Cerebello-Cerebral Pathways Contribute to Written Word Production

Abstract

Written language production is a fundamental aspect of daily communication, yet the neural pathways supporting it are far less studied than those for spoken language production. This study evaluated the contributions of speech-production pathways to written word production, specifically focusing on the central processes of word spelling rather than the motor production processes that support handwriting. Seventy-three English-speaking, neurotypical adults completed a spelling-to-dictation task and underwent diffusion MRI scans. The bilateral cerebello-thalamo-cortical pathways (CTC) and frontal aslant tract (FAT) were identified in individual participants using probabilistic tractography and automated segmentation tools. Fractional anisotropy (FA) values were computed along the trajectory of each tract and entered into correlation analyses with the spelling accuracy scores. A significant correlation was found between spelling accuracy scores and FA in the left CTC, which connects the left cerebellar hemisphere with the right cerebral hemisphere. This effect remained significant after controlling for spoken production measures. A similar trend was observed in the right homologous tract. In contrast, no significant correlations were identified between spelling accuracy scores and FA in the bilateral FAT. These findings demonstrate, for the first time, the involvement of cerebello-cerebral connections in spelling processes, aligning with the growing recognition regarding the role of the cerebellum in higher-order language functions. This effect did not generalize to the FAT, which may be relevant for more peripheral aspects of language production.

Keywords: cerebellum; diffusion MRI; spelling; superior cerebellar peduncle; white matter; written word production.

Figure 1.

The spelling-to-dictation task. (A) Behavioral paradigm: Participants typed each word after listening to it in isolation and in the context of a carrier sentence. The stimuli included 40 low-frequency English words, each 8–10 letters long, with one-to-many phoneme-to-grapheme mappings (adapted from Burt & Tate, 2002). (B) Scoring example: Each response was classified as either correct or incorrect. Spelling accuracy was calculated for each participant as the proportion of correct items out of the total number of trials.

Figure 2.

Tracts of interest. Bilateral tracts identified with probabilistic tractography are displayed in six representative participants, overlaid on each participant’s T1 image (P1: F, 20 yr; P2: M, 21 yr; P3: M, 21 yr; P4: F, 22 yr; P5: F, 18 yr; P6: F, 24 yr). CTC = cerebello-thalamo-cortical tract (left = purple, right = lilac), FAT = frontal aslant tract (left = blue, right = light blue), P = participant, F/M = female/male, yr = years of age, L = left, R = right.

Figure 3.

FA profiles along the tracts of interest. For each tract of interest, FA values are plotted along 30 equidistant nodes between two waypoint ROIs. Each colored line in the line-graphs represents an individual participant, and the thick black line indicates the mean FA-profile, averaged across participants. On the left of each plot, the relevant tract is visualized in a single participant (F, 18 yr), with dotted lines indicating the locations of the two ROIs. FA = fractional anisotropy, CTC = cerebello-thalamo-cortical tract, FAT = frontal aslant tract, L = left, R = right.

Figure 4.

Correlations between spelling accuracy and FA along the tracts. Spearman’s correlation coefficients are visualized in 30 equidistant nodes along the tract cores. An arrow indicates the location of the cluster (nodes 3–9) that shows a significant correlation in the left CTC (p < 0.05, family-wise error corrected for 30 nodes). FA = fractional anisotropy, CTC = cerebello-thalamo-cortical tract, FAT = frontal aslant tract, L = left, R = right.

Figure 5.

Correlations with spelling accuracy in the bilateral CTC tracts. Scatter plots depict the association between spelling accuracy and mean FA within the cluster of nodes showing a significant correlation in the left CTC (left scatter plot) and its homologous location in the right CTC (right scatter plot). Gray lines represent the best linear fit, enclosed by the 95% confidence interval (dashed lines). These scatter plots are for visualization purposes (significance and FWE correction were calculated along the tract; see Materials and Methods). The cluster locations are marked on the tractograms with framed shaded regions. CTC = cerebello-thalamo-cortical tract, FA = fractional anisotropy, L = left, R = right.