Summary: A new study comparing stroke survivors and healthy older adults shows that language difficulties after stroke arise not from slower hearing but from weaker neural integration of speech sounds. Although patients detected sounds at the same speed as controls, their brains encoded speech features far less strongly—especially when words were ambiguous or difficult to hear.
Healthy listeners extended neural processing when speech was uncertain, helping them resolve ambiguous words. In contrast, people with post-stroke language impairments tended to stop processing phonetic information earlier, which may prevent successful word recognition. These findings identify specific brain activity patterns that support verbal comprehension and suggest faster, story-based diagnostic approaches for aphasia and related language disorders.
Key Facts
- Weakened phonetic integration: Stroke survivors show much weaker neural encoding of speech sound features, even though basic sound detection speed remains intact.
- Shortened persistence under uncertainty: When word identity is unclear, people without stroke maintain phonetic encoding longer than those with post-stroke language deficits.
- Diagnostic promise: Natural story-listening EEG recordings could become a faster, more natural alternative to lengthy behavioral testing for language disorders such as aphasia.
Source: SfN
Background: After a stroke, some individuals develop aphasia, a language disorder that interferes with processing speech sounds and understanding spoken words. To examine how stroke changes the brain’s handling of speech, researchers led by Laura Gwilliams (Stanford University) and Maaike Vandermosten (KU Leuven) compared neural responses in people with post-stroke aphasia and healthy, age-matched controls.
The team recorded EEG while participants listened to about 25 minutes of natural story material. The study included 39 people with post-stroke aphasia and 24 healthy older adults. Researchers analyzed the timing, strength, and sensor locations of phoneme and phonetic-feature encoding in the EEG data to reveal how speech is represented across time and brain regions.
The recordings revealed three main outcomes. First, healthy older adults encoded phonetic features reliably in their EEG responses, confirming that phoneme-level information is preserved in normal aging. Second, people with aphasia showed a marked reduction in the strength of phonetic-feature encoding after an initial shared processing window (roughly 80–250 ms after phoneme onset). This reduction was particularly pronounced over left-hemisphere electrode sites, consistent with the left-lateralized language network. Third, although the temporal speed of neural pattern evolution was similar across groups, only healthy listeners prolonged phonetic encoding when word identity was uncertain. People with aphasia did not show this sustained encoding under ambiguity, suggesting a failure to maintain low-level speech information long enough to resolve lexical identity.
Taken together, these results indicate that aphasia after stroke may not primarily reflect slower auditory detection but rather a diminished and short-lived neural representation of phonetic detail. In everyday listening—where words are often unclear or embedded in noisy or rapid speech—this deficit in maintaining phonetic information could lead to missed or incorrect word recognition and therefore impaired comprehension.
First author Jill Kries highlights the translational potential: passive story-listening paradigms combined with EEG may offer a quick, naturalistic diagnostic tool that captures the neural signatures of speech processing, reducing the reliance on hours of behavioral testing currently used to assess language disorders.
Key Questions Answered
Q: Why do some stroke survivors struggle to understand spoken language?
A: Their auditory systems detect sounds normally, but the brain’s encoding of phonetic features is weaker, which reduces the information available to identify words.
Q: What happens when spoken words are unclear?
A: Healthy listeners extend phonetic encoding to resolve ambiguity, whereas people with aphasia tend to stop encoding too early and therefore miss the cues needed to recognize difficult words.
Q: How could this research change diagnostics for language disorders?
A: Recording brain responses while people listen to stories could become a fast, ecologically valid method to identify language-processing impairments and track recovery or treatment effects.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The associated journal paper was reviewed in full.
- Additional context and clarifications were added by staff writers.
About this stroke and speech processing research news
Author: SfN Media
Source: SfN
Contact: SfN Media
Image credit: Neuroscience News
Original Research: Closed access. “The Spatio-Temporal Dynamics of Phoneme Encoding in Aging and Aphasia” by Laura Gwilliams et al., Journal of Neuroscience. DOI: 10.1523/JNEUROSCI.1001-25.2025
Abstract
The Spatio-Temporal Dynamics of Phoneme Encoding in Aging and Aphasia
Successful language comprehension depends on the brain’s ability to encode speech sounds (phonemes) through evolving neural patterns. This study tested whether those dynamics are altered in people with post-stroke aphasia. EEG responses were recorded from 39 individuals with aphasia and 24 healthy age-matched controls during 25 minutes of natural story listening. Analyses measured the duration and strength of phonetic-feature encoding, the speed of neural pattern evolution, and the spatial distribution across EEG sensors. Results show robust phonetic encoding in healthy older adults, a significant decrease in encoding strength in the aphasia group after an initial shared processing window (0.08–0.25 s), and left-lateralized reductions in the aphasic participants. Crucially, healthy controls prolonged phonetic encoding when word identity was uncertain, a mechanism absent in those with aphasia. These findings support the idea that aphasia involves failure to retain low-level speech information long enough to identify lexical items, offering targets for diagnostics and rehabilitation focused on sustaining phonetic representations during natural listening.