In this study, we found that acute phase shifts selectively affect recall of the hippocampal-dependent contextual fear conditioned behavior, regardless of whether we phase shifted the mice before or after training. Our findings are consistent with previous studies on phase shifts and cognition 
, but one critical difference is that we applied the phase shift only once to mice that were naïve to such manipulations. Hence, we were able to demonstrate that a rapid shift in the lighting cycle produces a dramatic reduction in recall without a significant effect on acquisition ( &
). The duration and severity of jet lag depends on the number of time zones crossed. For the circadian system, the larger the phase shift, the longer the duration required for re-synchronization to the new lighting schedule. For example, several studies suggest that the circadian system would require more than 6 days to recover from the 6 hr phase advance used in the present study 
. We were able to demonstrate that the larger the phase shift, the larger the impact on recall () with even a 3 hr phase advance having some impact on recall compared to untreated controls. Using this same behavioral assay, we previously found that a mutation in one of the key clock genes (Period2
) as well as the loss of vasoactive intestinal peptide, a signaling molecule critical for coupling within the central clock, reduced recall, but not acquisition, of conditioned fear 
. Collectively, our findings are consistent with a role for the circadian system in the consolidation of memory.
Several lines of evidence indicate that phase advances of the LD cycle are more disruptive than phase delays. In general, an organism's behavioral activity-rest cycle can re-synchronize to a phase delay of the LD cycle rapidly while synchronization to a phase advance is much more gradual. For example, in mice, re-synchronization to a 6 hr phase delay occurs within a couple of days, while re-synchronization to a 6 hr phase advance may take 5–6 days 
. In older mice, repeated phase advances can increase mortality, an effect not seen with phase delays 
. These studies suggest that phase advances may be more disruptive to cognitive processes than phase delays. In the present study (), we found that both advances and delays disrupted the recall of the conditioned fear. The impact of the phase advance was larger than the phase delay at all time points tested, so it is possible that future work will find more robust differences. Earlier work with rats also found that both advances and delays of the LD cycle disrupted memory 
. Perhaps the difference between advances and delays on cognitive processes lies more in the duration of the disruption than its magnitude.
One downside of using the 12 hr phase shift ( &
) as a drastic disruption of the circadian system is the possibility that the re-entrainment could take place via phase advances or delays. Our series of different durations of phase shifts described in confirmed that although the 12 hr phase shift has the most disruptive effect on memory, 6 and 3 hr shifts also have a significant impact on recall of fear-conditioned behavior, and the 6 hr protocol was hence used in all subsequent mechanistic experiments to allow interpretation of the effects of direction of shift and other factors that could affect memory. A further possible confound of the post-training phase shift experiment described in is the difference in the lighting conditions between training (light phase) and testing (dark phase post-shift). We have previously shown that recall is higher in the day than in the night 
. While we cannot rule out some direct effect of dark reducing recall in this one experiment, most of our experiments were carried out with shorter phase shifts in which both training and testing were carried out under the same lighting conditions that continued to affect recall (e.g. &
). Furthermore, the experiment described in shows no difference in freezing between the non-shifted animals tested at non-24 hr intervals in the dark (18 hr) and light (30 hr) phase. Similarly, the mice subjected to a phase shift prior to testing do not show a difference in freezing between the non-24 hr interval testing phases in the light (18 hr) and dark (30 hr). For these reasons, we do not feel that the acute effects of lighting conditions were an interpretational problem for these studies.
Peak performance still occurs 24 hrs after training in the phase shifted group
Behaviorally, there is a long history of work demonstrating that peak performance in the recall of a number of behavioral tasks varies with time of day 
and circadian time 
. This type of research has led in most behavioral learning protocols to keep the interval between training and testing at 24 hrs. This prior work also raises the possibility that the 6 hr advance in the LD cycle induced an immediate 6 hour shift in the peak of recall. If this were the case, then the peak of recall would be 18 hrs after training in the phase advanced group while remaining at 24 hrs after training in the control group. We examined this possibility by training mice that were on a stable LD cycle and then testing them at 18, 24, and 30 hrs after training (). The control mice showed a clear peak of recall of training 24 hrs after training, confirming prior work. Interestingly, the phase advanced cohort also showed a peak in recall 24 hrs after training. The 6 hr advance did not shift the peak in performance to 18 hrs after training. Therefore, the “time-stamp” of 24 hr for peak recall was not affected by phase shifts, and confirmed that the reduced recall we observed after a phase shift is not due to a shift in the timing of the peak recall.
