Infratentorial white matter integrity as a potential biomarker for post-stroke aphasia

Traditionally, neuroimaging studies of post-stroke aphasia focus on supratentorial brain regions related to language function and recovery. However, stroke-induced lesions often distort these areas, posing a challenge for neuroimaging analyses aimed at identifying reliable biomarkers. This study seeks to explore alternative biomarkers in regions less affected by direct stroke damage, such as white matter regions below the tentorium, to overcome these methodological limitations. Diffusion tensor imaging was accomplished on 55 participants with chronic post-stroke aphasia. Focusing on regions below the tentorium, correlations were analysed between Western Aphasia Battery-Revised scores and average fractional anisotropy values. The volume of intersection between each participant's lesion and their left arcuate fasciculus was also analysed for correlations with Western Aphasia Battery-Revised scores as well. Linear regression analyses were then conducted using regions showing significant correlations as univariate predictors. After applying multiple comparisons corrections, we found that average fractional anisotropy in the middle cerebellar peduncle was positively correlated with aphasia quotient (P = 0.004), spontaneous speech (P = 0.005), auditory verbal comprehension (P = 0.004), naming and word finding (P = 0.005) and repetition (P = 0.013). Average fractional anisotropy in the left inferior cerebellar peduncle positively correlated with spontaneous speech (P = 0.018) and auditory verbal comprehension (P = 0.018). Average fractional anisotropy in the left corticospinal tract positively correlated with aphasia quotient (P = 0.019) and spontaneous speech (P = 0.005). The volume of intersection between the left arcuate fasciculus and participant lesion was negatively correlated with aphasia quotient (P = 0.014) and repetition (P = 0.002). Through linear regression analyses, average fractional anisotropy of the middle cerebellar peduncle significantly predicted aphasia quotient and all subscores. Average fractional anisotropy of the left inferior cerebellar peduncle significantly predicted all scores except repetition. Average fractional anisotropy of the left corticospinal tract significantly predicted all scores except for auditory verbal comprehension. The volume of intersection between the left arcuate fasciculus and lesions significantly predicted all scores except for auditory verbal comprehension. These findings underscore the potential of infratentorial white matter regions as biomarkers of aphasia severity, encompassing overall and specific subdomain impairment. By shifting the focus to below the tentorium, it becomes possible to find more robust targets for further research and therapeutic interventions. This approach is not only able to sidestep analytical complications posed by cortical lesions, it also opens new doors for understanding complex cerebellar mechanisms that underlie language function and recovery post-stroke.