OB respiratory potential during the natural sleep-wake cycle.

a. Schematic representation of electrode placement in the brain and electromyogram recordings from neck muscles (EMG) and diaphragm (dEMG). Bottom: representative example of a respiratory signal and the OB respiratory potential. b. From top to bottom: recordings of respiratory activity derived from the dEMG, cortical electrical activity, and EMG during Wake, NREM, and REM sleep. Note the presence of respiratory potentials in the OB during wakefulness and their absence during both sleep states. c. Representative example of normalized power spectra of the OB (orange), respiration (blue), and coherence spectrum between the two signals (black) during wakefulness and sleep. Note the presence of coherence during wakefulness, demonstrating the internalization of the respiratory signal, and its drop during sleep. The x-axis is the same as in panels e and f. Note that the c plots extend to lower frequencies than the panels below, due to their higher frequency resolution. d. Representative example of instantaneous respiratory frequency across different states of the sleep-wake cycle. e. Phase-amplitude coupling comodulogram across the sleep-wake cycle for a representative animal. The y-axis shows the amplitude of high-frequency activity. The x-axis shows the phase frequency of the respiratory signal extracted from the dEMG (top panel) and the respiration-locked potential in the OB (bottom panel). Note the amplitude modulation of high-frequency activity during wakefulness, which is lost during sleep. f. Same as in e, but averaged across multiple animals (n = 7). g. Quantification of respiration-gamma (60-95 Hz) modulation using the modulation index. * p < 0.05. Resp: Respiration. M1: Primary Motor Cortex. S1: Primary Somatosensory Cortex. V2: Secondary Visual Cortex. N.P.: Normalized Power. PAC: Phase–Amplitude Coupling. Amp. freq.: Amplitude Frequency.

Internalization of the respiratory rhythm during the urethane-activated state and its loss during the slow-wave state.

a. Electrophysiological patterns induced by urethane administration. From top to bottom: respiratory activity extracted from the diaphragmatic EMG (dEMG), chest movement recorded with a motion transducer, cortical electrical activity, and neck EMG. Electrocorticographic traces clearly show an activated state (ASt) with low amplitude and high frequency resembling wakefulness, and a slow-wave state (SWSt) with high amplitude and low frequency resembling NREM sleep. Note the presence of respiratory potentials in the olfactory bulb (OB) during ASt and their absence during SWSt. b. From top to bottom: 1 – Compressed S1 cortical recording. Note that amplitude fluctuations in the raw trace distinguish the two urethane-induced states. 2 – Power of the olfactory bulb. 3 – Power of diaphragmatic respiration. Note that changes in respiration are coupled with fluctuations in OB power. 4 – Coherence between respiration and OB. 5 – Coherence between respiration and gamma-band amplitude (60-95 Hz) in the OB (Cross-frequency coupling). Note that gamma modulation occurs during ASt but not during SWSt. c. Normalized power spectrum (mean ± std) of respiration in both states. The normalization was made to the maximum value of SWSt (/SWSt). d. Instantaneous respiratory frequency averaged across states (n = 7). * p < 0.05. e. Normalized power spectrum of the OB. f. Spectral coherence between respiration and OB. g. Maximum coherence values between respiration and OB EEG, across states. Note the significantly higher coherence during ASt, which decreases during SWSt, resembling the pattern observed during wakefulness and sleep. h. Phase–amplitude coupling (PAC) comodulogram for both states. The y-axis shows the amplitude of high-frequency activity; the x-axis shows the phase frequency of respiration. Note the strong coupling in ASt and its absence in SWSt. i. Modulation index quantification. Note the significant differences between the two states. ** p < 0.001 Resp: Respiration. M1: Primary Motor Cortex. S1: Primary Somatosensory Cortex. V2: Secondary Visual Cortex. N Power: Normalized Power.

