Speech pauses in production of action language in Parkinson’s disease: A potential marker of mild cognitive impairment (S51.006)

Abstract

<h3>Objective:</h3> To evaluate pauses before action utterances in Parkinson’s disease (PD) as a marker of cognitive function. <h3>Background:</h3> Speech is frequently impaired in PD, with longer pauses within and between utterances compared to controls. Action verb use is also impaired, but it is unclear if motor or cognitive dysfunction underlies this deficit. We hypothesized that production of utterances containing action verbs is more associated with cognitive than motor function, and may serve as a marker of mild cognitive impairment (MCI) in PD. <h3>Design/Methods:</h3> 93 participants with PD and 8 older controls were asked to describe the cookie theft picture. We identified action utterances (AU, utterances containing an action verb) and non-action utterances (nonAU). Pauses &gt; 2 sec. were measured using <i>Praat(v5.3.72)</i>. To control for the amount of speech produced, the total duration of pauses was divided by number of utterances. Wilcoxon rank-sum test was used to compare linguistic variables between PD and controls. In the PD group, Spearman’s correlations between linguistic variables, motor severity (MDS-UPDRS Part-III) and global cognitive function (MoCA) were calculated. We ran a logistic regression model with the dependent variable cognitive status (MCI or normal cognition), and predictors linguistic markers, controlled for age, sex, words per minute and UPRDS. <h3>Results:</h3> Compared to controls, PD participants had significantly longer pauses before AU (p&lt;0.04) but not nonAU. UPRDS was not correlated with linguistic markers but MoCA was inversely correlated with pauses before AU (rho= −0.34, p=0.0004) and before nonAU (rho=−0.38, p=0.0001). Pauses before AU (OR 30.6 (95% CI=1.81–519.50), p=0.018) but not before nonAU were predictive of PD-MCI. <h3>Conclusions:</h3> In PD, pausing before AU in spontaneous speech is associated with cognitive function independent of motor severity. This linguistic variable was sensitive to PD-MCI and could be a useful marker of early cognitive decline in PD. Future investigation should explore the cognitive basis of this action related language deficit. <b>Disclosure:</b> Dr. Nascimento Andrade has nothing to disclose. Dr. Smith has received personal compensation in the range of $0-$499 for serving as a Consultant for PureTech. The institution of Dr. Smith has received research support from NIH. Dr. Smith has received personal compensation in the range of $0-$499 for serving as a single time expert panel discussion contributor with Acadia. Ms. Manxhari has nothing to disclose.