This is a brief comment on a recent ALE-analysis of the SRT task (Janacsek et al.)
•
It concluded that the cerebellum does not contribute to sequence learning itself.
•
6 out of the 16 studies should be excluded (i.e., did not scan the cerebellum).
•
Reanalysis without said studies confirmed the original results.
•
Inferences about the role of cerebellum/basal ganglia should be treated with caution.
Abstract
An ALE meta-analysis focused on the serial reaction time task published in NeuroImage (Janacsek et al., 2019) demonstrated consistent activation of the basal ganglia across neuroimaging studies featuring sequence > random block contrasts and no consistent cerebellar activation. To enable valid conclusions regarding the role of the cerebellum in this context, some of the included studies should be excluded (e.g., because the cerebellum was explicitly not scanned). After omitting 6 of 16 studies/subject groups, 70% of the remaining studies did report cerebellar activation. While an ALE analysis of the remaining contrasts confirmed the original results, it may lack the power to detect cerebellar effects. We argue the conclusion that the cerebellum is not involved in sequence-specific learning should be treated with caution.