Respiration-locked potential and gamma activity are generated locally in the olfactory bulb.

a. Representative recording showing a transition from a slow-wave state (SWSt) to an activated state (ASt). From top to bottom: nasal airflow recorded with a thermistor (Term.), thoracic wall movement (Chest), local field potentials (LFPs) from superficial, middle, and deep layers of the OB recorded with a Neuronexus probe, and cortical activity from S1 and V2 used to determine brain states. b. Left: Coronal section of the olfactory bulb showing the trajectory of a DiI-labeled Neuronexus probe (red) and DAPI-stained nuclei (blue). Center and right: Average current source density (CSD) and LFP (black traces) aligned to the onset of nasal inspiration (dashed line) during ASt and SWSt, respectively. Prominent respiration-locked oscillations with alternating current sinks and sources are observed during ASt, while such modulation is nearly absent during SWSt. Bottom: Mean respiratory signals from the nasal thermistor (blue) and thoracic sensor (orange), aligned to inspiration onset. c. Spectral coherence maps between the respiratory signals (thermistor and chest) and OB LFPs across depth, during ASt and SWSt. Coherence is high during ASt and markedly reduced during SWSt. The same experiment as b. d. Group-level coherence between respiration (thermistor or chest) and either left OB epidural EEG or right OB CSD. For this analysis, the CSD was constructed using the three most distant and equally spaced channels. Coherence was significantly higher during ASt than SWSt for both respiratory signals and recording types. e. Phase–amplitude comodulograms showing the coupling between respiratory phase (thermistor or chest) and high frequency amplitude across OB layers. Same experiment that b. f. Modulation index (MI) quantifying phase–amplitude coupling between respiration and gamma-band (60–90 Hz) activity, computed from left OB EEG and right OB CSD. For this analysis, the CSD was constructed using the three most distant and equally spaced channels. Significant reductions in MI were observed during SWSt compared with ASt across both respiratory sources and recording types. *** p < 0.0001. Inter-Ch: Inter-Chanel. sup., superficial; mid., middle.

The internalization of respiration-locked potential in the OB requires both nasal airflow and cortical activation.

a. Schematic of the double tracheotomy preparation. The lower tracheal cannula was left open to the environment, allowing spontaneous lung ventilation. In contrast, the upper tracheal cannula was connected to an artificial sniffing system that delivered controlled nasal airflow independent of thoracic movement. b. Schematic dorsal view of the rat brain illustrating the anteroposterior distribution of epidural electrodes and LFP probe. primary motor cortex (M1), primary somatosensory cortex (S1), and secondary visual cortex (V2). In the right hemisphere, a 16-channel Neuronexus probe was inserted into the OB for laminar LFP recordings. c. Representative recording during an activated state (ASt). From top to bottom: cortical EEG from M1 and S1; four LFP channels from the OB; TTL signal from the microcontroller triggering the solenoid valve (Pump); thermistor signal (Term) monitoring nasal airflow; and chest movement signal (Chest) reflecting spontaneous respiration. Note that stimulation at 2 Hz during the first half of the recording evoked a stimulus-locked potential, which vanished upon deactivation. d-e. Nasal airflow remained constant throughout the recording, while the cortical state spontaneously transitioned from SWSt to ASt. Note the emergence of respiration-locked OB oscillations only during ASt. f. Coherence between OB LFPs and both respiratory signals (thermistor and chest) across depth. During ASt, nasal airflow (Term) was coherent with OB activity at the stimulation frequency, whereas chest movement showed no coherence in either state. g. Modulation index (MI) between respiratory phase and OB gamma amplitude (60–90 Hz). Robust phase–amplitude coupling was observed exclusively between nasal airflow and olfactory bulb gamma activity at the sniffing frequency, and only during the activated state (ASt). h. Population summaries (mean ± s.d.) of spectra in 0–5 Hz. Top row: normalized power of the respiratory signals (Term/Chest), normalized within each animal to the SWSt maximum for that signal. Middle row: coherence between each respiratory signal and the OB CSD. Bottom row: coherence between each respiratory signal and the OB gamma envelope (60–90 Hz). sup., superficial; mid., middle.

