The human cerebellum lies in the posterior cranial fossa and is composed of a core of white matter and a superficial cortex. The cerebellar cortex is characterized by a high degree of convolution, resulting in the formation of lobes and lobules. These lobes and lobules are then subdivided into smaller leaf-like structures called folia (Goldwyn, 1985). Accumulating evidence indicates progressive cerebellar atrophy with age (Luft et al., 1999; Hulst et al., 2015; Steele and Chakravarty, 2018; Han et al., 2020). Nevertheless, abnormal atrophy of the cerebellum can result in various neurological impairments, including movement coordination issues (Abdelgabar et al., 2019; Cortese et al., 2020), memory deterioration (Klockgether et al., 2019), and speech dysfunctions (Rusz et al., 2019).

The significance of the cerebellum has led to extensive investigations into its aging trajectory. Studies have confirmed a reduction in the volume of the vermis with age (Luft et al., 1999; Koppelmans et al., 2017). Similarly, the volume of bilateral Crus I decreases gradually during normal aging (Bernard and Seidler, 2013; Koppelmans et al., 2017). Longitudinal research by Raz et al. (2013) and Han et al. (2020) further revealed shrinkage in certain cerebellar subregional volumes with age. While the majority of these studies focused on aging trends in cerebellar subregional volumes, providing valuable insights into volumetric changes throughout the aging process, they offered limited information on morphological alterations within different cerebellar regions. Shape analysis, however, can be used to identify local alterations occurring through development or the aging process in brain regions (Shi et al., 2014; Ge et al., 2021). Consequently, this study aimed to evaluate the morphological changes in the cerebellum from different perspectives by integrating results from local volume measurements and shape analysis.

Previous research has shown that distinct regions within the cerebellum are involved in motor (Mottolese et al., 2013), cognitive (Guell et al., 2018), and emotional processes (Stoodley and Schmahmann, 2018; Schmahmann, 2019). Furthermore, neuroimaging studies have elucidated that particular sensorimotor tasks and advanced cognitive functions are associated with specific lobules within the cerebellum (Diamond, 2000). For instance, a functional neuroimaging study revealed activation related to motor tasks in lobules IV, V, VI, and VIII, whereas cognitive tasks elicited activation in lobules VI, VIIB, IX, and X, Crus I, and Crus II (Guell et al., 2018). The functional specificity of cerebellar regions was further corroborated by observations in patients with cerebellar lesions. Lesions in the anterior lobe of the cerebellum can result in characteristic cerebellar motor syndromes, such as gait ataxia, limb dyskinesia, and dysarthria (Schmahmann et al., 2012). Lesions in the posterior lobe of the cerebellum have been associated with the onset of cerebellar cognitive affective syndrome (CCAS), yet they do not significantly impact motor functions (van Harskamp et al., 2005; Stoodley et al., 2016). In neurodegenerative diseases such as Alzheimer’s disease and Parkinson’s disease, the cerebellum is a critically impacted region, with morphological alterations over time due to sustained damage (Iskusnykh et al., 2024). Consequently, examination of cerebellar regional changes in individuals who are cognitively normal during the aging process can provide valuable insights into the pathology of neurodegenerative disorders, such as Alzheimer’s disease and Parkinson’s diseases. Furthermore, a detailed exploration of cerebellar atrophy patterns inherent to normal aging processes is instrumental for guiding the development of precise therapeutic interventions.

To comprehensively understand cerebellar atrophy associated with aging, we utilized convolutional neural networks and shape analysis to investigate the morphological changes in the cerebellum in a cohort of 553 healthy individuals aged 20 to 80 years. These methodologies were complementary and collectively supported a detailed assessment of the structural changes in the cerebellum attributable to aging. By investigating cerebellar morphology from both global and regional perspectives, our study aimed to explore the morphological changes in the cerebellum during the aging process.

A detailed description of the morphological changes in each structure with age is given below. Given that the dataset did not include information on handedness, this factor was not accounted for in our study.

In our research, we utilized volumetric and morphological analyses to assess the progression of cerebellar atrophy with advancing age. We observed a marked reduction in the total absolute volume of the cerebellum, consistent with findings from previous studies (Luft et al., 1999; Sullivan et al., 2000; Raz et al., 2005; Romero et al., 2021). Notably, the effects of aging on various cerebellar subregions are heterogeneous (Han et al., 2020). Specifically, the results demonstrated a significant age-related decline in the absolute volumes of CM, AL, PL, lobule X, and vermis X (Figure 2). The results of shape analysis further demonstrated that atrophy in different regions of the cerebellum varies with age (Figure 3).

