Because the modulation of an oscillation is not an LFP oscillation itself, its expression does not necessarily appear as a specific power peak in the “primary” power spectrum but can be identified in the second-order spectrum defined as the spectrum of the power fluctuations for a given frequency band (see Figure S1 available online). We

therefore performed second-order spectral analysis (Drew et al., 2008) of LFP power in the integrated theta frequency band (4–11 Hz) and indeed observed the presence of a peak at 0.5–1 Hz, indicating robust expression of TPSM at this frequency in the tested behavioral conditions (Figure 2A for sleep, Figure 2B for open-field, Figure 2C for wheel running). These results were confirmed by additional analysis to exclude the possibility see more that the 0.5–1 Hz peak in the second-order theta power spectrograms might derive from smoothing of theta power in the 1 Hz range under the influence of the selected time window for theta power quantification. First, similar results were obtained using wavelet analysis to quantify instantaneous theta power (Figure S2). Second, we observed that VE-822 clinical trial the second-order theta power spectrum had a peak at the same frequency for varying multi-taper time-window sizes (range 0.6–1.6 s; Figure S2). For further analysis, we kept a window size of 1 s, yielding good peak-to-noise ratio (Figure S2). And third, a similar 0.5-1Hz peak was also

observed in the spectrogram of theta power fluctuations measured as the peak-to-trough amplitude of each theta cycle in the filtered 2-30Hz EEG trace (Figures 2A–2C), excluding the Thymidine kinase potential interference with any filtering or smoothing of theta oscillations (4–11 Hz). Run durations in the maze (typically 3 to 5 s) were too short for spectral analysis below 0.5Hz with accuracy above 0.2 Hz, precluding reliable peak detection in the range 0.5–1 Hz. Nevertheless, the histogram of TPSM cycles durations shows a peak around 1 s, indicating a similar 0.5–1 Hz dominant frequency for TPSM in the maze (Figure 2D). What is the relationship between TPSM and other factors known to modulate theta oscillations?

It has been shown recently that in the behaving monkey auditory cortex (Lakatos et al., 2005), delta oscillations could modulate theta power. However, in our recordings, delta oscillations were typically expressed in the 2–4 Hz frequency range instead of 0.5–1 Hz for TPSM (Figure 3A, wheel). As illustrated in Table S1, we identified and compared the periods of TPSM and delta oscillations (delta power threshold, mean −0.1 SD) in each behavioral condition. During open field and wheel running, we observed that the proportion of TPSM time relative to total time was similar to its proportion relative to delta time, indicating that the expression of delta had no incidence on the expression of TPSM in these behavioral conditions, and vice versa.