Targeted memory reactivation during sleep with closed-loop auditory stimuli: Comparing the effects of slow oscillatory up-phase and down-phase cueing on sleep-dependent declarative memory consolidation

In the last decade, an increasing number of studies have demonstrated that using a relatively new technique called ‘targeted memory reactivation’ can modulate and enhance memory consolidation during sleep. This technique is performed by matching specific externally applied sensory stimuli (e.g., sound cues) with target information (e.g., associated word pairs) during wakefulness and then presenting the learned cues during subsequent non-REM sleep. The present study combines this technique with an auditory closed-loop stimulation (ACLS) algorithm, which detects slow oscillations (SOs) and triggers the presentation of the auditory cues precisely during a specific phase of a SO once the EEG signal passes a certain amplitude threshold. SOs are high-amplitude, lowfrequency undulating EEG signals, which represent the hallmark oscillations during slowwave sleep (SWS). Following the widely accepted assumption that SOs play a central role in the consolidation of memory by driving memory reactivation processes during SWS, the present study aimed to identify the optimal time window for the presentation of auditory cues within a single SO during SWS in order to optimize targeted memory reactivation and thus overnight consolidation. The main focus of the study lay thereby on the comparison of the differential benefits on declarative memory retention during two stimulation conditions: SO up-phase and down-phase. For this purpose, auditory stimuli were presented to subjects during non-REM sleep, timed to set in with SO negative peaks during down-phase stimulation conditions, and with SO positive peaks during up-phase stimulation conditions. As SO up states are assumed to represent a state of increased neuronal firing, the hypothesis of this study proposes that auditory cueing in-phase with online detected SO up states leads to better declarative memory consolidation than cueing during SO down states. During the evening of each experimental condition, subjects performed on a paired-associate learning task (PAL task), during which they were presented with each word pair on a screen while the first syllable of the first word was played simultaneously over in-ear headphones. They were first tasked to learn to match syllables with corresponding word pairs and then to recall a word pair when presented with the matching syllable only. Following this learning session of declarative memory contents after reaching a criterion of 60% correctly recalled word pairs, memory performance was tested once immediately and once during a delayed recall session the next morning. The syllables used in the task where of 600-ms length each and were later presented as cues during post-learning nonREM sleep. Contrary to expectations, analyses of behavioral results during up-phase conditions showed no significant declarative memory benefit compared to down-phase conditions. Also, no significant cueing effect could be observed during both conditions. Analyses of subjects’ averaged auditory evoked potential signals showed the typical immediate evoked potential responses for up-phase and down-phase stimulations, indicating correctly performed auditory cueing during both experimental conditions. Additionally, subjects performed on additional cognitive tests and psychometric assessment tasks, the results of which showed no significant differences for both conditions. The hypothesis of the present study, stating that up-phases represent a superior time window for auditory stimulations compared to down-phases of SOs, was not validated. However, the study facilitated new insights into the relatively new field of closed-loop targeted memory reactivation for potential memory enhancement. As was demonstrated in previous research, targeted memory reactivation combined with precise closed-loop auditory stimulation seems not only to represent a non-invasive and effective method to enhance memory consolidation during sleep, but also to hold various potential benefits, particularly in regard to clinical applications in the field of psychotherapy and neurological disorders. As a result, the present study has undeniably demonstrated that further research is still necessary and might benefit from the present study’s findings as many questions remain unanswered, particularly in regard to specific cueing conditions, underlying neural mechanisms, long-term effects, limitations, and potential side effects of the cued memory reactivation technique