Several experimental studies have altered the phase relationship between photic and non-photic environmental, 24 h cycles (zeitgebers) in order to assess their role in the synchronization of circadian rhythms. To assist in the interpretation of the complex activity patterns that emerge from these “conflicting zeitgeber” protocols, we present computer simulations of coupled circadian oscillators forced by two independent zeitgebers. This circadian system configuration was first employed by Pittendrigh and Bruce (1959), to model their studies of the light and temperature entrainment of the eclosion oscillator in Drosophila. Whereas most of the recent experiments have restricted conflicting zeitgeber experiments to two experimental conditions, by comparing circadian oscillator phases under two distinct phase relationships between zeitgebers (usually 0 and 12 h), Pittendrigh and Bruce compared eclosion phase under 12 distinct phase relationships, spanning the 24 h interval. Our simulations using non-linear differential equations replicated complex non-linear phenomena, such as “phase jumps” and sudden switches in zeitgeber preferences, which had previously been difficult to interpret. Our simulations reveal that these phenomena generally arise when inter-oscillator coupling is high in relation to the zeitgeber strength. Manipulations in the structural symmetry of the model indicated that these results can be expected to apply to a wide range of system configurations. Finally, our studies recommend the use of the complete protocol employed by Pittendrigh and Bruce, because different system configurations can generate similar results when a “conflicting zeitgeber experiment” incorporates only two phase relationships between zeitgebers.
Citation: Oda GA, Friesen WO (2011) Modeling Two-Oscillator Circadian Systems Entrained by Two Environmental Cycles. PLoS ONE 6(8): e23895. https://doi.org/10.1371/journal.pone.0023895
Editor: Paul A. Bartell, Pennsylvania State University, United States of America
Received: April 26, 2011; Accepted: July 26, 2011; Published: August 19, 2011
Copyright: © 2011 Oda, Friesen. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by grant (06/61276-0; GAO) from FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo), and grant (IOB-061563; WOF) from USA National Science Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Synchronization of the physiology and behavior of organisms to the earth's periodic environment is achieved in part by the entrainment of circadian oscillators to species-specific combinations of daily photic and non-photic environmental cycles, known as “zeitgebers” –.
The dominant zeitgeber that entrains circadian oscillators appears to be the light/dark cycle; however, the importance of food ,  and temperature in circadian entrainment ,  has received increasing recognition. In particular, temperature entrainment in the circadian organization of both ectotherms and endotherms –, together with light effects, uncovers the necessity of considering the simultaneous action of two zeitgebers. The complexity of circadian systems comprising multiple oscillators entrained by two, or more zeitgebers can be understood through theoretical studies, where the contributions of zeitgebers and the internal circadian structure can be dissected. Such studies can also guide experiments and provide interpretations of the complex activity patterns of organisms.
Several experimental studies have altered the phase relationship between zeitgebers that occurs in nature, in order to assess their role in the entrainment of circadian rhythms –. The common experimental procedure is to artificially generate a phase difference of 12 h, in effect subjecting organisms to conflicting environmental time cues. In such “conflicting zeitgeber” experiments, it is commonly assumed that the circadian oscillator is phase-locked to the strongest zeitgeber. Intuitively, one might expect that some oscillators would follow the phase-displaced zeitgeber, whereas others would remain phase-locked to the unaltered zeitgeber.
To the best of our knowledge, Pittendrigh and Bruce (1959) performed the most complete set of conflicting zeitgeber experiments, revealing complex dynamics in the phase of the overt rhythm. Their experiments were intended to assess the relative strengths of light/dark and temperature cycles in the entrainment of the circadian oscillators controlling adult eclosion in Drosophila pseudoobscura. Populations of flies were raised under 24 h light-dark and temperature cycles that are zeitgebers in this species , the latter being successively phase-shifted by 2 h steps relative to dawn (Fig. 1). The eclosion rhythm tracks the phase of temperature cycle during the first 6 conditions, but tracks the phase of light cycle in the subsequent 6 conditions. This observed switch in the phase association of eclosion rhythm, first to the temperature and then to the light cycle, has made the identification of the stronger zeitgeber for the control of eclosion in Drosophila uncertain until now.
