Functional characterization of wake/sleep cycles

This report presents a novel methodology for the functional characterization of
wakefulness and sleep cycles based on large-field-of-view two-photon calcium
imaging. The approach aims to investigate spatiotemporal patterns of neuronal
activity without relying on traditional electrophysiological techniques such as EEG
or LFP. Using head-fixed mice expressing the calcium indicator GCaMP6f in cortical
Layer 2/3 neurons, brain states were inferred through pupil size dynamics, which
provided a reliable marker to differentiate between wakefulness and NREM sleep.
Machine-learning-based analytical methods, including Hidden Markov Models
(HMM) for identifying latent brain states, were developed to extract distinct
patterns of neuronal activity. The results revealed that neuronal activity alternates
between different modes both during wakefulness and NREM sleep, suggesting
that cortical activity across the sleep cycle is more heterogeneous than previously
thought. While neuropil-based analysis did not reliably distinguish wakefulness
from sleep, the combination of pupil-based classification and HMM successfully
captured the fragmented nature of rodent sleep. The methodology represents a
significant advancement in the study of sleep-related neuronal dynamics, offering
a scalable framework for investigating brain states in both animal models and
cerebral organoids. Future applications will further explore the spatial organization
of activity modes and their relationship across wake and sleep, with potential
implications for understanding sleep disruptions in neurodegenerative conditions
such as Parkinson's disease.