Stress has significant effects on sleep in humans as well as rodents [28
]. Rapid eye movement sleep (REMS), in particular, changes dramatically after fear conditioning (FC) [28
]. The effects of FC on REMS have generally been assessed using standard measures of REMS macroarchitecture, i.e., number and average duration of REMS episodes and total amount of REMS time. However, the results of several studies demonstrate the importance of investigating REMS microarchitecture as well, by partitioning total REMS time into sequential REMS (seq-REMS, inter-REMS episode interval ≤ 3 min) and single REMS (si-REMS, inter-REMS episode interval > 3 min) [2
]. Episodes of seq-REMS, which are on average of shorter duration than those of si-REMS, occur in clusters. Studies have demonstrated that various stressors affect seq-REMS specifically [2
]. For example, Amici et al. [2
] found an increase in the number of seq-REMS episodes in the recovery sleep of rats after exposure to cold stress. Studying REMS microarchitectural changes following FC in rats can yield important new insights into the mechanisms of stress-induced sleep disturbances [7
The REMS response to stress differs among rat strains. Wistar-Kyoto rats (WKY), known according to a range of criteria to be stress-sensitive [22
], respond to FC with fragmentation of REMS, defined as a shift in the distribution of seq-REMS and si-REMS episodes towards seq-REMS [7
]. This response is in contradistinction to the responses shown by Sprague-Dawley and Wistar rats (WIS) [7
], in which a preponderance of longer duration si-REMS episodes following FC suggests that FC is less disruptive to REMS.
Mechanisms of REMS fragmentation, i.e., the neurophysiological substrates of a form of REMS that occurs in short duration episodes separated by short duration intervals, require explanation. Electroencephalographic activity in the gamma frequency range, which some suggest to be associated with focused attention during waking (W) [4
], also is prominent during REMS [4
], a state of hyperalerting to stimuli of endogenous origin [21
]. In an initial attempt to explore the relationship of gamma power to REMS [14
], we demonstrated that relative gamma power (measured as percent of total power) during REMS was lower in WKY compared to WIS, both before and after a FC procedure. However, there was no effect of FC on gamma power in REMS, an effect that would have been expected if gamma power signals brain mechanisms serving to maintain continuous, focused REMS, and the specific neural interactions that characterize the beginning and ending of a REMS episode [10
]. Bassi et al. [3
] have emphasized the importance of looking at the transitions into and out of REMS, suggesting that at these time points REMS is most likely to fail initially to consolidate properly or to terminate early, respectively. Therefore, we examined gamma power changes at the transitions into and out of REMS.
Animals from our previous study [7
] provided data with which to explore the effects of FC on gamma range activity during REMS transitions in WKY compared to WIS. WKY and WIS underwent a cued FC procedure in that study. They were presented with ten tones (800 Hz, 90dB, 5 s duration), each co-terminating with a mild electric foot shock (1.0 mA, 0.5 s duration) at 30 s intervals. The effects of FC on sleep were examined by analyzing electroencephalographic and electromyographic traces obtained in the light phase (11 AM to 3 PM) after animals were re-exposed to three tones, without shock, one day, and again 14 days, after FC. Only WKY demonstrated an increase in REMS fragmentation, defined as a shift in the distribution of seq-REMS and si-REMS toward seq-REMS [7
]. Also, WKY continued to freeze to tone presentations until the end of the test period, 14 days, whereas WIS did not. To determine that the alterations in REMS microarchitecture were specific to FC and not due to a lasting effect of shock stress (SS) alone, the effects of foot shock alone were studied in an additional group of animals. DaSilva et al. [7
] reported that SS had no effect on REMS microarchitecture in either strain, suggesting that the changes in REMS microarchitecture were due to the FC procedure and not a residual effect of SS.
To understand better the mechanisms of REMS fragmentation in the WKY strain, we examined, in the present study, changes in relative gamma power in the electroencephalogram (EEG) during those times when REMS was beginning (35 s before to 105 s after REMS onset (ARO)) and terminating (85 s before REMS termination (BRT) to 35 s later). We hypothesized that relative gamma power at REMS onset and termination would be differentially altered in WKY compared to WIS following FC, with decreases in relative gamma power both ARO and BRT in WKY only. This would suggest different alterations of neural activity in REMS after FC in WKY compared to WIS. Changes in relative gamma power could be a useful measure of changes in the neural activity underlying REMS.