Low-frequency cortical responses to natural speech reflect probabilistic phonotactics

Area of Science:

• Neuroscience
• Psycholinguistics
• Auditory Processing

Background:

• Human speech comprehension is remarkably robust to noise and variations.
• The auditory system integrates sensory input with prior linguistic knowledge, including phonotactic rules.
• Phonotactic probabilities influence word likelihood in a language, but their neural processing in natural speech is not fully understood.

Purpose of the Study:

• To investigate the neural encoding and integration of phonotactic probabilities during naturalistic speech perception.
• To determine if phonotactic information influences brain responses to continuous speech.
• To explore the timing and nature of phonotactic processing in the brain.

Main Methods:

• Utilized non-invasive electroencephalography (EEG) to measure brain responses.
• Employed linear modeling to analyze the relationship between speech stimuli and EEG signals.
• Included phonotactic probabilities as a key descriptor of the speech stimuli.

Main Results:

• The inclusion of phonotactic probabilities significantly improved the model describing the speech-EEG relationship.
• Low-frequency cortical signals (<9 Hz) were found to reflect the integration of phonotactic information.
• Phonotactics-related neural signals showed significant interactions with speech between 100-500 ms post-stimulus.

Conclusions:

• Low-frequency brain activity integrates phonotactic information during natural speech perception.
• Phonotactic processing occurs at a pre-lexical stage, influencing early auditory responses.
• This study provides a method for measuring individual-level phonotactic processing using EEG.