Perceptual learning of multiple talkers requires additional exposure

Area of Science:

• Psycholinguistics
• Auditory Perception
• Speech Processing

Background:

• Listeners maintain distinct generative models for different talkers to interpret speech.
• Phonetic recalibration allows listeners to adapt to idiosyncratic speech sound productions.
• Previous research shows talker-specific recalibration, but simultaneous adaptation is less understood.

Purpose of the Study:

• To investigate the perceptual cost of simultaneously updating two distinct generative models compared to updating one at a time.
• To examine how listeners adapt to idiosyncratic speech productions when exposed to multiple talkers concurrently versus sequentially.
• To explore the cognitive mechanisms underlying talker-specific phonetic recalibration.

Main Methods:

• Listeners were exposed to two talkers with atypical /s/ and /∫/ productions.
• Experiment 1: Talkers were blocked by voice (sequential exposure).
• Experiment 2: Talkers were intermixed during learning (simultaneous exposure).
• Lexical information disambiguated phoneme identity in ambiguous contexts.

Main Results:

• Listeners recalibrated more quickly when talkers were presented sequentially (Experiment 1).
• Simultaneous exposure to two talkers required significantly more exposure trials for adaptation (Experiment 2).
• A perceptual cost exists when updating multiple generative models concurrently.

Conclusions:

• Simultaneously updating distinct generative models for multiple talkers incurs a perceptual cost.
• Listeners may need to select which generative model to update, slowing adaptation.
• This finding has implications for understanding real-world speech perception with multiple unfamiliar talkers.