Presentation

Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams

Cerebellar role in lexical processing during continuous speech perception

Poster E46 in Poster Session E, Saturday, October 8, 3:15 - 5:00 pm EDT, Millennium Hall

Hannah Mechtenberg¹, Christopher Heffner^2,3, Emily Myers^1,2, Sara Guediche²; ¹Department of Psychological Sciences, University of Connecticut, ²Department of Speech, Language and Hearing Sciences, University of Connecticut, ³Department of Communicative Disorders and Sciences, University at Buffalo

INTRODUCTION In the last few decades, research into the role of the cerebellum has expanded beyond motor functions, with specific subregions of the cerebellum—namely lobule VI and Crus I—implicated in higher cognitive functions that include language processing. One proposed function of the cerebellum in spoken language processing is in generating predictions about upcoming words from contextual cues (e.g., semantic and/or sentence context; Lesage et al., 2016; Moberget & Ivry, 2019). Generating accurate lexical predictions during spoken word recognition depends on word properties, such as lexical frequency and phonological neighborhood density, that affect both lexical access and selection. While words with higher lexical frequency have been shown to facilitate lexical access (due to increased activation compared to low frequency words), lexical selection is thought to be impaired for words with higher phonological neighborhood densities, due to increased competition from words that are similar in sound properties. Although the cerebellum has been implicated in lexical processing more generally (e.g., Guediche et al., 2014; Lesage et al., 2016; Pleger & Timmann, 2018), remarkably little research to date has explored its sensitivities to these lexical properties while listening to natural, continuous speech. The current fMRI study is the first, to our knowledge, to independently examine the effects of lexical frequency and phonological neighborhood density on cerebellar activity during continuous speech perception. METHODS Seventy-nine participants passively listened to a 10-minute sample of an unaltered podcast during fMRI. We extracted the lexical frequency (SUBTLWFUS, Balota et al., 2007) and phonological neighborhood density (IPhOD, Vaden et al., 2009) for each content word. After implementing a standard preprocessing pipeline, subject-level analyses used an amplitude-modulated regressor, time-locked to the onset of each word, to model the lexical properties (i.e., lexical frequency and phonological neighborhood density) of the unfolding speech signal. The output allowed us to examine cerebellar activation scaled to our lexical regressors of interest. RESULTS and DISCUSSION At the group level, sensitivity to lexical frequency and phonological neighborhood density was localized to cerebellar regions Crus I and Crus II. We then used Diedrichsen et al.’s (2009) probabilistic atlas of the cerebellum to plot each participant’s activation map. Overall, cerebellar sensitivity to lexical frequency and phonological neighborhood patterned in opposite directions. Across participants, we observed a positive correlation with lexical frequency that patterned with an extensive lexical-semantic cortical network (Binder et al., 2009) including the anterior temporal lobe, inferior frontal gyrus, and the tempo-parietal junction. Conversely, there was a negative correlation with phonological neighborhood density, which tracked with cortical regions sensitive to acoustic-phonetic information (i.e., superior temporal gyrus). These findings contribute to accumulating evidence that the cerebellum, specifically Crus I, is sensitive to factors that affect word recognition. The cerebellum has been implicated in prediction error learning, across multiple perceptual domains. Given that lexical information can be used to generate predictions about the incoming speech signal, the findings of the current study are consistent with a similar role for the cerebellum in spoken language processing.

Topic Areas: Speech Perception, Meaning: Lexical Semantics