The jet lag protocol alters the magnitude of the stress response but not baseline levels of corticosterone
Stress and the release of corticosterone is an important modulator of learning and memory 
. With contextual fear conditioning, increasing corticosterone can facilitate consolidation 
or interfere with recall 
. Corticosterone is a hormone secreted with a robust circadian rhythm, with peak secretion during the late day, ~ ZT 10, in nocturnal rodents 
. Anatomical studies have provided evidence that the paraventricular nucleus (PVN) receives innervations from the SCN. Release of corticotrophin releasing factor by neurons within the PVN is the critical step in stimulating adrenocorticotropic hormone (ACTH) release from the pituitary and thus the activation of the hypothalamic-pituitary-adrenal axis 
. SCN-lesioned rats show a loss of daily rhythm in ACTH and corticosterone 
. In this study, we determined the impact of the 6 hr phase advance on the levels of corticosterone in the mice. While we only sampled at one time of day, we did not see evidence that unstimulated, baseline levels were increased by the phase shift (). In contrast, the stress (foot shock) evoked responses were significantly larger in the phase shifted group. So it is possible that higher corticosterone levels during recall played a role in the reduced memory in the phase shifted groups. In flight crews who habitually experience travel between more distant time zones, there is evidence for both higher salivary cortisol and reduced performance of vigilance tasks 
The jet lag protocol alters the temporal distribution but not total amounts of sleep
Sleep immediately after a training session has been shown to be critical for consolidation of contextual fear conditioned memory 
as well as many other learned behaviors 
. In humans, sleep disturbances are a common complaint after jet travel crossing a number of time zones 
. To examine the possibility that the 6 hr phase advance caused sleep deprivation in mice, we examined pre- and post-phase shift sleep/wake patterns using EEG recording in freely moving mice. We found no significant differences in the amount of NREM or REM sleep before and after the phase advance (). Surprisingly, we could not find other studies that had examined the impact of experimental jet lag on sleep in mice. In rats, there has been one report that phase advances of the LD cycle led to an increase in NREM and REM sleep 
. Our phase advance protocol results in one shorted day, and it has been shown that rats and hamsters housed under short photoperiod (8
16 LD) show altered sleep patterns but the short photoperiod does not affect sleep homeostasis 
. To further explore the sleep/wake patterns, we turned to behavioral measures of sleep 
. We measured the patterns of sleep/wake before and after a 6 hr phase advance. The results () clearly show a change in the temporal distribution of sleep but do not show an overall loss of sleep. Thus the impact of jet lag on recall occurred without producing sleep deprivation. Future studies will need to explore the relationship between misalignment of sleep on memory consolidation.
Experience can reduce the impact of jet lag on the conditioned fear
As a final experiment, we tried to further disrupt the circadian system by subjecting the mice to repeated phase shifts, but discovered that prior experience of phase shifts appears to ameliorate the adverse effects on recall. This observation may explain some apparent contradictions in the literature. A previous study by Craig and McDonald showed that chronic or serial jetlag in rats impairs acquisition, but in contrast to our findings, chronically and acutely jetlagged rats did not appear to show deficits in recall of contextual fear conditioned freezing 
. This discrepancy could be due to differences in application of “acute” phase shifts. In our study, all phase shifts were acutely applied to mice naïve to phase shifts, whereas Craig and McDonald applied serial phase shifts over several days to produce their acute jetlag model. In fact, their study agrees with our finding that multiple serial exposures to phase shifts can compensate for the negative effects of acute phase shifts on memory. The data suggest that it is possible to design treatments that can reduce the cognitive impact of circadian de-synchronization.
Phase shifts desynchronize the network of circadian oscillators: mechanisms
Previous studies have shown that when rodents are subjected to acute phase shifts of the LD cycle, de-synchrony results within core clock genes within the SCN 
, between different regions within the SCN 
and between the SCN and peripheral oscillators 
. Within circuits involved in learning and memory, it has been demonstrated that the amygdala takes longer to re-entrain to phase shifts of the LD cycle than the SCN 
. Nuclei within the amygdala (central and basolateral) and as well as the dentate gyrus region of the hippocampus exhibit rhythms in gene expression which are dependent on an intact SCN 
. The hippocampus also exhibits rhythms in clock gene expression 
that are independent of the SCN 
. By applying an acute phase shift, we are most likely uncoupling the tightly synchronized network of circadian oscillators, including regions of the brain responsible for learning and memory. We speculate that this disruption in the coordination of clock gene expression within different neural structures lies at the heart of memory deficits.
Consolidation of memory involves changes in gene expression 
and is prevented by inhibitors of transcription and translation. The molecular circadian clock regulates the temporal pattern of transcription and we believe that by this mechanism, disruptions in the molecular clock could also disrupt consolidation of memory. Previous work has also found evidence that levels of adenylyl cyclase 1 expression 
as well as cAMP and MAPK activity in the hippocampus 
exhibit daily oscillations. Previous work in Aplysia
implicates the circadian gating of the MAPK pathway as the mechanistic control point for circadian regulation of sensitization 
. These results also raise the possibility that jet lag evoked disruptions in intracellular signaling pathways may be an important part of the observed deficits in recall.
Conclusions and Significance
In the present study, we demonstrate that single acute phase shifts can reduce recall of a learned behavior, presumably through altering memory consolidation. Among other novel findings, we demonstrate that the 24-hr interval between training and testing still produces the strongest recall even in phase shifted mice. We were able to disassociate the impact of the circadian disruption from the total amount of sleep as the mice were not sleep deprived. The temporal distribution of sleep was disrupted and future studies will need to explore the importance of when
sleep occurs on memory consolidation. Our data adds to a body of studies that have shown that a functioning circadian system is important for long-term memory. Memory deficits have been found in several lines of mice with mutations impacting the generation of robust circadian rhythms in behavior 
. Similarly, environmental manipulations, including chronic phase shifts of the LD cycle, that disrupt circadian rhythms without genetic mutations also disrupt memory in different tasks 
. We think that the broader hypothesis that internal desychronization of a network of circadian oscillators results in memory deficits is clinically important. Patients with a variety of psychiatric and neurological disorders exhibit disruptions in their sleep and circadian rhythms. If our hypothesis is correct, these circadian disruptions may contribute to the cognitive symptoms experienced by a range of patients.