Sound Asleep: Processing and Retention of Slow Oscillation Phase-Targeted Stimuli

Abstract

The sleeping brain retains some residual information processing capacity. Although direct evidence is scarce, a substantial literature suggests the phase of slow oscillations during deep sleep to be an important determinant for stimulus processing. Here, we introduce an algorithm for predicting slow oscillations in real-time. Using this approach to present stimuli directed at both oscillatory up and down states, we show neural stimulus processing depends importantly on the slow oscillation phase. During ensuing wakefulness, however, we did not observe differential brain or behavioral responses to these stimulus categories, suggesting no enduring memories were formed. We speculate that while simpler forms of learning may occur during sleep, neocortically based memories are not readily established during deep sleep.

Phase-locked loop for precisely timed acoustic stimulation during sleep

“…The PLL is capable of adapting to individual sleep patterns and characteristics without the need to be further calibrated after the initial optimization using baseline information described in section 2.1.3. Whereas the method used by Cox and colleagues [26] for slow-wave phase targeting seems to perform in an inconsistent manner across different subjects, our PLL is very consistent (both within and between subjects) in how precisely the tones are delivered at a given target phase because of the inherent adaptive properties of the PLL. This section contains a detailed discussion of the phase tracking statistical performance of our algorithm.…”

“…Several algorithms to track and target the phase of the EEG during slow wave sleep have been proposed for online acoustic stimulation [22, 26]. Cox et al [26] described an algorithm based on Fast Fourier Transform (FFT) calculation, filtering in a given frequency band, measuring of power, Hilbert transform, and fitting to a sine function to create a predictive phase model of the EEG.…”

“…Cox et al [26] described an algorithm based on Fast Fourier Transform (FFT) calculation, filtering in a given frequency band, measuring of power, Hilbert transform, and fitting to a sine function to create a predictive phase model of the EEG. By inspection of the circular histograms in Cox et al [26] Fig. 1, it seems that the targeting of the phase is inconsistent from subject to subject, whereas in our case we have similar results for different subjects.…”

“…This stimulation method is more general (it can be applied easily to any particular target phase), flexible (it adapts automatically to the slow wave individual characteristics), accurate and precise (it targets the right phase much better and more consistently) than the slow-wave detection procedure described by Ngo and colleagues or the more recent one described by Cox and colleagues [26]. …”

The Fall of Sleep K-Complex in Alzheimer Disease

“…To this aim, recent studies enhancing slow oscillations in humans by means of electrical394041, transcranial magnetic42 and acoustic4344454647 stimulation appear promising. Notably, the boost of slow oscillation during sleep was also associated with beneficial effects on memory processes39404143.…”