¹King's College London, School of Biomedical Engineering and Imaging Sciences, London, United Kingdom

Background: Effective memory consolidation during sleep emerges from orchestrated communication across hippocampal-thalamo-cortical networks. Invasive recordings in rodents and humans have identified electrophysiological markers of effective cortical-hippocampal communication during sleep that correlate with memory performance. However, translating these findings using causal non-invasive approaches is essential for further understanding of the mechanisms underlying memory consolidation and for developing therapeutic applications in psychiatric and neurological conditions, which frequently exhibit disrupted sleep architecture and impaired consolidation processes.

Aims & Objectives: Our group has been working on the development and validation of Temporal Interference Stimulation (TIS) as a non-invasive method for selectively modulating activity in deep brain structures, including the hippocampus. More recently, we have applied this approach to sleep neuroscience, seeking to determine whether TIS can causally modify hippocampal-dependent consolidation processes during sleep. We specifically tested whether stimulation efficacy depends upon precise temporal alignment with memory reactivation events, and whether TIS modulates established neurophysiological indices of successful consolidation within thalamo-cortical networks.

Method: Twenty-nine participants underwent paired-associate encoding followed by daytime sleep monitoring. During stages N2 and N3, we applied TIS targeting the left anterior hippocampus using TIS at physiologically relevant frequencies. Three within-subject conditions were implemented: TIS co-occurring with auditory memory reactivation cues (TIS+Cue), TIS delivered several seconds prior to cues (TIS-before-Cue), and auditory cues without stimulation. High-density electroencephalography provided simultaneous recording of brain activity and stimulation-induced changes. Subsequent testing quantified retention for reactivated versus non-reactivated item pairs.

Results: Memory performance showed a main effect of condition, with participants forgetting fewer cued than non-cued items and, crucially, demonstrating superior retention when TI was synchronized with memory reactivation compared to TI delivered before cues. These behavioural effects were accompanied by modulations in sleep physiology, with spindle amplitude increased selectively in the TIS+Cue condition. Moreover, TI before cues moved spindles towards the pre-trough phase of slow oscillations, whilst TI+Cue maintained coupling at the peak, which has been associated with the optimal phase for consolidation. Multivariate analysis linked electrophysiological features to enhanced memory performance in the TI+Cue condition.

Discussion & Conclusions: These results provide causal evidence that non-invasive stimulation can bidirectionally modulate consolidation depending on temporal precision relative to brain activity. The timing-dependent effects suggest TIS operates by modulating ongoing hippocampal-thalamo-cortical coordination rather than through non-specific excitability changes. The ability to both enhance and disrupt consolidation opens new opportunities for the development of approaches applicable to different psychiatric and neurological conditions.