Cerebellar Contributions to Visual Attention and Visual Working Memory Revealed by Functional MRI and Intrinsic Functional Connectivity

Abstract

The study of cerebellum function has been traditionally limited to the motor domain. Recent research, however, has begun to characterize the cerebellum's role in cognition (see Schmahmann, 2010) and has demonstrated intrinsic functional connectivity between cerebral cortical networks and distinct cerebellar regions (Buckner et al., 2011). Here, in two separate fMRI experiments, we investigated whether cerebro-cerebellar connectivity of dorsal attention network (DAN) predicts cerebellar activation during visual attention and visual working memory (VWM) task performance. In experiment 1 (N=8), subjects performed a multiple-object tracking task. In experiment 2 (N=9), subjects performed a VWM change detection task using oriented bars. Memory load was varied across blocks (set size: SS0, SS1, or SS4). Both experiments employed resting-state functional connectivity analysis using cortical network seeds (Yeo et al., 2011) to parcellate cerebro-cerebellar networks in individual subjects. In experiment 1, a region-of-interest analysis revealed a robust attentional effect within cerebellar regions functionally connected to the cortical DAN (p<.01). Conversely, cerebellar regions functionally connected to the cortical default mode network (DMN) showed reliable deactivation (p<.001). In experiment 2, contrasting SS4 with SS0 and SS1 resulted in a similar pattern of competitive interaction between cerebellar nodes of the DAN and DMN. Load-dependent activation spatially corresponded with cerebellar DAN nodes (SS4-SS0: p<.005; SS4-SS1: p<.0001) and load-dependent deactivation was observed within cerebellar DMN nodes (SS4-SS0: p<.005; SS4-SS1: p<.0005). Across both experiments the strength of intrinsic functional connectivity, with either the cortical DAN or the cortical DMN, significantly predicted the response of individual cerebellar voxels (Experiment 1: rDAN =.67, rDMN =-.71; Experiment 2: rDAN =.60, rDMN =-.56). Our results indicate that cerebellar nodes of the DAN contribute to network function across a diverse range of attentive and working memory conditions. Meeting abstract presented at VSS 2015.