Our investigation revealed significant differences in the subregional absolute volume of the cerebellum. The cerebellar parcellation algorithm based on convolutional neural networks was effective in the segmentation of 28 cerebellar subregions. The analysis indicated a statistically significant reduction in the absolute volumes of 15 subregions with advancing age. The results of our investigation slightly diverge from those reported by Han et al. (2020), despite using the same segmentation method. We speculate that this discrepancy may be due to the wider age range of our sample, variations in sample characteristics and the use of different statistical methods. Our study revealed a significant age-related trend, specifically in the right lobules I-IV and V. A comparative study of voxel-based morphometry (VBM) in cerebellar patients and healthy individuals revealed that the pattern of cerebellar degeneration with age was similar, but not identical, to that of patients with cerebellar degenerative disease. The VBM results of the study indicated that the anterior cerebellar lobe (lobules I-V) exhibited the most pronounced cerebellar volume reduction in both patient datasets (Hulst et al., 2015). This study provides strong support for our results. The absolute volumes of bilateral Crus I and lobule VI showed a significant age-related declining trend, which was supported by the results of Ramanoel et al., who found substantial atrophy in Crus I among older adults using the SUIT toolbox (Ramanoel et al., 2018). In contrast to lobules I-VI, which primarily facilitate the coordination and fine regulation of bodily movements, lobule VI/Crus I predominantly participates in executing complex cognitive tasks, such as those associated with language processing and working memory (Guell et al., 2018). We hypothesize that morphological alterations in bilateral Crus I and lobule VI associated with aging will impact working memory functions, among others. The precise nature of the relationship between structural changes and functional outcomes necessitates further experimental exploration. In this study, bilateral Lobule VIIB demonstrate age-related decline and this decline is statistically significant. Lobule VII is a secondary sensorimotor structure (Debas et al., 2010). The left cerebellar lobule VIIB is involved in executive functioning, with impairments in this domain correlating with an elevated risk of pervasive psychopathology among children and adolescents (Buckner et al., 2011). Our study also revealed a significant reduction in the absolute volume of bilateral lobule IX with aging. A study of the cerebellum and cognition in Friedrich’s ataxia patients showed a direct correlation between the volume of lobule IX and impaired visual spatial function (Cocozza et al., 2020). Such findings may elucidate the reduction in visual spatial abilities observed in the elderly population. Lobule X mainly participates in maintaining balance, posture, and eye movements (Schniepp et al., 2017). The absolute volume of the right lobule X showed a significant age-related decrease. The impact of aging on the function of the right lobule X should be further investigated.

Several investigations have reported a decrease in the volume of the vermis with advancing age (Luft et al., 1999; Koppelmans et al., 2017). However, our study suggested that only the absolute volume reduction in vermis X exhibited a significant age-related decreasing trend. With respect to changes in the absolute volume of the vermis during the aging process, existing research does not provide a uniform conclusion (Bernard and Seidler, 2013; Bernard et al., 2015; Han et al., 2020). Variations in datasets, statistical methodologies, segmentation protocols, and analytical techniques may account for the discrepancies observed among the findings.

Our study revealed statistically significant sex-related variations in the absolute volumes of 16 out of the 28 segmented subregions in the cerebellum. This finding aligns with previous studies confirming cerebellar structural sexual dimorphism (Fan et al., 2010; Bernard et al., 2015). It is widely accepted that the brains of males tend to be larger than those of females within the same age group (Good et al., 2001; Luders and Kurth, 2020; Joel, 2021). Consequently, we analyzed an analysis of the normalized cerebellar volumes, as detailed in our Supplementary Materials. This analysis revealed significant sex-based differences in the relative volumes of only 4 out of the 28 cerebellar subregions examined. Contrary to our findings, Romero et al. did not detect statistically significant sex-based differences in the relative volumes of cerebellar subregions, despite noting significant differences in their absolute volumes (Romero et al., 2021). We postulate that the disparity between our study and Romero’s study is a result of our use of a more detailed segmentation approach and different sample sizes. Shape analysis results further suggested that while cerebellar atrophy with advancing age exhibits regional specificity in both sexes, the degree of atrophy varies between sexes, with male cerebellums showing marginally more pronounced atrophy than those of female (Figure 4).

Unlike previous studies (Fan et al., 2010; Bernard and Seidler, 2013; Bernard et al., 2015; Han et al., 2020), our research presents corroborative evidence of progressive cerebellar atrophy throughout the aging process obtained with two complementary approaches. While regional volume measurements provide a conventional quantitative description of cerebellar aging, shape analysis further characterizes cerebellar atrophy with age from an intuitive perspective, utilizing the thickness values of cerebellar surface vertices. This dual approach contributes to a closer understanding of the cerebellum ageing trajectories over time, with potential applications in neurodegenerative diseases.

Our study has several limitations. The smoothing steps inherent in shape analysis processing may obscure the visibility of cerebellar sulci on the surface, complicating the precise delineation of cerebellar subregions. Furthermore, neural network algorithms are predominantly used to assess the local volume of the cerebellum, in contrast to shape analysis, in which the cerebellum’s thickness, as defined by its vertices, is evaluated. Consequently, the correlation between these two methodologies is somewhat limited. Nonetheless, the application of shape analysis was intended to validate the progressive cerebellar atrophy associated with aging. Therefore, the integration of these methodologies is valuable. Given its cross-sectional design, this study inherently lacks the capacity to explore temporal changes at the individual level over time. Moreover, the diversity of sequence acquisition in publicly available datasets, coupled with the restricted clinical information on the subjects included and the absence of cerebrovascular quantification for the participants, could potentially influence the study’s findings, a prevalent challenge commonly encountered with the use of such datasets. Addressing such issues remains an essential future task.

In conclusion, in our research, cerebellar regional volume changes and the extent of cerebellar cortical atrophy across a normal population aged between 20 and 80 years were systematically examined. We identified region-specific atrophy within the cerebellum and found that it intensifies with age. Furthermore, our findings revealed that males experienced more pronounced atrophy in certain cerebellar regions than did females. This study provides comprehensive insights into cerebellar atrophy during normal aging and a structural basis for identifying potential therapeutic targets and devising strategies to address abnormal cerebellar atrophy.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.

The studies involving humans were approved by https://brain-development.org/ixi-dataset/. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

YW: Conceptualization, Data curation, Investigation, Methodology, Software, Writing – original draft, Writing – review & editing. YeT: Writing – review & editing. TL: Data curation, Investigation, Methodology, Software, Supervision, Writing – review & editing. YuT: Writing – review & editing. WL: Software, Writing – review & editing. WW: Investigation, Software, Writing – review & editing. ZL: Writing – review & editing, Supervision. QX: Writing – review & editing, Supervision. FX: Methodology, Software, Writing – original draft, Writing – review & editing. SL: Writing – original draft, Writing – review & editing, Supervision.