Abstract
Using the parcellation tool, our results showed that volumes of motor and cognitive lobules impact both motor and cognitive performance, and that functional performance and cerebellar volumes distinguishes MS fallers from non-fallers. Future studies should explore the potential of cerebellar imaging to predict falls in PwMS.
Introduction
Falls represent a significant clinical challenge for individuals living with Multiple Sclerosis (MS), often resulting in injury, reduced mobility, and a diminished quality of life. While clinical assessments like walking speed and balance tests are standard, they often lack the sensitivity to predict which patients are at the highest risk of falling. The cerebellum, acting as the brain's "air traffic control" for both motor coordination and cognitive processing, is a key area of interest in MS pathology. However, its structural complexity has historically made detailed analysis difficult. This study investigates whether high-resolution, automated mapping of cerebellar sub-regions can identify specific structural signatures that correlate with fall history and functional decline.
Methods
The study involved 60 participants: 31 people with MS (PwMS) and 29 age- and sex-matched healthy controls (HC). The PwMS group was further divided into "Fallers" (n=15), defined as having at least one fall in the previous month, and "Non-Fallers" (n=16).
The technical core of the research was the application of ACAPULCO (Automated Cerebellar Anatomical Parcellation Using U-Net with Locally Constrained Optimization). This deep learning algorithm utilizes a convolutional neural network to segment the cerebellum into 28 distinct regions, including various lobules and vermal sections. Participants also completed a comprehensive clinical battery, including the Timed 25-Foot Walk (T25FW), the Symbol Digit Modalities Test (SDMT) for cognitive processing speed, and the Six Spot Step Test (SSST) for coordination.
Results
Clinical and Functional Performance
Clinical data revealed significant differences between the groups. MS Fallers exhibited poorer performance across nearly all motor measures compared to both healthy controls and MS Non-Fallers.
As shown in Table 1, MS Fallers had significantly higher (worse) scores on the Timed Up and Go (TUG) test and slower walking velocities. Interestingly, MS Non-Fallers often performed similarly to healthy controls in several domains, highlighting that functional decline is not uniform across the MS population.
Structural "Faller Signature"
The most critical finding was that regional cerebellar volume loss, rather than global atrophy, distinguished fallers from non-fallers. While PwMS generally showed lower volumes in the corpus medullare and lobules I-III and V compared to controls, these differences were predominantly driven by the faller subgroup.
Table 4 details the statistical significance of these regional differences. MS Fallers showed significant volume loss in the corpus medullare (p = 0.027), lobule VI (p = 0.041), and vermis X (p = 0.017) when compared to MS Non-Fallers. The "Cognitive Lobules" composite score (comprising VI, VII, and Crus I-II) was also a strong differentiator (p = 0.007).
Functional Correlations
The study reinforced the integrated nature of cerebellar function. Cognitive performance (SDMT) was strongly correlated with Lobule VI volume (p < 0.001), while walking speed (T25FW) was significantly associated with Lobules I-III (p = 0.004). Notably, volume in Crus I—traditionally considered a cognitive region—was significantly associated with both walking speed and balance, suggesting that "cognitive" regions play a vital role in motor stability.
Discussion
These findings suggest that the "wiring" of the cerebellum, represented by the white matter core (corpus medullare), and specific lobules involved in both motor and cognitive processing are critical for maintaining stability in MS. The fact that these structural differences were detectable in a cohort with relatively low disability (mean EDSS of 3.8) suggests that cerebellar atrophy may serve as an early warning sign of functional decline.
The study's primary limitation is its cross-sectional and retrospective nature, relying on self-reported fall history. However, the high precision of the ACAPULCO tool provides a robust foundation for future longitudinal research.
Conclusion
Specific regional cerebellar volumes can distinguish MS fallers from non-fallers, even when global disability scores are relatively low. Future research should focus on longitudinal studies to determine if these volumetric measures can prospectively predict fall risk, potentially allowing for earlier clinical interventions such as targeted physical therapy.
Summary updated