State regulation of neuromodulatory control in the neural network

Free-moving
Fiber Photometry
Electrophysiology
Wide-Field Imaging
Revealing the Circuit Mechanisms by Which Dynamic ACh Oscillations During Sleep Cycles Regulate the Selective Long-Term Storage of Memory

Overview

How does the sleeping brain decide which experiences are worth keeping? The synaptic homeostasis hypothesis (SHY) proposes that sleep restores homeostasis through global synaptic downscaling — yet sleep is not a uniform state. Across the NREM–REM cycle, acetylcholine (ACh) oscillates dramatically: suppressed during NREM slow oscillations and hippocampal ripples, then surging to an isolated peak during REM theta. We study how these ACh dynamics serve as a phased control signal that drives global downscaling in NREM while selectively reinforcing learning-relevant synaptic connections during REM — translating sleep architecture into the selective long-term storage of memory.

Figure 1: Sleep cycle structure, ACh dynamics, and recording platform. Left: NREM–REM alternation in primates (N1/N2/N3/REM) and rodents. Center: across the wake–sleep transition, global synaptic strength (SHY theory) declines while specific connection maps are selectively preserved; ACh drops during NREM (slow oscillations, hippocampal ripples) then surges to an isolated peak during REM (theta oscillations). Right: multi-modal recording stack — wide-field imaging, sCMOS, electrophysiology, fiber photometry, and behavioral monitoring (sniff rate, pupil size, speed, face camera, diverse stimuli).

System Architecture

The platform is built around a learning–post-learning sleep paradigm that coordinates three recording modalities with circuit-level temporal resolution.

Fiber Photometry

Real-time readout of bulk neuromodulator signals (ACh, dopamine) and calcium activity from genetically defined cell populations during behavior.

Electrophysiology

Multi-site silicon probe recordings capturing single-unit and LFP activity across hippocampal and olfactory circuits simultaneously with behavior.

1P Imaging & Stimulation

One-photon calcium imaging and optogenetic stimulation with targeted single-cell resolution, phase-locked to the sniff cycle.

Behavioral Paradigm

Animals are free-moving and perform T-maze task:

Event Timing
T-maze Task Alternating trials with left/right arm choices, using visual/auditory cues
Reward Water delivery on correct lick
Laser trigger State-locked — fires at REM onset/offset (\(t_{stim}\))

state detection is critical — it aligns neural activity to a physiologically meaningful reference and enables comparison across trials and animals.

Continuous Monitoring Streams

Beyond sleep-stage classification, we continuously record four parallel signals:

  • ACh — cholinergic tone via fiber photometry, capturing the NREM–REM oscillation cycle in real time
  • LFP — local field potentials from hippocampus and cortex, resolving slow oscillations, sharp-wave ripples, and theta rhythms across sleep stages
  • Wide-field calcium imaging — mesoscale cortical activity via sCMOS, tracking spatiotemporal reactivation patterns across the wake–sleep transition
  • Behavioral state — sniff rate, pupil size, and locomotion speed as continuous arousal and state-transition markers

This multivariate state vector allows us to precisely assign each neural event — a ripple, a reactivation, a firing rate change — to its position within the NREM–REM cycle and the corresponding cholinergic phase, turning correlative observations into a causally interpretable record of how ACh orchestrates memory consolidation during sleep.

NoteSynchronization across modalities

All modalities are referenced to a common hardware clock. Sleep-stage transitions, optogenetic perturbations, and calcium or photometry frames are co-registered to the same timeline, enabling precise alignment of ACh dynamics, network oscillations, and behavioral state at the resolution of individual sleep cycles.


Interested in collaborating or accessing our behavioral platform? Get in touch.