A) Eclosion rhythm of Drosophila populations under 24 h light/dark and temperature cycles. Each horizontal bar corresponds to two successive days. Light/dark cycles are shown by shaded areas and temperature cycles by a continuous curve, the latter being displaced by 2 h in each successive bar. Dark bars indicate the number of flies that eclosed within a 2 h time window. From Pittendrigh and Bruce (1959) with permission. B) Schematic diagram of two coupled A and B oscillators, entrained, respectively by zeitgebers L and T.
In the present work, we describe our numerical simulations of limit-cycle oscillator models  that explain the main features of the complex behavioral results of Pittendrigh and Bruce (1959). More importantly, our simulations illuminate the spectrum of dynamics generated by conflicting zeitgeber experiments by revealing how coupled oscillators respond to progressive phase displacements between zeitgebers. Our simulations show how conflicting zeitgeber experiments disentangle the complex interactions between oscillators and zeitgebers if it is “complete”; that is, when the phase relationship between zeitgebers is progressively increased in small steps. We argue that a single phase displacement between two zeitgebers may generate misleading models of the circadian system.
We performed numerical simulations inspired by the model of Pittendrigh et al.  and the experimental protocol of Pittendrigh and Bruce  with a system of coupled limit-cycle oscillators (A and B), with each oscillator affected by one zeitgeber, L or T (Figure 1B).
The A and B oscillators were simulated by coupled Pittendrigh-Pavlidis equations (1–4), where R and S are state variables, and a, b, c and d, are parameters. Parameters CAB and CBA set the coupling strengths of oscillator A to B and of oscillator B to A, respectively. Zeitgebers L, for oscillator A, and T, for oscillator B, are square-wave functions with a 24 h period. These equations differ from the Pavlidis equations  by a variable K (Kyner), which is a small, nonlinear term that ensures numerical smoothness . The R variables are explicitly constrained to be positive. These equations were developed originally for studies of the oscillator controlling the eclosion rhythm in Drosophila  and were employed in our modeling of the general properties of mammalian circadian oscillators –.
As in former applications of this model, for the sake of simplicity and to better evaluate the effects of varying inter-oscillator coupling or zeitgeber strength, we assume that the oscillators are identical by fixing the parameters such that , , and . This parameter set generates an oscillator with intrinsic period ≈ 24 hr. Short, 1 h pulses (single pulse T-cycles with h ) were used in the simulations because they are the simplest mathematical models of daily zeitgebers. Zeitgeber amplitude () and coupling strengths ( to 0.18) were chosen in a range that allowed a single zeitgeber to entrain the weakly coupled oscillators when alone. This choice of default values enabled exploration of coupling values that had the same effective magnitude as the zeitgeber strength.
Phases are defined with respect to the 24-hour day as follows (Fig. 2):
Time course of oscillator A, zeitgeber L (upper panel), oscillator B and zeitgeber T (lower panel). S and R variables of each oscillator are represented, respectively, by heavy and light lines. Zeitgebers L and T are represented by rectangular pulses with periods = 24 h. The phase of A, with respect to the acrophase (upper, filled squares) is given by φA, that of B (lower, filled circles) is given by φB. The phases of the zeitgeber pulses are indicated by ΦL (upper, open squares) and ΦT (lower, open circles). The phase difference between oscillators A and B is represented by φAB, while the phase difference between zeitgebers L and T is represented by ΦLT.
φA = acrophase of oscillator variable SA, phase at which SA takes its maximum value.
φB = acrophase of oscillator variable SB.
= phase difference between coupled oscillators A and B.
ΦL = phase of zeitgeber L pulse onset.
ΦT = phase of zeitgeber T pulse onset.
