A comparison study of two different segmentation methods to measure intelligibility in Mandarin / 中华耳鼻咽喉头颈外科杂志

ABSTRACT

<p><b>OBJECTIVE</b>To confirm alternative methods for Mandarin hearing in noise test (MHINT) to get the performance-intensity (PI) function, and to set up right test rules for scoring sentence intelligibility for subjects.</p><p><b>METHODS</b>Sentence speech reception thresholds were obtained for 30 normally hearing subjects with the MHINT test using several adaptive rules based on either character or word segmentation of the sentences. Each adaptive rule was intended to measure the threshold at a different point on the PI function. By also measuring sentence intelligibility at each threshold, the accuracy with which the PI function was estimated could be evaluated for each type of segmentation. The results of different segmentation were compared.</p><p><b>RESULTS</b>There was no significant difference in the reception threshold of sentences (RTS) between the group which used character segmentation and the group which used word segmentation (P > 0.05); and also no significant difference in their speech recognition score (SRS, P > 0.05). When the same subject used both character segmentation and word segmentation to do the test, there was no significant difference in their RTS between character segmentation and word segmentation (P > 0.05); and also no significant difference in their SRS (P > 0.05).</p><p><b>CONCLUSIONS</b>This method could be used not only with normally hearing individuals and people with hearing loss to evaluate their sentence intelligibility, but also with people who had aided equipment.</p>