Delta power determines internalization of the respiratory cycle.

a. Left: Example from a single animal showing the negative correlation between delta power and nasal respiration – OB LFP coherence. Orange, black, and blue circles represent ASt, transition, and SWSt epochs, respectively; color-transparent dots show non-significant windows (surrogate analysis). Second panel: Significant linear fits (p< 0.001) of this relationship across all animals and experiments, with the mean trend highlighted in red and the distribution of R² values shown in the inset. Third panel: Example from a single animal showing the relationship between delta power and the modulation index (MI; OB gamma–respiration coupling). Rightmost panel: Same relationship across all animals and experiments, with mean trend and R² distribution inset. b Group-level comparisons of delta power, respiration–OB coherence, and OB gamma–respiration MI across ASt, transition, and SWSt epochs. The last two panels display the percentage of significant windows for the delta–coherence and delta–MI relationships, respectively. Lines link data from the same animal. Statistical significance: **p < 0.01; ***p < 0.001; n.s., not significant.

Extraction of respiratory signal from diaphragm EMG with ECG artifact removal.

a. Schematic representation of the experimental setup. A bipolar electrode was implanted in the diaphragm muscle to monitor breathing-related activity (dEMG). b. From top to bottom: raw diaphragmatic EMG (blue) and the same signal after high-pass filtering above 10 Hz (orange). Detected ECG peaks are marked (×) for subtraction. Post-subtraction trace, absolute value, median filter, and final band-pass filtered signal (0.5–10 Hz), which isolates the respiratory rhythm. c. Illustration of the ECG subtraction process. Left: an individual raw dEMG epoch. Middle: mean ECG waveform computed across ± 10 detected R-peaks (R-peak moving window). Right: signal after template subtraction. d. The processed signal allows accurate estimation of the respiratory frequency, clearly separated from the heart rate peak (respiration in blue, heart rate in orange).

Respiratory-coupled potentials in the OB re-emerge during brief REM sleep periods and microarousals.

a-d. Example traces of respiration (Resp), olfactory bulb (OB), visual cortex (V2), and EMG during REM and NREM d. In each panel, the left trace shows a period with minimal OB respiratory potential. In contrast, the right trace highlights a brief reappearance of the signal, either during short REM sleep periods (a–c) or a microarousal (d). These events illustrate that the respiration-locked potential, while generally absent during NREM and REM sleep, can transiently return during momentary sleep periods.

Phase–amplitude coupling (PAC) in the olfactory bulb across sleep-wake states.

Comodulograms from seven rats (R1–R7) showing phase-amplitude coupling between low-frequency phase (x-axis) and high-frequency amplitude (y-axis) in the olfactory bulb during Wake, NREM, and REM sleep. During wakefulness, strong PAC is observed between the respiratory-locked potential (∼2-10 Hz) and gamma-band amplitude (∼60–110 Hz), which disappears during both NREM and REM sleep. This suggests a state-dependent decoupling of olfactory bulb circuits from respiratory drive during sleep. Comodulograms were computed using the Modulation Index (MI) method (Tort et al., 2010). au, arbitrary units.

State-dependent olfactory bulb tracking of artificial nasal stimulation during urethane-activated state (ASt).

Example recording during ASt with double tracheotomy and artificial sniffing in a chronically implanted rat. The artificial sniff frequency varied throughout the session (top trace), while chest movements and thermistor signals (2nd and 3rd panels from the top) indicate complete respiratory independence between them. The OB LFP (4th panel) and its coherence with the thermistor signal (5th panel) show that stimulus-locked potentials are tightly coupled to artificial sniffing. OB gamma-band activity (6th panel) also follows the artificial sniffing input, suggesting preservation of sensory transmission. In contrast, coherence between chest and OB or thermistor signals (bottom two panels) remains low, confirming effective decoupling of thoracic respiration and nasal air-flow. These findings demonstrate that during ASt, the OB selectively tracks artificial nasal inputs, supporting a gating mechanism that is largely independent of thoracic motor drive. au, arbitrary units.