Regulatory, Integrative and Comparative Physiology

Circadian system of mice integrates brief light stimuli

Anthony N. Van Den Pol, Vinh Cao, H. Craig Heller


Light is the primary sensory stimulus that synchronizes or entrains the internal circadian rhythms of animals to the solar day. In mammals photic entrainment of the circadian pacemaker residing in the suprachiasmatic nuclei is due to the fact that light at certain times of day can phase shift the pacemaker. In this study we show that the circadian system of mice can integrate extremely brief, repeated photic stimuli to produce large phase shifts. A train of 2-ms light pulses delivered as one pulse every 5 or 60 s, with a total light duration of 120 ms, can cause phase shifts of several hours that endure for weeks. Single 2-ms pulses of light were ineffective. Thus these data reveal a property of the mammalian circadian clock: it can integrate and store latent sensory information in such a way that a series of extremely brief photic stimuli, each too small to cause a phase shift individually, together can cause a large and long-lasting change in behavior.

  • phase shifts
  • biological clock
  • entrainment
  • suprachiasmatic nuclei
  • activity rhythms

most animals have daily rhythms of behavior and physiology. These rhythms free run with periodicities close to 24 h (circadian) in the absence of environmental cues. Circadian rhythms are generated by endogenous pacemakers that can be entrained to the 24-h day by external geophysical cues. Entrainment by external cues involves phase shifting of the rhythm of the endogenous pacemaker to bring it into synchrony with the geophysical cycle. The most important entraining stimulus is exposure to light. For example, in most mammals that have been studied, exposure to light in the early dark period will phase delay the pacemaker and exposure to light in the late dark period will phase advance the pacemaker (8).

In mammals, the circadian pacemaker resides in a group of hypothalamic neurons, the suprachiasmatic nucleus (SCN) (10, 17), that receives direct connections from retinal ganglion cells (6, 11). In this study we investigated the sensitivity of the circadian system of mice to multiple extremely brief photic stimuli to determine if the system will integrate successive brief stimuli and thereby interpret them as being equivalent to a much longer, continuous exposure to light.


Pigmented adult C57/Bl6 male and BALB/c albino female mice were housed in individual clear plastic cages in the same room and fed food and water ad libitum. Light flashes were generated by a DynaLite Flash 2040 driven by a DynaLite M1000X supply at a distance of 10 feet from the mice. Mice were left in their home cage during light stimulation to reduce any complication in data interpretation that might arise if mice were handled and transported to a stimulation chamber. Light intensity of 2-ms flashes was measured with a Graseby Optronics photometer with sensitivity maximum at 500 nm light. A mean light of 1 μJ/cm2 was measured in the animals’ cages. Rods recover from light slower than cones. Based on constants derived elsewhere (9), this flash would result in about 7,000 photoisomerizations ⋅ rod−1 ⋅ flash−1, which would allow recovery in a few seconds. For comparison, light of 30,000 photons/rod would allow recovery in ∼10 s, or 106 photons in ∼1 min (14).

Each cage was equipped with a running wheel. Revolutions of the wheel were detected with a reed switch coupled to a computer in another room. Activity was monitored continuously with DataQuest computer software from Mini-Mitter and analyzed with Circadia software. Behavioral data were divided into 10-min bins for analysis and display.

The animals free ran in constant darkness for several weeks before exposure to light, and they were therefore out of phase with respect to each other at the time of the experiment. The experimental protocol was to expose the entire population to simultaneous photic stimulation consisting of either 300 2-ms flashes of light that occurred at 1-s intervals for 5 min (total light duration = 600 ms) or 60 2-ms flashes of light that occurred at 5-s intervals for 5 min (total light duration = 120 ms). In addition, to determine if the phase shift due to multiple light flashes may have been due simply to activated photoreceptors that took several seconds to recover (9), we examined the effect of 60 2-ms light flashes occurring at intervals of 1 min for 1 h (total light duration = 120 ms). Because the free running rhythms of the mice were out of phase with each other, the photic stimulation was experienced by different animals at different phases of their circadian rhythms, enabling us to construct a phase response curve for the population by plotting each animal’s phase response as a function of its circadian phase at the time the stimulus was presented.


The mice (n = 86) were held under conditions of constant dark for periods of 1–7 mo. Under these constant environmental conditions, mice demonstrated circadian rhythms with a mean period of 23.5 h ranging from 23.2 to 24.3 h. Most individual mice showed characteristic circadian cycles with periods that varied from one day to the next by only a few minutes per day. Based on wheel size and revolutions per day, most mice ran 2–4 miles/day. Some long-distance runners ran as many as 15 miles during their active periods.

