Delta-range coupling between prefrontal cortex and hippocampus supported by respiratory rhythmic input from the olfactory bulb in freely behaving rats

Abstract

An explosion of recent findings firmly demonstrated that brain activity and cognitive function in rodents and humans are modulated synchronously with nasal respiration. Rhythmic respiratory (RR) coupling of wide-spread forebrain activity was confirmed using advanced techniques, including current source density analysis, single unit firing, and phase modulation of local gamma activity, creating solid premise for investigating how higher networks use this mechanism in their communication. Here we show essential differences in the way prefrontal cortex (PFC) and hippocampus (HC) process the RR signal from the olfactory bulb (OB) allowing dynamic PFC-HC coupling utilizing this input. We used inter-regional coherences and their correlations in rats, breathing at low rate (∼2 Hz) at rest, outside of the short sniffing bouts. We found strong and stable OB-PFC coherence, contrasting OB-HC coherence which was low but highly variable. PFC-HC coupling, however, primarily correlated with the latter, indicating that HC access to the PFC output is dynamically regulated by the responsiveness of HC to the common rhythmic drive. This pattern was present in both theta and non-theta states of waking, whereas PFC-HC communication appeared protected from RR synchronization in sleep states. The findings help to understand the mechanism of rhythmic modulation of non-olfactory cognitive processes by the on-going regular respiration, reported in rodents as well as humans. These mechanisms may be impaired when nasal breathing is limited or in OB-pathology, including malfunctions of the OB epithelium due to infections, such as in COVID-19.

Figure 1.

A. Sample recording of respiratory rhythm (black) derived from diaphragmal EMG (grey) along with LFPs in OB, PFC, and HC and neck muscle EMG in QW state. B. Group averages of dia autospectra in different states. Note narrow RRO peaks in all recordings at ~2 Hz. Power is shown in arbitrary units after normalization of autospectra in individual recordings setting maxima equal to 1. C. Group averages of dia-OB coherence spectra, in different states. Note coherence peaks constrained to RRO frequencies (dia spectral peaks) in sleep and in a wider range, up to 6 Hz in wake states. In AW, dia-OB coherence does not have a clear RRO peak on the group average due to interindividual variability of the respiratory rates (see in Fig. S2).

Figure 2.

Comparison of state-dependent RRO coherences in PFC and HC transferred through OB. A. Peak coherence at RRO frequency between rhythmic dia activity and LFP in the OB and between OB and cortical (PFC) and hippocampal (HC) networks during different sleep-wake states. Note strong state dependence and nearly identical dia-OB and OB-PFC coherences in all recordings and considerably lower OB-HC coherence. Squares: group averages, dots: individual experiments; same colors identify individual rats. B. Variability of coherence values in individual experiments in different states. Coefficient of variation (top) and CV ratio (bottom) of coherences in the OB-HC vs. the other two signal-pairs (dia-OB and OB-PFC). Note high variation of OB-HC in waking (AW and QW) and REM sleep, 2–3 times exceeding CV of the other pairs. C. Relationship between peak RRO coherences connecting dia to OB (Dia-OB) and those connecting OB to neural networks of PFC (left) and HC (right) in different states. Trend-lines with non-significant correlations (p>0.1) are shown in grey; solid lines show theta, dashed line show non-theta states. Note significant positive correlation of between dia-OB and OB-PFC, but no positive correlation between dia-OB and OB-HC coherences.

Figure 3.

Coupling of PFC and HC networks by RRO. A. PFC-HC coherences in different states. B-C. PFC-HC coherences at RRO in awake (QW, AW) and sleep states (SWS, REM) at RRO (1–3 Hz) and at theta (6–8 Hz) frequencies in theta states (AW, REM). Squares: group averages, dots: individual experiments; same colors were used for individual rats in different states (same colors as in Fig. 2A). D. Correlation between peak RRO coherences connecting PFC and HC vs. RRO coherences connecting OB to HC, dia, and PFC signals in AW (filled symbols and solid trendlines) and QW (open symbols and dashed trendlines) recordings at RRO frequency RRO. Significant correlations are shown in the color of the corresponding dots, trendlines of non-significant correlations are shown in grey.