Neural organization of speech production: A lesion-based study of error patterns in connected speech

Brielle C Stark; Alexandra Basilakos; Gregory Hickok; Chris Rorden; Leonardo Bonilha; Julius Fridriksson

Our results extend prior behavioral research which described a divergence between paraphasias made during connected speech and naming (Nicholas et al., 1989; Mayer and Murray, 2003; Fergadiotis and Wright, 2015) and our VLSM results establish a foundation for continued brain research comparing connected speech with naming. While we found that underlying brain structures are shared between paraphasias across tasks, at least as measured by lesion analysis, our behavioral analysis reflects subtle changes in linguistic load across tasks. While logically, a paraphasia should arise from damage to a brain area associated with the category of the paraphasia (i.e. left hemisphere anterior temporal lobe and semantically related errors), and therefore brain damage associated with paraphasias in connected speech and naming should be similar, there is a body of literature in neurotypical adults and adults with neurogenic communication disorders that connected speech involves the interaction of brain areas important for cognitive processes not usually found in single-word retrieval, such as inhibition and working memory (Coelho et al., 1995; Wright et al., 2014; Schnur, 2017) . The multiword environment present during connected speech is prone to many processes not at play during single-word retrieval, including lexical bias (Nozari and Dell, 2009) , phonemic similarity (Dell, 1984) , and perhaps even semantic interference (Schnur et al., 2006; Oppenheim et al., 2010) . We suggest and have first shown here that studying the neural substrates of paraphasias in connected speech share brain areas also associated with paraphasias of naming, but we further suggest that studying paraphasias in connected speech extends our understanding of the neurobiology of language. In the present study, we employed lesion-symptom methods largely focused on identifying the convergence and divergence of brain damage associated with paraphasias in both connected . CC-BY-NC-ND 4.0 International license is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/544841 doi: bioRxiv preprint 22 speech and naming, and we cannot fully elaborate on the recruitment of other linguistic or extra-linguistic areas to paraphasias in connected speech. Future work evaluating functional and structural connectivity may elaborate on the relative contribution of extra-linguistic cognitive processes to paraphasias produced during connected speech. In the current study, we analyzed a single naturalistic measure of connected speech, picture description, but we acknowledge that other forms of connected speech may be more naturalistic, such as story retelling and procedural descriptions, and encourage future work to evaluate these tasks in order to more comprehensively evaluate the spontaneous language system.