Abstracts

Rationale: Single-pulse electrical stimulation (SPES) of cortex can evoke electrophysiological responses at distant sites, including phase-locked potentials called cortico-cortical evoked potentials (CCEP), or non-phase-locked cortico-cortical spectral responses (CCSR). CCSRs occur in a variety of frequencies, including high gamma (HG) bands, which are widely thought to index changes in neuronal population firing rates. Thus, CCSRHG may reflect changes in neural firing rates evoked via direct or indirect anatomical connections with the stimulation site. Propagation through these pathways may be changed during sleep. Here we investigated whether different brain states (e.g. sleep stages) affect the propagation of HG activity evoked by SPES. Methods: Twelve patients underwent invasive presurgical evaluation for intractable partial epilepsy, using subdural electrodes (IRB No.443 in Kyoto University). Among them, repetitive SPES at alternating polarity was delivered to the cortical surface in frontal (7 sites in 6 patients) or parietal lobe (5 sites in 3 patients) during wakefulness, light sleep, slow wave sleep, and REM sleep. Channels showing significantly increased CCSRHG across any sleep stage were considered to be involved in the network affected by stimulation. Channels showing significantly decreased CCSRHG right after stimulation were removed, not expecting much propagation from decreased CCSRHG. Then, we applied event-related causality (ERC) analysis based on short-term direct directed transfer function (SdDTF) to estimate the dynamics, directionality, and magnitude of propagation of neural activity from sites with CCSRHGs with an increase in high gamma power. Finally, we separated the channels into three regions: frontal, parietal, and temporal-occipital lobe, and compared the neural propagation among the regions’ networks across sleep stages. Results: ERC analyses revealed different patterns of neural activity propagation evoked by SPES applied to different anatomical regions. These patterns were also dependent on sleep stages. For example, compared to wakefulness stimulation of frontal lobe during slow wave sleep elicited larger propagation towards parietal lobe at latencies between +200 and +300 ms. On the other hand, stimulation of parietal lobe elicited smaller propagation towards frontal lobe in the same period. Furthermore, propagation within frontal lobe was smaller during REM sleep than during wakefulness at latencies between +100 and +200 ms. Conclusions: Stimulus-induced propagation of neural firing across human brain networks is modulated by global brain states, including different sleep stages. Greater propagation from parietal to frontal regions during wakefulness, and from frontal to parietal regions during slow wave sleep, may reflect processes related to memory encoding and consolidation. Increased propagation within frontal regions during wakefulness, relative to REM sleep, suggests that computation within the frontal lobe is reduced during REM sleep, possibly leading to bizarre and disorganized experiences in dreaming. Funding: National Institute of Neurological Disorders and Stroke (R01NS091139), Kyoto University Foundation, the Japan Epilepsy Research Foundation,MEXT KAKENHI (15H05874, 16K19510, 17H05907, 26282218),JSPS KAKENHI (17K16120)