Learning effect observed for the speech reception threshold in interrupted noise with normal hearing listeners

Area of Science:

• Audiology
• Psychoacoustics
• Speech processing

Background:

• Speech reception threshold (SRT) is traditionally measured using stationary noise.
• Non-stationary masking noises, such as interrupted noise, are increasingly used.
• Previous research suggested potential learning effects in SRT with interrupted noise.

Purpose of the Study:

• To investigate if stable speech reception thresholds (SRTs) are reached with repeated measurements in stationary versus interrupted noise.
• To determine the presence and extent of a learning effect in SRT measurements using different noise types.
• To assess SRT stability for both female and male speech materials.

Main Methods:

• Listeners with normal hearing completed five repeated SRT measurements.
• Measurements were conducted using either stationary speech-shaped noise or 8-Hz interrupted noise.
• SRTs were determined for both female and male speech materials.

Main Results:

• A significant improvement in SRT was observed with repeated measurements in interrupted noise.
• SRTs in interrupted noise improved by 3-4 dB after five replications.
• Stable SRT thresholds in interrupted noise were reached after approximately two replications.
• No significant improvement was found for SRTs measured in stationary noise.

Conclusions:

• A substantial learning effect exists for SRT measurements in interrupted (non-stationary) noise.
• Stationary noise does not appear to induce a significant learning effect for SRT.
• A repeated measures design is recommended when using non-stationary noise for SRT assessment.