The human cerebellum is thought to interact with distributed brain networks to support cognitive abilities such as episodic memory and semantic prediction. Hippocampal and fronto-temporo-parietal networks that respectively support episodic memory versus semantic prediction have been associated with distinct endogenous oscillatory activity frequency bands: theta (~3–8 Hz) versus beta (~13–30 Hz) respectively. We sought to test whether it is possible to toggle cerebellar participation in episodic memory versus semantic prediction by noninvasively stimulating with theta versus beta rhythmic transcranial magnetic stimulation. In human subjects of both sexes, cerebellar theta stimulation improved episodic memory encoding but did not influence neural signals of semantic prediction, whereas beta stimulation of the same cerebellar location increased neural signals of semantic prediction but did not influence episodic memory encoding. This constitutes evidence for double dissociation of cerebellar contributions to semantic prediction versus episodic memory based on stimulation rhythm, supporting the hypothesis that the cerebellum can be biased to support these distinct cognitive abilities at the command of network-specific rhythmic activity.

SIGNIFICANCE STATEMENT The cerebellum interacts with several distinct large-scale brain networks for cognitive function, but the factors governing selectivity of such interactions for particular functions are not fully understood. We tested the hypothesis that cerebellar contributions to cognition are guided by neural oscillations with function-specific frequency bands. We demonstrated that matching noninvasive stimulation to network-specific frequencies selectively enhanced episodic memory versus semantic prediction. These findings suggest that cerebellar contributions to cognitive networks are selected based on corresponding activity rhythms and could be used to develop cerebellar stimulation interventions for specific neurocognitive impairments.

人们认为小脑与分布式脑网络进行交互以支持认知能力，例如情节记忆和语义预测。分别支持情景记忆和语义预测的海马和额颞顶网络已经与不同的内源性振荡活动频带相关：分别为theta（〜3–8 Hz）vs beta（〜13–30 Hz）。我们试图通过θ与β节律性经颅磁刺激无创性刺激来测试小脑参与情节记忆与语义预测之间的切换是否可能。在男女受试者中，小脑theta刺激都能改善情节记忆编码，但不会影响语义预测的神经信号，而相同小脑位置的beta刺激会增加语义预测的神经信号，但不会影响情景记忆编码。这构成了小脑对语义预测和基于刺激节律的情节记忆的双重分解的证据，支持了小脑可以偏向于支持特定于网络的节律活动来支持这些独特的认知能力的假设。

意义声明小脑与几个截然不同的大规模大脑网络相互作用以实现认知功能，但是控制这种相互作用对特定功能选择性的因素尚不完全清楚。我们检验了小脑对认知的贡献是由具有特定功能频段的神经振荡指导的这一假设。我们证明，将无创刺激与网络特定的频率相匹配可以选择性地增强情节记忆与语义预测之间的关系。这些发现表明，小脑对认知网络的贡献是根据相应的活动节律选择的，可用于开发针对特定神经认知障碍的小脑刺激干预措施。