Spatiotemporal patterns of sleep spindle activity in human anterior thalamus and cortex

Abstract Sleep spindles (8 - 16 Hz) are transient electrophysiological events during non-rapid eye movement sleep. While sleep spindles are routinely observed in the cortex using scalp electroencephalography (EEG), recordings of their thalamic counterparts have not been widely studied in humans. Based on a few existing studies, it has been hypothesized that spindles occur as largely local phenomena. We investigated intra-thalamic and thalamocortical spindle co-occurrence, which may underlie thalamocortical communication. We obtained scalp EEG and thalamic recordings from 7 patients that received bilateral deep brain stimulation (DBS) electrodes to the anterior thalamus for the treatment of drug resistant focal epilepsy. Spindles were categorized into subtypes based on their main frequency (i.e., slow (10±2 Hz) or fast (14±2 Hz)) and their level of thalamic involvement (spanning one channel, or spreading uni- or bilaterally within the thalamus). For the first time, we contrasted observed spindle patterns with permuted data to estimate random spindle co-occurrence. We found that multichannel spindle patterns were systematically coordinated at the thalamic and thalamocortical level. Importantly, distinct topographical patterns of thalamocortical spindle overlap were associated with slow and fast subtypes of spindles. These observations provide further evidence for coordinated spindle activity in thalamocortical networks. Highlights Sleep spindles were measured in human anterior thalamus and on the scalp Both fast and slow spindles occurred in the anterior thalamus > 25% of spindles spanned multiple channels in thalamus and cortex A novel statistical approach confirmed that spindle co-occurrences were not random Cortical spindle patterns depended on thalamic involvement and spindle frequency