You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster A6, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

Fluent Speech in the Presence of Severe Verbal Working Memory Dysfunction

Christopher Barkley1, Zhenhong Hi2, Angela Birnbaum1, Ilo Leppik1, Susan Marino1;1University of Minnesota, 2University of Florida

It has been well established that verbal working memory (VWM) processes play a central role in language production (Daneman, 1991; Acheson, 2009), though the nature of the VWM mechanisms necessary for production has yet to be fully characterized. Here we describe the results of a clinical protocol that utilizes the tools of pharmacology to generate findings that constrain accounts of the interactions between VWM and production processes. Topiramate (TPM), a broad-spectrum anti-seizure drug used to treat conditions ranging from epilepsy to migraine, frequently causes cognitive deficits, the most common being severe word-finding difficulties (Mula et al., 2003). Lorazepam (LZP), a benzodiazapene used to treat anxiety and sleep disorders, is also associated with cognitive impairments but, in contrast to TPM, there is no evidence in the literature that LZP disrupts language. Due to the tight link between VWM and language production, we predicted that subjects on TPM would show VWM impairments. As LZP appears to spare the language system, we predicted that performance on a VWM task would remain intact after taking LZP. Twenty-nine healthy subjects received either 100, 150, or 200 mgs of TPM, 2mg of LZP, and PBO in a randomized, double-blind, crossover study design. Four hours after drug administration, they completed a Sternberg VWM task (three memory loads: 1, 3, and 5 syllables). Reaction time and accuracy were recorded, and the extent of drug-induced impairment was calculated by comparing each treatment to PBO ((drug-PBO)/PBO), thus normalizing across individual accuracy rates. Working memory capacity (WMC) was assessed by using data collected during the PBO session to calculate Cowan’s k scores, averaged across memory loads (Cowan, 2001). Subjects also completed a neuropsychological battery at 3 time points during each session. The battery included multiple tests of executive function, as well as the timed phonemic and semantic controlled oral word-association tasks (COWA) to assess verbal fluency. Compared to PBO, TPM caused pronounced deficits on all neuropsychological tests, including both COWA tasks (all p < .05). In contrast, while LZP affected all non-language tasks (all p < .05), performance on the COWAs did not differ from PBO. On the VWM task, TPM administration led to decreases in accuracy for all three memory loads (all p<.005). The magnitude of these accuracy decreases was correlated with subjects WMC (r=-.49, p=.011). Intriguingly, performance on the VWM task during the LZP session declined even more severely. Accuracy again decreased for all loads (all p<.005), but reaction times also increased significantly compared to both PBO and TPM (all p<.005). However, the severity of these performance deficits was not correlated with WMC (r=-.18,p=.33), suggesting that LZP impaired a VWM process distinct from the one crucial to production that was negatively impacted by TPM. These data suggest that individuals can produce normal language in the face of at least some severe disruptions of the VWM system, a finding not predicted under most descriptions of the architecture of the language production system. This underscores the utility of using pharmacological probes to understand the structure of the language system.

Topic Area: Control, Selection, and Executive Processes

Back to Poster Schedule