The Concurrent OLSA Test: A Method for Speech Recognition in Multi-talker Situations at Fixed SNR

Area of Science:

• Auditory neuroscience
• Speech perception research
• Cognitive psychology

Background:

• Assessing speech recognition in noisy environments is challenging.
• Existing methods may not accurately reflect real-world listening conditions.
• High signal-to-noise ratios (SNR) are crucial for precise auditory measurements.

Purpose of the Study:

• To introduce a novel multi-talker paradigm for speech recognition testing.
• To investigate the impact of attentional processes on speech-recognition scores.
• To enable auditory measurements at high signal-to-noise ratios (SNR).

Main Methods:

• Simulated a three-talker conversation using the German matrix test OLSA.
• Implemented an adjustable overlap time for talker alternation.
• Participants performed a dual-task: detecting a call sign ('Kerstin') and repeating target words.

Main Results:

• Tested with 22 young normal-hearing participants (YNH) across three overlap times (0.6s, 0.8s, 1.0s).
• Median target word recognition scores were 88%, 82%, and 77% for the respective overlap times.
• Significant differences in scores were observed, indicating an effect of overlap duration.

Conclusions:

• The multi-talker paradigm effectively simulates complex listening situations.
• Decreasing overlap time between talkers improves speech recognition performance.
• The results suggest that the paradigm induces a significant cognitive load, impacting performance.