= phase difference between zeitgebers L and T. All phase values are given with respect to ΦL, which is assigned a value of 12 h. Thus, for example, a temperature onset phase of 8 h means that the temperature parameter T was set to 2, 4 h before light onset.
Simulations were performed with the CircadianDynamix software, which was developed to explore problems related to coupled and forced oscillators in chronobiology. It is an extension of Neurodynamix II , . We used the Euler method for numerical integration, with 1000 integration steps per 24-hour day.
The phases of each oscillator (φA and φB) were evaluated under a series of entrainment conditions by successively increasing ΦT, in one-hour steps, from +12 to +24 h, and then from 0 to +12 h. The final state of each entrained condition was used as the initial state of the subsequent condition. The reverse sequence, from +24 to +12 h and from +12 h back to 0 h was also employed in order to test for dependence on initial conditions.
Furthermore, we focused on how φA and φB, at each ΦT, are affected by changes in the strength and symmetry of the inter-oscillator coupling and zeitgeber strength. The reference system was completely symmetrical, i.e. the oscillators, zeitgebers and inter-oscillator coupling were identical; asymmetry was added by incremental changes in relative strengths of inter-oscillator couplings or zeitgeber amplitude.
Symmetric System: identical oscillators, zeitgebers and coupling
To learn how the phase difference between zeitgebers (ΦLT) affects the phases of the A (φA) and B (φB) oscillators, as well as their phase relationship (φAB), we set Pittendrigh-Pavlidis oscillator parameters to the simplest configuration: two identical oscillators with equal bidirectional coupling (). This system was subjected two identical, independent zeitgebers, whose relative phases were stepped from 0 to 24 h.
We first investigated the effects of coupling strength between the two oscillators. The output of the symmetric system is shown in Figure 3, where the inter-oscillator coupling strength was increased from (panel A), 0.01 (panel B), 0.07 (panel C), 0.15 (panel D) and 0.18 (panel E). In the absence of coupling, each oscillator is phase-locked to its zeitgeber and the phase relationship between zeitgebers and oscillator acrophases remains constant (Fig. 3A). We found that even very weak coupling between oscillators was sufficient to modulate φA and φB, as revealed by the changing relationships between the zeitgebers and the oscillators (Fig. 3B, C). These effects become increasingly prominent with increasing inter-oscillator coupling strength (while the zeitgeber strength was maintained); eventually, the oscillators appear more strongly influenced by each other than by their zeitgebers (Fig. 3D, E).
Steady-state oscillator phases (φA and φB) are represented by filled squares and circles, respectively; zeitgeber phases (ΦL and ΦT) are represented by open squares and circles, respectively (as in Fig. 2). In each of the 24 successive horizontal bars ΦT was increased by 1 h with respect to the phase of zeitgeber L, which was fixed at ΦL = 12 h. The duration and amplitude of both zeitgebers were fixed. Coupling strengths C were set to: A) 0.0; B) 0.01; C) 0.07; D) 0.15; and E) 0.18. Pittendrigh-Pavlidis model parameters here and in the remaining figures: a = 0.85, b = 0.3, c = 0.8, d = 0.5. T = L = 2.
A particularly interesting phenomenon is observed when zeitgebers T and L are near antiphase (h). When inter-oscillator coupling is weak relative to the zeitgeber strength (Fig. 3A–C), oscillators attain large phase differences, up to 12 h, by non-linear, but smooth φA and φB changes. These phase differences are shown more clearly in Figure 4, where simulation data for the symmetric systems are replotted as φAB as a function of ΦLT. These curves are sigmoidal, but with a sharp transition zone (for or greater). The maximum coupling value for which the antiphasic relationship between oscillators occurs (h when h) is (Fig. 4A). Below this coupling value, φAB is a nearly linear function of ΦLT, but above it the oscillators become tightly phase-coupled, acting more as a single system. Under this tight coupling, large values of φAB are not allowed (Fig. 3D and E, Fig. 4B).
