Levichkina, E.; Vidyasagar, T. R.; French, C.

Several clinical attempts to use deep brain stimulation applied to nucleus basalis of Meynert to reduce cognitive decline in different types of dementia report inconclusive effects. At the same time, experiments in rodents have largely demonstrated cognitive improvement. The hypothesised basis for this difference was the application of different patterns of stimulation - tonic continuous over multiple sleep-wake cycles in human clinical trials vs intermittent bursts over shorter intervals in animal studies. However, since no systematic testing of the effects of different stimulation patterns applied in different states of vigilance was conducted, it remains unclear what specific attributes of cortical responses are associated with these patterns, and the effect of vigilance state on pattern response. We hypothesised that communication between nucleus basalis and its cortical targets involves frequencies in both low and high ranges, making burst-type intermittent patterns potentially more effective. Therefore we studied preferred frequencies of neuronal synchronization within nucleus basalis as well as between nucleus basalis and its cortical targets in B6 mouse strain (N=9, age 4-6 month) by recording neuronal spiking and local field potentials (LFP) from 2 sites of the nucleus basalis and 4 cortical sites simultaneously across multiple sleep-wake cycles. Spike-field coherence (SFC) analysis revealed natural synchronization tendencies of nucleus basalis - all animals demonstrated significant SFC at delta-theta range and gamma coherence peaking between 57 and 88 Hz (SFC > 95% Jackknife confidence interval at frequency range > 2*bandwidth). SFC between cortical cells and local field potentials of nucleus basalis, which reflects cortical feedback, peaked at delta-theta frequencies. Based on these results we constructed "nested" patterns of stimulation that included both delta-theta and individually assessed gamma peak for each animal (N=8). The effects of nested stimulation were compared to the ones of random and tonic (20Hz) stimulation patterns. These patterns of stimulation were applied across multiple sleep-wake cycles to freely behaving animals, and spectral power density in response to stimulation compared between patterns for different states of vigilance. We report significant reduction of delta power during slow wave sleep and theta power in REM for all three patterns. Slow wave suppression was also present during wakefulness. This indicates all patterns may improve levels of alertness but highlights strong likelihood of sleep disturbances in response to stimulation during sleep. Thus, deep brain stimulation of nucleus basalis may be counterproductive during sleep. The nested pattern of stimulation alone demonstrated preservation or even enhancement of gamma activity in the cortex while suppressing low frequencies. Broad-band suppression of cortical activity was shown in response to tonic stimulation, and to some extent to the random pattern as well, suggesting that tonic stimulation may be unsuitable for deep brain stimulation of the nucleus basalis of Meynert.