Each of the three multiple photic stimulus regimens described above produced phase shifts of up to several hours in the circadian wheel running rhythms (Fig. 1). These were permanent phase shifts that lasted as long as the rhythms of individual animals were followed or until the animal experienced another phase shifting stimulus. Both male and female mice showed changes in their circadian rhythms in response to these brief light stimuli.

Fig. 1.

A: an example of a plot of mouse wheel running activity before and after a series of 2-ms light flashes spaced at 5-s intervals for a total of 5 min (total duration light exposure = 120 ms). This male mouse shows a phase delay. For ease of viewing and interpretation, the data are double plotted, with each 24-h period printed twice in succession on the same horizontal line.B: 2-ms pulses at 5-s intervals for 5 min can generate a large phase advance. This record is from a female mouse. C: typical example of a male mouse, showing little change in pattern of circadian rhythm with a single 2-ms light flash, but several weeks later showing a substantial phase shift in response to a 2-ms flash every 1 min for 1 h. ○, Time of photic stimulation; circle is placed within the day of light stimulation. Straight diagonal line indicates the computer-generated start of activity for the circadian period.

In control experiments we found that single flashes of 2-ms duration were ineffective in causing phase shifts. This is similar to the previous finding in hamsters (12) that single pulses of 3 ms duration were ineffective in causing phase shifts. These data indicate that the circadian system is capable of integrating extremely brief light pulses falling as much as 1 min apart so as to interpret them as being the equivalent of continuous, long-duration photic stimuli.

The effect of photic stimuli on circadian rhythms is typically investigated by exposing animals to flashes of light usually of durations ranging from minutes to 1 h at different times in the circadian cycle. Plotting the resulting phase shifts as a function of time of light exposure produces a phase response curve. A typical photic phase response curve for an animal held in constant dark would show no phase shifts during the subjective day, phase delays during the early subjective night, and phase advances during the late subjective night.

When the phase responses of all of the mice were plotted as a function of the phases of their rhythms when they received the stimulation, a typical mouse photic phase response curve resulted (Fig.2 A). During the period from circadian time (CT) 0000 to 1200 light flashes had little effect, consistent with the previous demonstration of this light-insensitive period of the circadian cycle (5). Between CT12 and CT18, light flashes caused a robust phase delay. Between CT18 and CT24 a modest phase advance was generated by the light flashes (Fig.2 A). Each species has its own typical phase-response curve (8). The one shown in Fig.2 A is typical for mice, where the phase advance is of smaller magnitude than the phase delay (3, 8). In Fig. 2 B a substantial phase delay was generated by one 2-ms flash every 1 min for 1 h; little phase advance is seen. This lack of phase advance could be due to the small sample size during the critical phase advance period or to a decreased phase-advance sensitivity to the brief intermittent light stimulation.

Fig. 2.

A: combined data for several experiments, including both the stimuli at 1 flash and 5 flashes/5 s. Endogenous circadian time (CT) for each animal was defined on the basis of the start of running wheel activity being CT12. Light flashes caused either phase delays or modest advances depending on the mouse’s CT. Each point represents the phase shift in the circadian rhythm, comparing a control period 7 days before the light flash with the 7 days after the light stimulation. Little effect of light was found between CT0 and CT12, the subjective day for these nocturnal rodents, a time when they are inactive. Light flashes consistently caused a phase delay between 1200 and 1800. B: a single 2-ms pulse of light has little effect on the phase of the circadian rhythm. In striking contrast, a series of 2-ms photic pulses at 60 1-min intervals produced a substantial phase delay, but little apparent phase advance.

The multiple flashes given between CT14 and CT18 caused a phase shift of 1.63 h (±0.33, SE) based on 11 mice during this period. These results are significantly different from the absence of a significant phase shift in seven mice that received a single flash during the same light-sensitive time period, CT14 to CT18 (t-test,P < 0.002).

Because the data from all trials in Fig.2 A create a typical mouse phase-response curve, it is unlikely that the observed phase shifts of individual animals were random events induced by stress or other perturbations. Occasionally, the same mice were exposed to nonphotic disturbances associated with cage cleaning and to control experiments in which foreign mice were introduced into the cages for 1 h. In no case did these disturbances cause phase shifts of the circadian wheel running rhythms of the experimental mice, suggesting that the responses found here are selective to stimulation of the visual pathway.


