In humans, chronic exposure to light at inappropriate times of day can lead to sleep and circadian disorders, cognitive defects and depression. In this regard, the impact of ultradian light–dark cycle, as experienced with shift work, remains poorly understood. Here, we show that ultradian light–dark cycle induces alterations of cognitive and mood functions that are not due to direct photic effects but are relying on a non-physiological regulation of the circadian clock delaying the circadian sleep phase.
First, we confirmed that exposure to such ultradian light cycle, even if it does not induce an abolishment of the circadian function, causes a period lengthening (25-h) leading to a day-by-day phase-delay of circadian rhythms of both general activity and sleep18, 21. This period lengthening results in a complete inversion of the circadian phase after 12 days of T7 cycle exposure, and seems particularly non physiological given that the endogenous circadian clock period is lower than 24 h in mice22. These effects seem to rely both on direct effects of light on behavior and on light-induced circadian modulation.
Indeed, we observed under T7 cycle a direct sleep-promoting effect of light and a wake promoting effect of darkness, corresponding respectively to the ancient terminology of negative and positive masking17. In addition, under T7 cycle we also found out a short-component 7-h long rhythm of activity, with higher activity scores under darkness and conversely higher amounts of sleep during light pulses. In that respect, according to our data, the repeated alternation of light and dark pulses of short duration was sufficient to modulate activity levels but also sleep–wake distribution throughout the entire T7 cycle exposure. These results are in contradiction to those previously published by Samer Hattar’s group23, who showed no effects of light on sleep during T7 cycle (possibly explained by the large age range of mice used in their study; n = 4, from 4 to 12 months), although they described a significant effect on wheel-running activity. Furthermore, our findings are consistent with multiple data from the literature that reported the somnogenic effect of light pulses during ultradian cycles of different durations20, 24,25,26,27. In addition, one can assume that, if the direct effects of light were the only factor to modulate behavior under T7 cycle, activity scores and sleep amounts would rely on real light conditions only (ie on 3.5 h-long light and dark pulses), and a progressive circadian desynchronization would be observed due to these direct effects.
In sum, our data highlight that, under T7 cycle, activity and sleep modifications do not depend on direct effects of light only. Indeed, the findings reveal that the ultradian T7 cycle is also able to entrain the circadian system over a lengthened 25-h long period. The maintenance of circadian variations explains that the difference between activity amounts observed under light and dark exposures was milder under T7 cycle than under T24 cycle, given that activity levels remained overall higher during subjective nights than subjective days. In the same way, the greatest light-induced promotion of sleep appeared in the trough of sleep circadian rhythm (corresponding to the T7-subjective night), when the circadian drive favors waking rather than sleep, whereas the effects of light–dark alternation were attenuated during the periods of long-duration of sleep (corresponding to the subjective day). More than a “ceiling effect” of the sleep-promoting direct influence of light, the data suggest that the overall effect relies on the circadian drive getting the upper hand over waking-promoting direct effect of darkness. Last, when animals are exposed to DD after T7 cycle, the endogenous period calculated over 7 days was very close to 24 h, which is higher than classically described endogenous activity period of C57Bl/6 mice28, and the circadian phase of activity was delayed by approximately 12 h. This delay corresponds to the 1-h period lengthening observed for the 12 preceding days under T7 cycle and demonstrates that T7 cycle, besides inducing direct light effects, modifies activity and sleep through the entrainment of circadian clocks with longer circadian period20, 24,25,26,27.
