0051 Direct Stimulation to the Thalamus Evokes Sleep Spindles in Humans

Abstract Introduction The thalamus plays a key role in supporting sleep and can be a target of therapeutic stimulation. We investigated whether electrical stimulation delivered to the thalamus in humans can evoke sleep spindles and the effect of stimulation delivered to different subnuclei. Methods Participants (N=5) were patients with refractory epilepsy and semi-chronic clinically implanted depth electrodes, including at least one in the thalamus, as part of their epilepsy pre-surgical evaluation. Single pulse electrical stimulation (SPES; Bipolar 233microseconds, 7mA) was delivered to the thalamus (2-4 sites) with simultaneous intracranial EEG recordings (100-200 channels) during N2 sleep (assessed from simultaneous scalp EEG). Individual contacts were identified and localized from combined post-operative CT and pre-operative MRI. Anatomical labelling included anterior (ANT), centromedial (CM), and pulvinar (PUL) thalamic subnuclei. Results Stimulation to the thalamus generated responses across the brain during wake and sleep. Stimulation delivered to each subthalamic nuclei had specific response connectivity, with ANT primarily connected to frontal regions, CM to central and other subcortical regions, and PUL to temporal and parietal regions. This connectivity distribution was similar during wake and sleep, although during sleep, the number of channels with a response to stimulation was reduced by 46%. Thalamic SPES evoked K-complexes followed by spindles in all patients and from all thalamic subnuclei. Both K-complexes and spindles were time and phase-locked to the stimulation. Stimulation delivered to the ANT evoked the highest percentage of spindles with the longest duration, especially in frontal regions. Interestingly, spindles alone (without K-complex) were also evoked in frontal regions (N=2). Evoked oscillations were consistent across individual trials, producing spindles in up to 95% of the trials. Mean central frequency of evoked spindles was 12Hz and mean duration was 600ms. Spectral power in the spindle band (10-16Hz) was higher following stimulation (0.025-1.5s) than in the pre-stimulation baseline (p<< 0.001). Conclusion Direct stimulation to the thalamus evokes individual spindles in humans. Different subnuclei may have differential effect on sleep oscillations. Understanding where in the brain and with which stimulation spindles could be evoked could improve our mechanistic understanding of spindle generation and may have implications for sleep integrity and memory consolidation. Support (if any) NIH NIA 5R01AG090302