Transcranial slow oscillation stimulation during NREM sleep enhances acquisition of the radial maze task and modulates cortical network activity in rats

Abstract

Slow wave sleep, hallmarked by the occurrence of slow oscillations (SO), plays an important role for the consolidation of hippocampus-dependent memories. Transcranial stimulation by weak electric currents oscillating at the endogenous SO frequency (SO-tDCS) during post-learning sleep was previously shown by us to boost SO activity and improve the consolidation of hippocampus-dependent memory in human subjects. Here, we aimed at replicating and extending these results to a rodent model. Rats were trained for 12 days at the beginning of their inactive phase in the reference memory version of the radial arm maze. In a between subjects design, animals received SO-tDCS over prefrontal cortex (PFC) or sham stimulation within a time frame of 1 h during subsequent non-rapid eye movement (NREM) sleep. Applied over multiple daily sessions SO-tDCS impacted cortical network activity as measured by EEG and behavior: at the EEG level, SO-tDCS enhanced post-stimulation upper delta (2-4 Hz) activity whereby the first stimulations of each day were preferentially affected. Furthermore, commencing on day 8, SO-tDCS acutely decreased theta activity indicating long-term effects on cortical networks. Behaviorally, working memory for baited maze arms was enhanced up to day 4, indicating enhanced consolidation of task-inherent rules, while reference memory errors did not differ between groups. Taken together, we could show here for the first time an effect of SO-tDCS during NREM sleep on cognitive functions and on cortical activity in a rodent model.

Keywords: EEG; consolidation; reference memory; sleep; slow oscillation stimulation; tDCS; working memory.

Figure 1

Behavioral measures (mean ± s.e.m.). (A) Reference memory errors. (B) Working memory errors. (C) Working memory errors made on baited arms (= re-entries into baited arms where the bait has been consumed already within the ongoing trial). (D) Working memory errors made on never baited arms (= re-entries into never baited arms within the ongoing trial).• Represent STIM condition, ◦ represent SHAM condition. ANOVAs for repeated measures followed by post-hoc t-tests. ^*p < 0.05, ^#p < 0.1.

Figure 2

Total sleep time (TST; mean ± s.e.m.) across the daily 2h-recording period in minutes. • Represent STIM condition, ◦ represent SHAM condition. There were no differences between the conditions, but TST decreased over experimental days.

Figure 3

Upper delta power within the 10 s intervals of post-stimulation NREM sleep. (A) Mean upper delta power (2.08–4.03 Hz) within the first and the last post-stimulation interval of the day across all 12 experimental days. ANOVAs for repeated measures followed by post-hoc t-tests. ^*p < 0.05. (B) Mean power spectra within all 10 s intervals of post-stimulation NREM sleep across all 12 experimental days. Upper delta band is marked in light gray. Note the peak at ~2.5 Hz in the STIM, but not in the SHAM condition.

Figure 4

EEG power during the acute (sham)stimulation. A significant group difference and an interaction could only be seen for the theta band (5.00–9.03 Hz). Note, SO power could not be analyzed due to frequency overlap with SO-tDCS. ANOVAs for repeated measures followed by post-hoc t-tests. ^*p < 0.05, ^#p < 0.01.