Adaptive paradigms for identifying the neural substrates of phonological encoding in individuals with and without aphasia

Abstract

Phonological encoding is an important stage of speech production in which lexical forms are mapped onto articulatory plans. Previous research has shown that phonological encoding generally depends on brain regions in the left supramarginal gyrus and left ventral precentral gyrus. Localizing these brain regions in individual participants would be valuable for neuroplasticity research as well as presurgical language mapping. However, existing language mapping paradigms are not effective in activating these components of the language network. We previously developed a semantic matching task that successfully activates the well-attested frontotemporal components of the language network with good reliability and sensitivity. However, this task does not activate phonological brain regions because it does not rely on speech sound processing to perform. Therefore, we developed two paradigms that do load on phonological processing: a rhyme judgment task, in which participants decided whether or not two pseudowords rhymed, and a syllable counting task, in which they decided whether or not two pseudowords had the same number of syllables. We found that both paradigms effectively identified phonological encoding regions in individual participants, with the rhyme paradigm generating more lateralized patterns of activation. Having developed these paradigms, we then set out to investigate their effectiveness in identifying language regions in patients with post-stroke aphasia. We scanned 36 individuals with aphasia with fMRI as they performed the rhyme judgment and semantic matching tasks. Activation approximated the neurotypical language network where possible given the damaged regions. The rhyme paradigm successfully activated the precentral and supramarginal regions that the literature has shown to be implicated in phonological encoding. Repetition score, a behavioral measure that was highly correlated with perceptual ratings of the patients’ phonological ability, was correlated with activity in the dorsal attention network in these patients, suggesting that patients with aphasia performing a linguistic task are more reliant than neurotypical controls on cognitive processes nonspecific to language. This study constitutes the first robust evidence for domain-general networks contributing to language-specific processes in individuals with aphasia.