What is remarkable about the results of this study is that maximal phase delays were produced with extremely brief total exposures to light. Previous experiments have commonly used 15-min periods of light, which were considered short, to generate maximal phase shifts in circadian rhythms (16). Experiments with hamsters demonstrated that pulses of strong light of 3-s duration produced relatively small phase shifts (12). In a study of nocturnal flying squirrels, 1-s light pulses at 24-h intervals were sufficient to entrain the animals’ activity rhythms (4). A phase-response curve was not reported in that study, however, and therefore we do not know if the 1-s pulses used were capable of producing large-magnitude phase shifts such as the ones shown in the present study. It is possible that a free running rhythm could entrain to a stimulus that would not be of sufficient magnitude to produce a maximal phase shift if it were delivered at appropriate circadian times. The present results differ from all preceding studies in that they demonstrate maximal (for the mouse) phase delays in response to extremely brief exposures to light, amounting to only 120 ms in total duration given 2 ms at time over 5 or 60 min. These results suggest that the total duration of light can be less important than the integration of time over which the light intermittently appears. Furthermore, the results indicate that neurons can store information derived from brief photic stimuli for at least 60 s and probably can accumulate and integrate the brief sensory information over a substantially longer period.

The observation that a series of 2-ms pulses of light causes a permanent alteration in the timing of circadian rhythms is an interesting case of temporal summation or integration of environmental information by the circadian system. The effect of this temporal summation is the induction of large and long-lasting phase shifts of the endogenous activity cycle. It is interesting to note that a phase shift means that the circadian system will respond differently to the same stimulus applied 24 h later than it would have responded if no phase shift had occurred. Thus the phase shift is a response modification that resulted from experience, which may be considered a simple form of learning (18). There are other instances in which long-term behavior modification has been associated with phase shifting of circadian rhythms. Circadian “aftereffects” were found in a paradigm in which animals were entrained to a short (22 h) light-dark cycle. After being released into constant conditions, they continued to show relatively short free running periods for several weeks (15). In another study, Pavlovian conditioning of the circadian system was demonstrated. When nonphotic sensory stimuli were paired with light cues, they became capable of inducing phase shifts when tested alone (1).

Interesting questions posed by the results of this study are where in the circadian system is the temporal integration of brief photic stimuli occurring and what are the cellular mechanisms? Summation of the discrete photic stimuli could be occurring in the retina itself or in the SCN or in the intergeniculate leaflet of the thalamus, which receives retinal innervation and projects to the SCN (2) and may be involved in some aspects of phase shifting (7). After photic stimulation, retinal rod and cone photoreceptors may not fully recover to baseline levels within several seconds, potentially leading to an ongoing change in retinal activity if a subsequent flash of light causes a continuing deviation from baseline retinal output (9, 14). However, we found a substantial phase shift even if light flashes were spaced 1 min apart, an interval probably too long to be accounted for by continuous retinal activation after photic stimulation in mice at the level used in this experiment (9). If continuous retinal signals were to be sent for up to 1 min due to long-lasting rod or cone activation, then a phase shift should be generated by a single flash, but in our experiments a single flash was ineffective in causing phase shifts. Thus the photoreceptors in the retina do not appear to be the site of integration of the trains of photic stimuli delivered in our experiments.

It is most likely that a central nervous system site, either the SCN or the intergeniculate leaflet of the thalamus, is responsible for the temporal summation of photic information observed in these experiments. Candidate mechanisms have been described that may act as a substrate for neuronal changes in response to photic stimulation. Neuromodulators that may be involved in photic signaling can exert long-term actions on both intracellular calcium levels and on electrical activity of SCN neurons, probably through G protein-coupled second messengers (13, 19). The mechanism for this neuromodulatory action appears to be a long-term depression of synaptic activity. Perhaps such a mechanism enables the SCN to integrate over time extremely brief stimuli that individually would be ineffective in inducing phase shifts.

This study points out what is probably a major adaptive difference between the visual system and the circadian system. It is adaptive for the visual system to register brief photic stimuli as discrete events to be maximally responsive to rapid changes in visual information. For the circadian system of a nocturnal animal, however, it is more important to be able to register the presence of low levels of light even when the animal is active in an environment (e.g., under tree or brush cover) that precludes continuous exposure to the stimulus.

There may be a practical application for the fact that the circadian system temporally integrates photic stimuli. Bright light therapy is being used to entrain the human circadian system in situations ranging from jet lag to treating seasonal affective disorder. The treatment involves the individual remaining in front of a bank of high-wattage electric lights for an extended period. It may be possible to use strobe flashes to achieve the same effect. This would be more portable and require less space, less energy, and be less constraining to the individual.


We thank Drs. E. Pugh, T. Kilduff, F. Rieke, and D. Edgar for helpful discussion or use of equipment.


  • Address for reprint requests: A. N. van den Pol, Dept. of Neurosurgery, Yale Medical School, 333 Cedar St., New Haven, CT 06520.

  • This work was supported by grants from the National Institute of Neurological Disorders and Stroke (NS-10174, NS-31573, and NS-34887), National Science Foundation, and Air Force Office of Scientific Research.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract