Abstracts

Rationale:
We previously found a strong association between behavioral state (wake, REM, and NREM) and network excitability as determined by optogenetic population discharge thresholds. In order to understand the mechanisms responsible for this association and determine the influence of the synchronous activity characteristic of NREM sleep states, we examined the effect of pre-stimulus oscillatory phase on the population discharge probability.
Method:
We measured epileptiform population discharge probability in chronically implanted transgenic mice using optogenetic dose-response curves. Our chronic preparation consists of an optrode above area CA1 of Thy1-ChR2 (line 18) mice and 7 individually placed microwires in hippocampal and peri-hippocampal structures to monitor the network wide propagation of optogenetic stimulation. A range of light intensities was presented in randomized order, occasionally producing all-or-none high amplitude oPDs. The conditional probability of oPD given each light intensity was calculated and fit with the Boltzmann equation to determine an I50, or the light intensity at which oPDs occur 50% of the time. This metric, the I50, can be used to precisely track network excitability in freely moving mice. Using long-term oPDT recordings, we segmented the data by behavioral state using a convolutional neural network to automate sleep scoring. Pre-stimulus oscillation phase was determined using the filter-Hilbert method. Phases were binned and the conditional probability of an oPD was calculated for each bin. In addition, an unbiased estimate of the influence of the pre-stimulus period was generated by averaging the pre-stimulus period of stimuli near the I50 (where the P(oPD) ~ 0.5) separated by outcome (oPD and non-oPD) and by behavioral state (wake, NREM, and REM). 
Results:
Long-term (8hr) oPDT recordings revealed a strong effect of the phase of ongoing sleep oscillations. Separating the pre-stimulus period by sub- or supra-threshold responses reveals an effect of ongoing phase on PD probability that varies by behavioral state (wake, REM, NREM). During wake periods, there was no clear association between phase and probability, but during NREM, a strong effect of the phase of delta was detected with the highest probability of oPD on the rising phase/peak and lowest probability during the trough. During REM, theta modulates the oPDT with oPDs more likely when stimulation occurred during the peak. The greatest differences in the average pre-stimulus period for oPD vs. non-oPD outcomes were observed in entorhinal cortex, dentate gyrus, CA3, and subiculum.  
Conclusion:
We have previously demonstrated a strong effect of behavioral state (wake, NREM, REM) on network excitability. In the current study, we show that this modulation is due, in part, to the synchronizing effect of naturally occurring sleep oscillations.
Funding:
:NINDS 1RO1NS105005-01, NIAAA 2 R01 AA016852-11, NCTIC Pilot Award