeus muscle decreases in cooperation with a transient decrease in subneoglottal pressure only for [h] production. As a result of the transient decrease in subneoglottal pressure, the EMG activity for [h] decreased more strongly than that for [s]. These findings indicate that both neoglottic adjustment and pulmonary adjustment are required for [h] production in TE speech.