Driving sleep slow oscillations by auditory closed-loop stimulation-a self-limiting process

Abstract

The <1 Hz EEG slow oscillation (SO) is a hallmark of slow-wave sleep (SWS) and is critically involved in sleep-associated memory formation. Previous studies showed that SOs and associated memory function can be effectively enhanced by closed-loop auditory stimulation, when clicks are presented in synchrony with upcoming SO up states. However, increasing SOs and synchronized excitability also bear the risk of emerging seizure activity, suggesting the presence of mechanisms in the healthy brain that counter developing hypersynchronicity during SOs. Here, we aimed to test the limits of driving SOs through closed-loop auditory stimulation in healthy humans. Study I tested a "Driving stimulation" protocol (vs "Sham") in which trains of clicks were presented in synchrony with SO up states basically as long as an ongoing SO train was identified on-line. Study II compared Driving stimulation with a "2-Click" protocol where the maximum of stimuli delivered in a train was limited to two clicks. Stimulation was applied during SWS in the first 210 min of nocturnal sleep. Before and after sleep declarative word-pair memories were tested. Compared with the Sham control, Driving stimulation prolonged SO trains and enhanced SO amplitudes, phase-locked spindle activity, and overnight retention of word pairs (all ps < 0.05). Importantly, effects of Driving stimulation did not exceed those of 2-Click stimulation (p > 0.180), indicating the presence of a mechanism preventing the development of hypersynchronicity during SO activity. Assessment of temporal dynamics revealed a rapidly fading phase-locked spindle activity during repetitive click stimulation, suggesting that spindle refractoriness contributes to this protective mechanism.

Keywords: auditory stimulation; closed-loop control; declarative memory consolidation; fast spindles; sleep; slow oscillations.

Figure 1.

Stimulation protocols. Both the Driving stimulation (left) and the 2-Click stimulation protocols (right) relied on the identification of the negative half-wave peak of an SO in the ongoing EEG (prefiltered between 0.25 and 4 Hz) recorded from AFz during SWS. Thus, an SO-negative peak was identified whenever the EEG signal crossed an initial threshold value of less than −80 μV, which was followed by presenting a first click ∼0.5 s later. The initial threshold value was fine-tuned based on the largest negative amplitude during the preceding 5 s interval, but always <−80 μV. The interval between detection of a negative SO peak and presentation of the first click was individually determined based on recordings in the adaptation night, such that the click occurred at the estimated succeeding positive SO peak. In the Driving stimulation protocol, presentation of the click was followed by the detection of a succeeding SO-negative peak within a 1 s post-click window. For each succeeding click, the threshold was lowered to 80% of the previous click. If the EEG signal did not cross this threshold within the 1 s post-click window, the detection algorithm paused for 2.5 s. During the Sham condition SO detection was performed in an identical manner, however, without presentation of clicks. In the 2-Click stimulation protocol (adopted from Ngo et al., 2013b) presentation of the first click was always followed by presentation of a second click with a fixed interval of 1.075 s. Then the algorithm paused for 2.5 s.

Figure 2.

Number of identified SOs and presented clicks in trains. A, Mean number (±SEM) of SO trains of one, two, three, or four SO cycles in Study I (Driving stimulation vs Sham; left) and clicks presented in Study II (Driving vs 2-Click; right) during SWS within the 210 min stimulation interval. Compared with Sham stimulation, Driving stimulation prolonged SO trains as indicated by reduced numbers of single SO cycles and parallel increased amount of trains consisting of two, three, or four SO cycles (F_(1,14) = 6.548, p = 0.023 and F_(3,42) = 26.367, p < 0.001 for Stimulation main effect and Stimulation × Train length interaction in Study I). B, Mean (±SEM) auto-event correlations across all subjects determined from off-line-detected SO events for Study I (top), Driving stimulation (red line) versus Sham stimulation (black line) and for Study II (bottom), Driving stimulation (red line) versus 2-Click stimulation (dotted black line). *p < 0.05, **p < 0.01, ***p < 0.001, for comparisons between conditions.

Figure 3.

