[A comparison study of two different segmentation methods to measure intelligibility in Mandarin]

Area of Science:

• Audiology
• Speech-Language Pathology
• Psychoacoustics

Background:

• Assessing hearing in noise is crucial for understanding speech intelligibility.
• The Mandarin Hearing in Noise Test (MHINT) is used to evaluate performance-intensity (PI) functions.
• Standardized test rules are needed for accurate sentence intelligibility scoring.

Purpose of the Study:

• To validate alternative methods for the MHINT to determine the PI function.
• To establish optimal test rules for scoring sentence intelligibility.
• To compare character and word segmentation approaches within the MHINT.

Main Methods:

• 30 normally hearing subjects participated in the study.
• Speech reception thresholds were measured using the MHINT with adaptive rules based on character or word segmentation.
• Sentence intelligibility was assessed at each threshold to evaluate PI function estimation accuracy.

Main Results:

• No significant differences were observed in reception thresholds (RTS) between character and word segmentation groups (P > 0.05).
• No significant differences were found in speech recognition scores (SRS) between segmentation methods (P > 0.05).
• Individual subject comparisons also showed no significant differences in RTS or SRS between segmentation types.

Conclusions:

• The MHINT, using either character or word segmentation, accurately estimates the PI function.
• The findings support the use of the MHINT for evaluating sentence intelligibility in various populations.
• This method is applicable to normally hearing individuals, those with hearing loss, and users of aided listening devices.