Inhibition recruitment in prefrontal cortex during sleep spindles and gating of hippocampal inputs

Abstract

During light slow-wave sleep, the thalamo-cortical network oscillates in waxing-and-waning patterns at about 7 to 14 Hz and lasting for 500 ms to 3 s, called spindles, with the thalamus rhythmically sending strong excitatory volleys to the cortex. Concurrently, the hippocampal activity is characterized by transient and strong excitatory events, Sharp-Waves-Ripples (SPWRs), directly affecting neocortical activity--in particular the medial prefrontal cortex (mPFC)--which receives monosynaptic fibers from the ventral hippocampus and subiculum. Both spindles and SPWRs have been shown to be strongly involved in memory consolidation. However, the dynamics of the cortical network during natural sleep spindles and how prefrontal circuits simultaneously process hippocampal and thalamo-cortical activity remain largely undetermined. Using multisite neuronal recordings in rat mPFC, we show that during sleep spindles, oscillatory responses of cortical cells are different for different cell types and cortical layers. Superficial neurons are more phase-locked and tonically recruited during spindle episodes. Moreover, in a given layer, interneurons were always more modulated than pyramidal cells, both in firing rate and phase, suggesting that the dynamics are dominated by inhibition. In the deep layers, where most of the hippocampal fibers make contacts, pyramidal cells respond phasically to SPWRs, but not during spindles. Similar observations were obtained when analyzing γ-oscillation modulation in the mPFC. These results demonstrate that during sleep spindles, the cortex is functionnaly "deafferented" from its hippocampal inputs, based on processes of cortical origin, and presumably mediated by the strong recruitment of inhibitory interneurons. The interplay between hippocampal and thalamic inputs may underlie a global mechanism involved in the consolidation of recently formed memory traces.

Fig. 1.

Interneuron and superficial pyramidal cell recruitment during spindles. (A and B) Example of neuronal recruitment by spindles by layers. (A) Deep LFP trace (Upper) shows a transient bout of waxing-and-waning oscillations at about 15 Hz, the spindle oscillations, denoted by the horizontal gray bar; below: raster of discriminated units spiking activity in the superficial layers (II/III) and deep layers (V/VI). Firing from two FS cells, one in each layer, is depicted here in color and RS cell spikes are displayed in black. Note the δ-wave preceding the spindles (green asterisk). (Lower) An enlarged picture of the spindle oscillation: filtered deep LFP trace (10–15 Hz) is shown together with the spiking activity of the same FS cells as above. (B) Phase histogram of the example FS cells from A relative to spindle oscillations (same color code). (C) Percentage of significantly phase modulated cells. (D) Average relative change in firing rate during spindle oscillations compared with nonspindle SWS for significantly phase-locked cells (Mod., purple) or not (Non mod., gray) and for all of the cells (All, green) changed to green. Error bars display SEM. (E) Distributions of preferred firing phase of cells. Black bars indicate preferred phase for all cells; distributions for significantly phase-modulated cells only are shown in color lines.

Fig. 2.

Interrelationship between SPWRs, spindles, and the neocortical slow oscillation at various timescales. (A) Example of simultaneous unfiltered mPFC LFP (Upper trace, green asterix indicates a δ-wave) and hippocampal LFP in the pyramidal layer filtered in the ripple band (Lower trace, red asterix indicates a detected SPWR; signal is z-scored). (B) Cross-correlograms of δ-waves (reference) and SPWRs occurrence time in function of depth of the δ-wave (expressed in z-score of the filtered LFP in the 0.1–4 Hz band) (SI Materials and Methods). All δ-waves above 4 SD are grouped in the last row. Colors indicate occurrence rate. (B) Same as A but for spindle peaks. (C) Cross-correlograms between SPWRs (reference) and spindle peaks at various timescales, from fine (Top) to long timescales (Bottom). The thin gray line in the bottom panel is the time-reversed correlogram displayed to emphasize the difference between negative (Left) and positive (Right) time-lag differences. Only long SWS episodes (>12 min) were used for the last cross-correlation.

Fig. 3.

Prefrontal cells enhanced responsiveness to hippocampal SPWRs during rather than outside spindle episodes. (A) Average (Upper) and individual (Lower) z-scored cross-correlograms of superficial FS cells relative to SPWRs occurrence time during spindle oscillation episodes (Left) and nonspindle state (Right). Cells are ordered relative to the strength of activation during non spindle states. (B) Summary statistics of FS cell (Left) and RS cell (Right) modulation by SPWRs. Cross-correlograms were z-scored on a ± 400-ms window around SPWRs and averaged over a ± 40-ms window (*P = 0.026, **P = 0.0018, ***P < 10⁻⁹; paired t test).

Fig. 4.

Example of phase-relationship between spindle oscillations and γ-burst in the mPFC. (A) Raw LFP trace (Top, black) filtered in the spindle band (Middle, blue), and 30- to 140-Hz γ-band (Bottom, red). The green asterisk indicates a δ-wave. (B) Time-frequency analysis in the γ-band reveals the occurrence of γ-bursts at different frequencies that tend to appear at the peak of the deep-layer spindle oscillations, as shown in the enlarged panel.

Fig. 5.

γ-Recruitment in the mPFC during spindles and SPWRs. (A and B) γ-Modulation by spindle oscillations: broadband (30–140 Hz) γ-power (average of z-scored cross-correlograms, A) and LFP spectogram triggered on spindle peaks (z-scored for each frequency band in a ± 750-ms window), in deep (blue) and superficial (red) layers. Inset in A shows the power spectra (in arbitrary units) of the averaged γ-power shown in A. Shaded areas indicate 95% confidence intervals (Jackknife bootstrap). (C1 and D1) Same as A and B but triggered on SPWRs during spindling episodes. (C2 and D2) Same as C1 and D1 but during nonspindling epochs. (E) Wide-band γ-response to SPWRs outside spindle episodes in function of SPWR strength quartile (SI Materials and Methods) in the superficial (Upper) and deep (Lower) layers. (F) Zero-lag γ-power (averaged in a ± 20-ms window) in function of SPWR strength quartiles. (G) Distribution of correlation coefficients between γ-amplitude change at 0-lag (computed as in F) with SPWR quartiles across all LFP signals measured in the superficial (Upper) and deep layers (Lower). Shaded areas in A, C, and E, and error bars in F display SEM.