The lesion correlates of impaired content word fluency during spoken discourse in aphasia.

Human conversations are strongly dependent on the production of informative and fluent connected speech. In people with aphasia, this skill is affected to varying degrees, which can negatively impact social communication and quality of life. However, available coding systems to assess connected speech production based on discourse tasks are labour-intensive and very time-consuming, limiting their utilization in clinical and research contexts. In this study, we investigated and further validated a recently developed, accurate, time-efficient and clinically applicable measure of content word fluency (CWF) during spoken discourse in a large, unselected sample of 76 participants with chronic aphasia following left hemisphere stroke. We report the first identification of lesion correlates of impaired content word production in spoken discourse using state-of-the-art lesion-symptom mapping methods, including voxel-wise disconnection, multivariate lesion-symptom mapping and tract-wise analyses. Discourse responses elicited using composite picture description were analysed using (i) CWF to assess content word production in spoken discourse using a pre-specified checklist without transcription or quantitative analysis and (ii) 'correct information unit' (CIU) following the standard transcription and quantitative analysis protocol of discourse samples. We showed a significant, strong positive correlation between CWF scores and the number of CIUs. Item Response Theory analysis revealed that the one-parameter logistic model best fits the data, indicating that items on the CWF checklist are homogeneously measuring a single underlying construct. Moreover, the items on the checklist were found to have distributed difficulties in the sample over a large functional range, indicating that the CWF approach is sensitive to variations in performance across a broad spectrum of aphasia severity. The neuroimaging findings indicated overlapping lesion correlates between CWF and CIU in the left frontal and parietal regions, and anterior dorsal white matter pathways, specifically the middle frontal gyrus, inferior parietal lobe, frontal aslant tract and superior longitudinal fasciculus. These results reveal strong convergence between CWF and CIU, and they provide behavioural, neurological and psychometric validation of the CWF approach, an efficient tool for assessing communication deficits in people with aphasia in both clinical and research settings. These insights have potential clinical implications, from improving targeted rehabilitation strategies to predicting recovery outcomes.