Immediate effects of auditory stimulation. A, Mean (±SEM) EEG signal at Cz for Driving stimulation (red lines) and Sham condition (black lines) of Study I, time-locked to the presentation of the first click (0; n = 15). Stimulations are categorized according to the number of successively identified SOs (1, 2, 3, and 4 SOs). Bottom of each part indicates significant differences between conditions. Vertical dashed lines indicate time points of click presentation. Note, intervals between successive clicks were standardized to 100 bins to account for the temporal jitter, because presentation of each click in a train was precisely timed to the prior identification of an SO-negative half-wave peak. Compared with Sham stimulation, Driving stimulation enhances SO amplitudes. B, Mean (±SEM) interval between SO trains categorized according to the number of SO cycles present in the preceding SO train for the Driving stimulation (red line) and Sham condition (black line) in Study I. *p < 0.05, **p < 0.01, and ***p < 0.001 for comparisons between conditions. C, Corresponding average EEG signals for Study II (n = 13) comparing effects of the Driving stimulation specifically for trains with two successive clicks (2 SOs, red line) with the 2-Click stimulation (dotted black line). Compared with 2-Click stimulation, SO amplitude during Driving stimulation is higher following the first click but lower following the second click. Note also the unexpected significant increase in SO-positive half-wave amplitude preceding the presentation of the first click, specifically in trains comprising two SO cycles, which might be linked to refractoriness in the SO-generating networks, i.e., with Driving stimulation initial clicks occurring at higher depolarization inducing a stronger initial SO response that might subsequently lead to a faster ceasing of the SO train. Note, 100 bins on the x-axis correspond on average to 973.0 ± 11.7 ms.

Figure 4.

Immediate effects of stimulation on fast spindle responses. A, Mean (±SEM) fast spindle activity (12–15 Hz, rms EEG signal) at Cz for Driving stimulation (red lines) and Sham condition (black lines) of Study I, time-locked to the presentation of the first click (0; n = 15). Stimulations are categorized according to the number of successively identified SOs (1, 2, 3, and 4 SOs). Bottom of each part indicates significant differences between conditions. Vertical dashed lines indicate time points of click presentation. The gray line on top illustrates an SO train. Note, intervals between successive clicks were standardized to 100 bins to account for the temporal jitter between succeeding SOs. Driving stimulation profoundly enhances the spindle response only to the first click (with this response peaking around the occurrence of the second click in trains with >2 SOs), whereas later clicks in a train remain almost entirely ineffective. B, Topographical distribution of fast spindle power (12–15 Hz). Difference maps are shown between Driving Stimulation and Sham condition, exemplified for the case of trains with three SO cycles. Black line on top indicates averaged rms spindle activity for the corresponding train and the vertical lines mark time points of click presentation. Significant (p < 0.05, corrected for multiple comparisons) differences between Driving stimulation and Sham condition at specific electrode locations are indicated by filled white circles. C, Corresponding mean (±SEM) fast spindle activity for Study II (n = 13) comparing effects of the Driving stimulation specifically for trains with two successive clicks (2 SOs, red line) with the 2-Click stimulation (dotted black line). Compared with the 2-Click stimulation, Driving stimulation increases spindle responses to the first click, but reduces spindle responses to the second click. Note, 100 bins on the x-axis correspond on average to 973.0 ± 11.7 ms.

Figure 5.

Auditory stimulation enhances SO power and memory retention. A, B, Mean (±SEM) spectral power of the individually determined SO peak (A) and mean (±SEM) SO peak frequency (B) during SWS within the 210 min stimulation period averaged across all subjects and EEG electrodes. C, Mean (±SEM) retention of word pairs across the 7 h nocturnal sleep period (expressed as difference in recalled word pairs at retrieval testing after sleep minus performance at an immediate recall test before sleep). Data are presented separately for the Driving stimulation (black bars) and Sham condition (empty bars) of Study I and the Driving stimulation (black bars) and 2-Click stimulation condition (gray bars) of Study II. *p < 0.05, **p < 0.01 for pairwise comparisons between conditions.