Predictive language comprehension in Parkinson's disease

Abstract

Verb and action knowledge deficits are reported in persons with Parkinson's disease (PD), even in the absence of dementia or mild cognitive impairment. However, the impact of these deficits on combinatorial semantic processing is less well understood. Following on previous verb and action knowledge findings, we tested the hypothesis that PD impairs the ability to integrate event-based thematic fit information during online sentence processing. Specifically, we anticipated persons with PD with age-typical cognitive abilities would perform more poorly than healthy controls during a visual world paradigm task requiring participants to predict a target object constrained by the thematic fit of the agent-verb combination. Twenty-four PD and 24 healthy age-matched participants completed comprehensive neuropsychological assessments. We recorded participants' eye movements as they heard predictive sentences (The fisherman rocks the boat) alongside target, agent-related, verb-related, and unrelated images. We tested effects of group (PD/control) on gaze using growth curve models. There were no significant differences between PD and control participants, suggesting that PD participants successfully and rapidly use combinatory thematic fit information to predict upcoming language. Baseline sentences with no predictive information (e.g., Look at the drum) confirmed that groups showed equivalent sentence processing and eye movement patterns. Additionally, we conducted an exploratory analysis contrasting PD and controls' performance on low-motion-content versus high-motion-content verbs. This analysis revealed fewer predictive fixations in high-motion sentences only for healthy older adults. PD participants may adapt to their disease by relying on spared, non-action-simulation-based language processing mechanisms, although this conclusion is speculative, as the analyses of high- vs. low-motion items was highly limited by the study design. These findings provide novel evidence that individuals with PD match healthy adults in their ability to use verb meaning to predict upcoming nouns despite previous findings of verb semantic impairment in PD across a variety of tasks.

Fig 1

A schematic of the trial sequence, (A) for predictive sentences and (B) for baseline sentences. The left-most panel depicts the fixation cross, which participants had to fixate in order to launch the trial. The middle panel depicts the preview period when participants simultaneously previewed the four items. The right-most panel shows the screen layout when the sentence audio began. Images remained onscreen for the full duration of the sentence. Images were displayed in a balanced quadrant in the center of a white background, with ~1.5 inches between images horizontally, ~0.5 inches between images vertically, and ~1 inch between the edge of the image and the edge of the white portion of the screen. Stimuli images displayed in Fig 1 are similar but not identical to the original images and are therefore for illustrative purposes only. Pool net and strawberry reprinted from Brodeur, Dionne-Dostie [117] under a CC BY license, with permission from Mathieu Brodeur, original copyright 2009, 2010. Cradle [match_0188] and bathtub [object_0075] reprinted from Kovalenko, Chaumon [118] under a CC BY license, with permission from Niko Busch, original copyright 2012. Rope reprinted from Brady, Konkle [119] stimulus set; canoe reprinted from Konkle and Caramazza [120]; drum reprinted from Konkle and Oliva [121] under a CC BY license, with permission from Talia Konkle, original copyright 2008, 2011, 2013. Quilt original photo under a CC BY license, with permission from H.R. Templeton C.D. Hancok, original copyright 2021.

Fig 2. Gaze data by group in predictive sentences.

Binned binomial gaze probabilities to each area of interest, averaged across subjects and trials over the duration of the predictive sentences for controls (left) and PD participants (right). The x-axis reflects the elapsed time in the trial in milliseconds, offset by 200ms to account for saccade programming/launching time. The solid vertical lines mark the onset of the agent, verb, and target words. The dotted vertical line marks the end of the verb statistical analysis window (which extends only partially into the post-verb article). Error bars represent standard error of the mean.

Fig 3. Agent window model fit in predictive sentences.

Actual vs. model-predicted log odds (logits) of fixations to the target interest area in (A) the agent window, (B) the verb window, and (C) the target window, averaged across subjects and trials over the duration of each window for controls (top) and PD participants (bottom). The x-axis reflects the elapsed time in each window in milliseconds, offset by 200ms to account for saccade programming/launching time.

Fig 4. Gaze data by group and motion content in predictive sentences.

Binned binomial gaze probabilities to each area of interest, averaged across subjects and trials over the duration of the predictive sentences for: Controls x Low motion (upper left), Controls x High motion (lower left), PD x Low motion (upper right), PD x High motion (lower right). The x-axis reflects the elapsed time in the trial in milliseconds, offset by 200ms to account for saccade programming/launching time. The solid vertical lines mark the onset of the agent, verb, and target words. The dotted vertical line marks the end of the verb statistical analysis window (which extends only partially into the post-verb article). Error bars represent standard error of the mean.

Fig 5. Gaze data by group in baseline sentences.

Binned binomial gaze probabilities to each area of interest (AOI), averaged across subjects and trials over the duration of the non-predictive sentences for control participants (left) and PD participants (right). The x-axis reflects the elapsed time in the trial in milliseconds, offset by 200ms to account for saccade programming/launching time. The vertical line marks the onset of the target word. Error bars represent standard error of the mean.

Fig 6. Target window model fit in baseline sentences.

Actual vs. model-predicted log odds (logits) of fixations to the target interest area, averaged across subjects and trials over the duration of the target window for controls (left) and PD participants (right). The x-axis reflects the elapsed time in the target window in milliseconds, offset by 200ms to account for saccade programming/launching time.