Neural representations used by brain regions underlying speech production
OA Version
Citation
Abstract
Speech utterances are phoneme sequences but may not always be represented as such in the brain. For instance, electropalatography evidence indicates that as speaking rate increases, gestures within syllables are manipulated separately but those within consonant clusters act as one motor unit. Moreover, speech error data suggest that a syllable's phonological content is, at some stage, represented separately from its syllabic frame structure. These observations indicate that speech is neurally represented in multiple forms. This dissertation describes three studies exploring representations of speech used in different brain regions to produce speech.
The first study investigated the motor units used to learn novel speech sequences. Subjects learned to produce a set of sequences with illegal consonant clusters (e.g. GVAZF) faster and more accurately than a similar novel set. Subjects then produced novel sequences that retained varying phonemic subsequences of previously learned sequences. Novel sequences were performed as quickly and accurately as learned sequences if they contained no novel consonant clusters, regardless of other phonemic content, implicating consonant clusters as important speech motor representations.
The second study investigated the neural correlates of speech motor sequence learning. Functional magnetic resonance imaging (fMRI) revealed increased activity during novel sequence productions in brain regions traditionally associated with non-speech motor sequence learning - including the basal ganglia and premotor cortex - as well as regions associated with learning and updating speech motor representations based on sensory input - including the bilateral frontal operculum and left posterior superior temporal sulcus (pSTs). Behavioral learning measures correlated with increased response for novel sequences in the frontal operculum and with white matter integrity under the pSTs, implicating functional and structural connectivity of these regions in learning success.
Description
Thesis (Ph.D.)--Boston University