A large number of neurons form cell assemblies that process information in the brain. Recent developments in measurement technology, one of which is calcium imaging, have made it possible to study cell assemblies. In this study, we aim to extract cell assemblies from calcium imaging data. We propose a clustering approach based on non-negative matrix factorization (NMF). The proposed approach first obtains a similarity matrix between neurons by NMF and then performs spectral clustering on it. The application of NMF entails the problem of model selection. The number of bases in NMF affects the result considerably, and a suitable selection method is yet to be established. We attempt to resolve this problem by model averaging with a newly defined estimator based on NMF. Experiments on simulated data suggest that the proposed approach is superior to conventional correlation-based clustering methods over a wide range of sampling rates. We also analyzed calcium imaging data of sleeping/waking mice and the results suggest that the size of the cell assembly depends on the degree and spatial extent of slow wave generation in the cerebral cortex.

Body rocking can either induce sleep or arousal. That is, the vestibular sense influences sleep-wake states. Neuronal interactions between sleep-wake systems and vestibular systems, however, remain unclear. In this study, we found that GABAergic neurons in the lateral part of the medial vestibular nucleus (LMVN), a primary vestibular afferent projection site, control sleep-wake states. Specific inhibition of LMVN GABAergic neurons revealed that the firing of LMVN GABAergic neurons underlies stable wakefulness and smooth transitions from non-rapid-eye-movement (NREM) sleep to rapid eye movement (REM) sleep and that LMVN GABAergic neurons do not affect body balance control in freely moving conditions. Selective axonal tracing of LMVN GABAergic neurons indicated that LMVN GABAergic neurons send axons not only to areas involved in vestibular and oculomotor functions but also to areas regulating sleep-wake states. Our findings suggest that LMVN GABAergic neurons stabilize wakefulness and gate the entry into REM sleep through the use of vestibular information.

Cortical neurons fire intermittently and synchronously during non-rapid eye movement sleep (NREMS), in which active and silent periods are referred to as ON and OFF periods, respectively. Neuronal firing rates during ON periods (NREMS-ON-activity) are similar to those of wakefulness (W-activity), raising the possibility that NREMS-ON neuronal-activity is fragmented W-activity. To test this, we investigated the patterning and organization of cortical spike trains and of spike ensembles in neuronal networks using extracellular recordings in mice. Firing rates of neurons during NREMS-ON and W were similar, but showed enhanced bursting in NREMS with no apparent preference in occurrence, relative to the beginning or end of the on-state. Additionally, there was an overall increase in the randomness of occurrence of sequences comprised of multi-neuron ensembles in NREMS recorded from tetrodes. In association with increased burst firing, somatic calcium transients were increased in NREMS. The increased calcium transients associated with bursting during NREM may activate calcium-dependent, cell-signaling pathways for sleep related cellular processes.

Rapid eye movement (REM) sleep is primarily regulated by the brainstem pons. In particular, the sublaterodorsal tegmentum (SubLDT) in the dorsal pons contains neurons whose activity is selective to REM sleep. Elucidation of the precise identities of these neurons and their roles in REM sleep regulation is challenging, however, due to the functional and molecular heterogeneity of the SubLDT. A recent study revealed that corticotropin-releasing hormone–binding protein (Crhbp)-positive neurons in the SubLDT projecting to the medulla play a crucial role in REM sleep regulation and that loss of these Crhbp-positive neurons underlies sleep deficits observed in Parkinson's disease. The firing patterns of these neurons during sleep/wake, however, remained unknown. Here, we used an opto-tagging method and conducted cell-type-specific recordings from Crhbp-positive neurons using a glass pipette microelectrode in unanesthetized male mice. We recorded 58 Crhbp-positive neurons and found that many of these neurons are REM sleep-active neurons (41.4%) and that the remaining neurons are mostly either wake-active, wake/REM sleep-active, or non-REM sleep-active. In addition, projection-specific recordings revealed that the medulla-projecting Crhbp-positive neurons are mostly REM sleep-active neurons (75.0%). Based on clustering analysis and spike waveform analysis, REM sleep-active Crhbp-positive neurons can be further divided into different subtypes according to their electrophysiological properties, suggesting that Crhbp-positive neurons play diverse roles in REM sleep regulation.

Despite the importance of sleep to the cerebral cortex, how much sleep changes cortical neuronal firing remains unclear due to complicated firing behaviors. Here we quantified firing of cortical neurons using Hawkes process modeling that can model sequential random events exhibiting temporal clusters. “Intensity” is a parameter of Hawkes process that defines the probability of an event occurring. We defined the appearance of repetitive firing as the firing intensity corresponding to “intensity” in Hawkes process. Firing patterns were quantified by the magnitude of firing intensity, the time constant of firing intensity, and the background firing intensity. The higher the magnitude of firing intensity, the higher the likelihood that the spike will continue. The larger the time constant of firing intensity, the longer the repetitive firing lasts. The higher the background firing intensity, the more likely neurons fire randomly. The magnitude of firing intensity was inversely proportional to the time constant of firing intensity, and non-REM sleep increased the magnitude of firing intensity and decreased the time constant of firing intensity. The background firing intensity was not affected by the sleep/wake state. Our findings suggest that the cortex is organized such that neurons with a higher probability of repetitive firing have shorter repetitive firing periods. In addition, our results suggest that repetitive firing is ordered to become high frequency and short term during non-REM sleep, while unregulated components of firing are independent of the sleep/wake state in the cortex. Hawkes process modeling of firing will reveal novel properties of the brain.