Single-Word Processing and Connected Speech Production in Aphasia: Integrating Neuropsychological, Psycholinguistic and Neuroimaging Approaches

Abstract

Functional communication relies heavily on connected speech production, which depends mostly on word retrieval. A common consequence of brain damage or neurodegenerative disorders is acquired language impairments (i.e., aphasia), which affects a complex array of cognitive domains, including connected speech production and more specific features of language processing. However, the precise nature of some of these deficits remains unclear. Specifically, there is no converging evidence on how these deficits are affected at different levels of language production (i.e., single-word versus connected speech), or how word properties (i.e., noun/verb differences and psycholinguistic features) impact language processing. In this thesis, I addressed these issues through six empirical studies based on 48 chronic left hemisphere post-stroke aphasia patients. The patients' performance was compared directly to an age/education matched neuro-typical group, to validate novel tests and analysis protocols. Initially, a verb semantic battery was developed, which showed that the psycholinguistic features of the word, rather than word class per se, influenced single-word processing. This battery serves as a useful addition to existing clinical tests, and can inform sensitive assessments and intervention planning. Neuroimaging analyses showed that single-word processing is mainly supported by left anterior and posterior temporal regions, whereas additional left frontal regions are engaged by increased semantic complexity. Post-stroke aphasia was also explored in terms of a graded disorder along multiple fundamental dimensions. This approach showed that performance across noun-verb processing tests was not dichotomous but reflected graded differences in behavioural performance and within the cortex. Conversely, connected speech production was associated with additional cognitive demands and extra brain regions, which can be split into three components: verbal fluency supported by left frontal regions; verbal quality related to attention, self-monitoring, and executive control processes underpinned by left frontal and parietal regions; and speech rate, which relied on the left motor cortex and sensory auditory regions presumably for online feedback. Generally, seven fundamental components of aphasia were identified in a unified model: phonological production, semantic processing, phonological recognition, executive functions, as well as the three connected speech components. In a systematic fashion, a lack of direct map between the effect of psycholinguistic properties and these components was observed. The evidence further revealed weak/moderate correlations between word retrieval at single-word level and connected speech, suggesting that single-word tasks may not provide a sufficient reflection of connected speech output. In contrast, a strong correlation was observed between metrics from different connected speech tasks, and strong discourse genre effects were found on quantity and diversity metrics for neuro-typical adults and people with aphasia. Storytelling narrative outperformed the other discourse genres across many contrasts, albeit at a cost of time. Therefore, a novel index was proposed and validated that allows for a quick yet efficient measure of content word production during connected speech. Further evidence indicated that expressive deficits in aphasia mainly reflect quantitative (fewer words) rather than qualitative (complex words) degradation. Overall, the current thesis provides converging neuropsychological, psycholinguistic and neuroimaging insights into the nature of language deficits post-stroke, as well as evidence on the organisation of language functions, and proposes that language is modulated by independent but interactive components supported by a wide range of left brain regions.

Date of Award	1 Aug 2019
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Matthew Lambon Ralph (Supervisor), Paul Conroy (Supervisor) & Ajay Halai (Supervisor)