Time-expanded speech and speech recognition in older adults

Area of Science:

• Audiology
• Cognitive Neuroscience
• Speech Processing

Background:

• Speech understanding deficits are prevalent in older adults, influenced by hearing sensitivity and cognitive changes.
• Processing speed, a cognitive function, is crucial for following rapid speech signals.
• Naturally slowed speech improves recognition, prompting interest in acoustic characteristics for time-expansion algorithms.

Purpose of the Study:

• To investigate the effects of time-expanded speech and consonant enhancement on speech recognition in older adults.
• To compare the performance of younger normally hearing, older normally hearing, and older hearing-impaired listeners.
• To evaluate the efficacy of specific signal processing techniques for improving speech recognition in aging populations.

Main Methods:

• Participants included younger normally hearing, older normally hearing, and older hearing-impaired listeners.
• Speech stimuli were processed using time expansion (increased duration) and consonant gain (increased intensity of unvoiced consonants).
• Listener performance was assessed on both unprocessed and processed speech in quiet and noise conditions.

Main Results:

• All listener groups performed best on unprocessed speech.
• In noise, consonant gain without time expansion improved performance with processed speech.
• In quiet, the slowest time-expanded rate yielded better performance with processed speech.

Conclusions:

• Signal processing techniques, including time expansion and consonant enhancement, show potential for improving speech recognition in older adults.
• The effectiveness of these techniques is condition-dependent (quiet vs. noise).
• Further research into acoustic characteristics of slowed speech can inform the development of assistive listening technologies.