High gamma activity modulated by the theta rhythm in the human anterior thalamus at rest

Introduction: Modulation of the activity of local neural assemblies firing in the gamma frequency range by slower theta frequency oscillations over longer spatial ranges is thought to facilitate communication and neural plasticity (Canolty and Knight, 2010). Such cross-frequency coupling (CFC) has been observed in the neocortex, hippocampus, and anterior thalamic nucleus (ATN), varying according to the task performed (Axmacher et al., 2010; Canolty et al., 2006; Sweeney-Reed et al., 2014). The thalamus has been proposed to play a role in switching between cognitive states (Di et al., 2013; Greicius et al., 2003). Here we investigated whether thalamic CFC arises only during performance of tasks involving an external attentional focus or whether it is also detectable at rest. Methods: We had the rare opportunity to record intracranial electrophysiological data directly from the ATN and dorsomedial thalamic nuclei (DMTN) from patients receiving intrathalamic electrodes implanted for deep brain stimulation for pharmacoresistant epilepsy. Data were recorded using a Walter Graphtek amplifier from 8 contacts (4 each side) against a nose reference, sampled at 512 Hz during rest, memory encoding (Sweeney-Reed et al., 2014), and a novelty oddball paradigm (Zaehle et al., 2013). Bipolar re-referencing rendered bilateral ATN channels for 7 and DMTN channels for 6 patients. 30 s of artifact-free data were analyzed from the two tasks and during rest. The data were decomposed using a 6-cycle Morlet wavelet, and theta (4-8 Hz) phase and gamma (33-203 Hz) amplitude series extracted. Theta phases were combined with the HG envelopes to generate composite signals, the absolute value of whose mean provided a modulation index (MI) (Canolty et al., 2006; Onslow et al., 2011). Significance was assessed against a distribution of MIs generated by shuffling the order of subsections of the amplitude series to eliminate temporal relationships between the two frequency bands. Results: CFC between theta (4-6 Hz) phase and high gamma (HG) (80-150 Hz) power was identified in the left ATN during rest in all 7 patients (p 0.05). Theta-HG CFC was only seen in 1 of the 6 patients who performed the memory encoding task and 1 of the 6 who carried out the novelty oddball task (not the same patient). On direct comparison, the CFC pattern differed according to the task performed (rest versus encoding: 2-D KS test: K-statistic = 0.75, p = 0.00003; rest versus novelty oddball: K-statistic = 0.61, p = 0.0068). The theta phase at which HG power peaked differed in 16 of 21 pairwise, between-patient comparisons (two-sample Kuiper test: p < 0.0024 – Bonferroni-corrected threshold for criterion q = 0.05). Over 5 s time windows for each patient, the mean angle at which HG peaked in each window varied less than 45° in at least 4 of the 6 windows. Conclusions: HG activity was coupled with theta phase in the ATN consistently across patients at rest. This pattern was not observed during either of the tasks involving an external focus of attention. The inter-patient variation of the phase at which HG peaked is likely to be due to individual differences in cognitive processing in an unconstrained resting state. Our findings are consistent with theta-HG CFC being an ongoing process in the ATN, which is modified when specific tasks, involving an external focus of attention, are performed. Modulation of theta-HG CFC provides a potential mechanism by which the thalamus could be involved in switching between cognitive states (Di et al., 2013; Greicius et al., 2003)