Mapping the human cerebellum

Collaborators: Joern Diedrichsen; Carlos Hernandez (Western University) and Richard Ivry (UC Berkeley)

In the media: Berkeley Press Release; Psychology Today; Knowable Magazine

The aim of this project was to build a novel atlas of the human cerebellum by characterizing how subregions of this “little brain” represent diverse processes from language and attention to social cognition. Leveraging large amounts of functional magnetic resonance imaging (fMRI) data and machine learning techniques, we built a comprehensive, first-of-its-kind, functional atlas of the cerebellum. Three important findings emerged from this work: First, our atlas provided a superior account of the underlying data, relative to other existing atlases of the cerebellum. Second, anatomical boundaries were of minimal utility in accounting for the functional patterns observed in the cerebellum. This raises an important caveat given that anatomically defined regions have served as the basis for just about all fMRI or lesion reconstruction studies of the cerebellum. Third, we were able to provide cognitive labels for each cerebellar region. To this end, we made use of an on-line ontology of cognitive terms associated with different tasks. Since publication, the functional atlas, along with the labels, has been widely adopted by the neuroscience community to analyze novel patterns of cerebellar activity, and to use as a guide for testing new hypotheses of cerebellar function.  

centering the developing cerebellum in cognition 

The cerebellum has been the focus of significant debate over the past five decades, and it has been implicated in a wide range of cognitive functions extending beyond sensorimotor control. Much of the empirical research on the function of the cerebellum has been centered on the ‘little brain’ in its mature, adult form. However, we are now starting to appreciate that the developing cerebellum offers a unique window into understanding how and what this structure contributes to cognition. In this review, I document the vulnerability of the developing cerebellum, and present recent work on the co- occurrence of neurodevelopmental disorders (e.g. autism spectrum disorder) and atypical cerebellar development. Building on these observations, I discuss the differences in cerebellar architecture through the lens of injury, and consider how cerebellar function is interpreted from infancy to adulthood.  

mODELING CONNECTIVITY BETWEEN CORTEX AND CEREBELLUM

Collaborators: Joern Diedrichsen; Ladan Shahshahani (Western University) and Richard Ivry (UC Berkeley)

Most neuroimaging studies investigating the cerebellum’s role in cognition do not consider the relative cortical contributions to cerebellar function. In order to meaningfully interpret cerebellar activation, it is critical to examine, in parallel, the relationship between the cortex and the cerebellum. I leveraged a large task battery to establish a quantitative model of connectivity to better understand domain-general functions of the cerebellum. To uncover robust and reliable patterns of cortico-cerebellar connectivity, I trained models on cortical activity patterns and tested them on held-out cerebellar data. I varied the degree of convergence between the cerebral cortex and cerebellum in order to adjudicate between different possible architectures. I made two important contributions with this work: 1) I demonstrated that cerebellar activation is not simply a reflection of a one-to-one mapping between the cortex and the cerebellum, instead it is the result of convergence of inputs from disparate cortical regions onto the same cerebellar region, and 2) I found that the degree of convergence between the cortex and the cerebellum differed across cerebellar regions with highest convergence observed in areas linked to cognition, working memory, and language.