Given the influence of T7 cycle on circadian rhythmicity of activity and sleep, we aimed to assess whether T7 cycle induces memory and mood alteration and whether those resulted from direct effects of light and/or from these modifications of the circadian organization. In the spatial novelty preference task, memory abilities were similar for animals tested during the day or the night under T24 cycle. Circadian modulation of learning and memory with time of day has been widely studied across species ranging from invertebrates to humans but no consensus has been found, since results vary depending on the task, the memory process or the type of memory long-term or short-term. term, for a review29. Although, our data were in accordance with some of the studies that showed no effect of circadian phase on memory abilities, especially when light condition is controlled throughout the behavioral testing30, 31, as in our experiments. Animals exposed to T7 cycle showed deficiency to recognize the familiar arm when tested during subjective night. This memory disruption was in accordance with the one observed by Legates et al.18 with object recognition and Morris water maze tasks after two weeks of T7 cycle exposure, although the information on the time of testing is missing in their publication. However, given the phase inversion observed after 12 days of T7 cycle, this one time point evaluation is not sufficient to conclude that aberrant light impairs learning through direct effects independent of the circadian system, as assumed by Legates et al.18. If memory disruption observed in T7 group was due to a direct effect of light, animal’s performances in T-maze should also be altered when assessed during subjective day. Yet, our results showed that memory abilities were unaffected at this circadian time point. These results, besides the preservation of memory performances observed after 2 h of light exposure at early night under T24, demonstrate that memory impairments observed in mice exposed to ultradian light cycle are not due to direct effects of light, contrary to the assumption of LeGates et to the.18but rather rely on non-physiological regulation of the circadian clock, which is in accordance with human data showing that the non-circadian direct influence of light improves alertness and cognitive performances1, 32, 33.
In the same way, the modification of mood-related behavior induced by exposure to ultradian light–dark cycle is also dependent on circadian timing. Indeed, in the FST, behavior was modified only during subjective night in mice exposed to T7 cycle, this group showing less time spent immobile compared to the three other groups. These results suggest that prolonged exposure to an ultradian light–dark cycle affects the reaction to stress, dependently of circadian-time, and could particularly modify the ability to cope with an acute inescapable stressor34, 35. The decrease of immobility time of the “T7-subj. night” group does not seem to depend on the greater amount of activity during subjective night given that it is not observed during the night under T24 cycle. This time-dependent effect could rely on a circadian rhythm-gated pathway, preferentially conducting light information to brains involved in mood regulation at certain moments of the day/night cycle36. Several studies using clock genes mouse mutants have shown circadian periodicity modulations along with mood-related behaviors alterations: especially, mice with a lengthened circadian period of activity showed a decreased immobility time in the FST whereas the inverse was observed in mice showing a shortened circadian period (for a review, see37). Then, lengthening the circadian period of activity seems to be associated with behaviors mimicking the one observed in manic state of bipolar patients. Of course, is our case, the use of the FST alone does not allow us to infer the same interpretation. However, our results demonstrate that exposure to T7 cycle promotes circadian variations of mood that are not due to a direct effect of light.
Among limitations of the present study, we choose not to monitor corticosterone levels as a previous report showed no difference between T7 and T24 cycles at time points that are close to the ones we used for behavioral evaluations18. Additional behavioral evaluations would be useful to assess the effect of T7 cycle exposure on other functions, including anxiety-like behaviors, and to conclude on the potential appearance of manic-like state. Furthermore, the same experiments should be performed with female mice in order to evaluate the influence of sex. Finally, it would also be interesting to test the effect of other ultradian light cycles, with various durations of light and dark pulses, on cognition and mood. However, the main strength of this study is in its design integrating sleep and circadian data along with behavioral evaluations at different circadian time points. This research design allowed us to disentangle the mechanisms, direct versus indirect (affecting the circadian function), by which ultradian light cycles affect behavior.
Therefore, clock period lengthening and chronic day by day phase delay promotes the emergence of circadian variations of cognitive functions and mood and drives to memory inability and modifications of strategies adopted in an aversive situation at certain times of day. This study helps to better understand cognitive and mood alterations associated with circadian disorder, especially delayed sleep phase disorder38or non-24 h sleep–wake disorder as well as those induced by shift working39. A better understanding of these mechanisms is crucial for improving the management and patient care of circadian disorders, and for optimizing the management of light exposure in today’s society.