Intersections between semantic cognition and affective processing: Insights from neuropsychology and neuroimaging

Semantic cognition, the basis for our understanding of the world, is supported by both the storage of semantic representations and the ability to flexibly retrieve them – semantic control. Semantic control dictates our ability to parse ambiguous information, and is supported by a distributed semantic control network (SCN). Patients with semantic aphasia (SA) experience multimodal semantic control deficits following left hemisphere stroke. Using cognitive neuropsychology and neuroimaging, this thesis further explores the neural bases of semantic retrieval and tests the relevance of semantic control to emotion, valence, and reward. Key findings were: (i) SA patients experience diffuse disconnection beyond lesion site. Lesion and structural disconnection of left-lateralised SCN nodes predict semantic impairment, while domain-general control impairment is predicted by lesion to adjacent fronto-parietal regions and by interhemispheric structural disconnection. (ii) SA patients present with impaired categorisation of facial emotion portrayals, susceptible to effects of cues and miscues. (iii) Valence congruency between words facilitates semantic matching, while semantic relatedness facilitates valence matching. SA patients present with impaired valence matching, exacerbated by semantic distractors. (iv) Impairments in the retrieval of weak associations can be facilitated in SA using cued extrinsic rewards. (v) The retrieval of both contextual and emotional associations is associated with activation in SCN, with these tasks being dissociated by default mode subnetworks. Effects of retrieval demands on the default mode network run orthogonally to effects of task. This thesis provides novel insight into the complex neural basis of retrieving meaning and emotion. It also contributes to our understanding of the impairments in SA, as well as how they can be ameliorated. Finally, this thesis supplements conceptually-focused models of emotion processing, stressing a role of semantic control. This work contributes to an increasingly clear understanding of controlled semantic retrieval and its contribution to cognition across domains.