Actions from glial cells could affect the readiness and efficacy of learning and memory. Using a mouse cerebellar-dependent horizontal optokinetic response motor learning paradigm, short-term memory (STM) formation during the online training period and long-term memory (LTM) formation during the offline rest period were studied. A large variability of online and offline learning efficacies was found. The early bloomers with booming STM often had a suppressed LTM formation and late bloomers with no apparent acute training effect often exhibited boosted offline learning performance. Anion channels containing LRRC8A are known to release glutamate. Conditional knockout of LRRC8A specifically in astrocytes including cerebellar Bergmann glia resulted in a complete loss of STM formation while the LTM formation during the rest period remained. Optogenetic manipulation of glial activity by channelrhodopsin-2 or archaerhodopsin-T (ArchT) during the online training resulted in enhancement or suppression of STM formation, respectively.STM and LTM are likely to be triggered simultaneously during online training, but LTM is expressed later during the offline period. STM appears to be volatile and the achievement during the online training is not handed over to LTM. In addition, we found that glial ArchT photoactivation during the rest period resulted in the augmentation of LTM formation. These data suggest that STM formation and LTM formation are parallel separate processes. Strategies to weigh more on the STM or the LTM could depend on the actions of the glial cells.

The structures of neurons, such as dendrites and axonal projections, are closely related to their response properties and their specific functions in neural circuits. Identified neurons, having genetically determined morphological features and pre‐ and postsynaptic partners, play significant roles in specific behaviors. Giant interneurons (GIs) are identified in the terminal abdominal ganglion of the cricket as mechanosensory projection neurons and are sensitive to airflow stimulation of the cerci. GIs are classified into ventral GIs (vGIs) or dorsal GIs (dGIs) depending on the location of their axons running within the connective nerve cord. Based on their response properties to airflow, vGIs are presumed to be involved in triggering the wind‐elicited escape response, whereas dGIs are thought to be airflow direction‐encoding neurons. The previous findings regarding airflow sensitivity point to possible differences in the morphology of the central projections that may correspond to their neural functions. However, the detailed morphologies of the GIs in the cephalic and thoracic ganglia of adult crickets remain unclear. In this study, we stained six GIs, namely, GI 8‐1 (medial giant interneuron, MGI), 9‐1 (lateral giant interneuron, LGI), 9‐2, 9‐3, 10‐2, and 10‐3, using intracellular iontophoretic or pressure injection of dyes. Staining revealed remarkable differences in the axonal branching patterns between vGIs and dGIs. The dGIs were further divided into subgroups based on the profiles of their axon collaterals and projection sites in the brain. The anatomical differences between the GIs’ central projections seemed to be related to their information encodement and behavioral functions.