In Spoken Word Recognition, the Future Predicts the Past.

Area of Science:

• Neuroscience
• Auditory Perception
• Speech Processing

Background:

• Speech perception relies on integrating contextual information to resolve ambiguity.
• The neural basis for integrating *subsequent* context in speech perception is largely unknown.
• Prior context influences speech perception, but how future information shapes interpretation is unclear.

Purpose of the Study:

• To investigate the neural mechanisms of integrating subsequent contextual information in speech perception.
• To determine how the brain uses later-arriving information to disambiguate early speech sounds.

Main Methods:

• Used magnetoencephalography (MEG) to record neural activity from the human auditory cortex.
• Analyzed brain responses to spoken words with ambiguous initial phonemes disambiguated later.
• Recruited 50 participants across two MEG experiments.

Main Results:

• Primary auditory cortex shows sensitivity to phonological ambiguity as early as 50 ms after sound onset.
• Subphonemic acoustic details are preserved in auditory cortex over extended periods.
• Phonological categorization occurs in parallel, resolving within approximately 450 ms.

Conclusions:

• Future speech input influences the perception of earlier sounds by maintaining sensory features.
• The auditory system preserves acoustic signals while forming hypotheses about word identity.
• This dual processing strategy enables rapid message comprehension and minimizes parsing errors.