Klinische Neurophysiologie 2004; 35 - 235
DOI: 10.1055/s-2004-832147

Functional Connectivity between Cortex and Motor Nuclei of the Basal Ganglia across the Human Sleep-Wake Cycle

F Salih 1, P Grosse 2, A Sharott 3, R Khatami 4, T Trottenberg 5, GH Schneider 6, A Kupsch 7, P Brown 8
  • 1Berlin
  • 2Berlin
  • 3Berlin
  • 4Berlin
  • 5Berlin
  • 6Berlin
  • 7Berlin
  • 8Berlin

Functional imaging has shown a basal ganglia involvement in the activity changes across the human sleep-wake cycle. Here we used frequency analysis to explore whether motor nuclei of the basal ganglia (Globus pallidus internus [GPi], subthalamic nucleus [STN]) are functionally connected to the motor or the visual cortex in wakefulness, NREM2, NREM3/4 and REM sleep. To this end, we examined patients undergoing deep-brain stimulation for Parkinson's disease (PD; n=5) or dystonia (n=5). We simultaneously recorded local field potentials either from GPi or STN using a bipolar montage as well as scalp EEG (F3/4-C3/4, Pz-O1/2). Frequency analysis of scalp EEG between motor and visual cortex in GPi patients confirmed previous studies in healthy subjects during sleep (Achermann & Borbély, 1998), whereas in STN patients additional peaks were present in the delta-/theta-frequency range. In the coherence spectra strong peaks can be shown at 12–16Hz between GPI and both motor and visual cortex in NREM2, corresponding to the sleep spindle frequency range. Further, a distinct peak could be identified in the higher alpha-range (10–12Hz). In contrast, in NREM3/4 the 12–16Hz-coherence peak was not present while the 10–12Hz-peak remained. For NREM2 and NREM3/4 the 10–12Hz peak disappeared after partialization of the surface EEG with GPi as the predictor. The directed transfer function (dtf) revealed that the peak at 12–16Hz is predominantly propagated from the visual cortex to GPi and motor cortex in NREM2 and NREM3/4. The dtf for GPi to motor and visual cortex showed a distinct peak in the higher alpha-range (10–12Hz) in NREM2, while in NREM3/4 distinct peaks at 10–12Hz were present between GPi and motor cortex as well as from motor to visual cortex. Principally, the same coherence patterns were also present in STN patients, however at a much lower level and less homogeneously than in GPi patients. There was no significant coherence in REM between STN or GPi and cortex. We can show functional connectivity between motor nuclei of the basal ganglia and both the motor and the visual cortex in NREM sleep in the frequency range of sleep spindles. The motor system seems to be additionally synchronized with a peak in the higher alpha-range (10–12Hz), propagated through the GPi to the cortex. Most likely, the synchronization at 10–12Hz can be regarded as a specific signature of the motor system in NREM sleep, which parallels the overall synchronization at 12–16Hz.