A) Increasing coupling strengths: filled circles,C = 0.07; triangles, C = 0.11; stars, C = 0.15. B) Strong coupling: C = 0.18. Δ1 and Δ2 indicate the interval where hysteresis occurs; arrows indicate the direction of the change in ΦLT. When ΦLT is below Δ1, phase jumps occurs at Δ2. When ΦLT is above Δ2 the converse path is taken and phase jumps occur at Δ1.This history dependence is the hallmark of hysteresis and gives rise to two φAB values (bistability) in the interval Δ1<ΦLT<Δ2. Pittendrigh-Pavlidis model parameters: T = L = 2.
Moderately abrupt changes in φAB occur at h, corresponding to the inflection point of the φAB vs ΦLT curve, henceforth called “inflection phase”. In this vicinity, small changes in ΦLT result in relatively large changes in φAB, hereafter described as a “phase jump”. The stronger the coupling, the steeper the phase jump, such as shown for .
For coupling values larger than , the results look rather different. In Figure 4B, (, corresponding to data in Fig. 3E), as ΦLT increases from 0 to 24 h, φAB remains near 0 h even beyond h, revealing an unexpected asymmetry in the graph. Then, as ΦLT increases further, at Δ2 (Fig. 4B) the value of φAB jumps suddenly from just above 0 h to just below 24 h. With greater ΦLT values, φAB remains below 24 h and a nearly linear function of ΦLT, attaining 0 h = 24 h for h.
A complete picture is revealed by considering the converse sequence of zeitgeber phase displacements, from h to h. Although most φAB values are identical to those obtained before, the phase jump occurs at h (at Δ1, Fig. 4B). Thus φAB attains two steady-states in the interval , depending on whether the previous ΦLT was larger or smaller than 12 h. The complete graph, generated by increasing and decreasing ΦLT (indicated by arrows, Fig. 4B), restores the symmetry that was apparently missing; the picture now comprises bistability and hysteresis .
Bistability implies the existence of two stable steady states φAB, for the same ΦLT values located in the “bistability zone” , as is also seen in the explicit φA and φB values (lines 12,13 and 14 of Fig. 3E). The phases attained by the system depend on initial conditions and this dependence on the path of parameter change is a hallmark of hysteresis (arrows, Fig. 4B). Preliminary exploration of alternative parameter sets of Pittendrigh-Pavlidis equations have shown, however, that the bistability zone is reduced as the free-running periods of oscillators deviate from 24 h.
- Asymmetry in inter oscillator coupling strengths
We next examine the role of asymmetry by retaining equal zeitgeber strengths, but introducing asymmetry in the relative inter-oscillator coupling strengths. Using Pittendrigh-Pavlidis equations and departing from the symmetric case of a weakly coupled configuration , CBA was reduced from CAB to zero.
When φAB is plotted against ΦLT (Fig. 5A), the main qualitative feature of the symmetric system (Fig. 4) is recognized; namely, the sigmoidal shape of the curve. However, as asymmetry is increased by having , a progressive shift of the inflection phase occurs from h to h (Fig. 5A). Maximum departure of the inflection phase from h occurs at the extreme, case, which corresponds to unidirectional coupling; that is, a “master-slave” configuration.
A) Phase difference between oscillators (φAB) as a function of the phase difference between zeitgebers (ΔΦLT). CAB = 0.07 and CBA varies as a fraction of CAB (see legend inside the figure). B) Oscillator and zeitgeber reference phases for a master-slave configuration, elicited by CAB = 0.07,CBA = 0. Oscillator phases (φA and φB) are represented by filled squares and circles, respectively; zeitgeber phases (ΦLand ΦT) are represented by open squares and circles, respectively. In each of the 24 successive horizontal bars ΦT was increased by 1 h with the phase of zeitgeber L fixed at ΦL = 12 h. The duration and amplitude of both zeitgebers were fixed. C) For a stronger value of unidirectional coupling CAB = 0.11 and CBA = 0, relative coordination (loss of stable entrainment) occurs in lines 9 and 10 (hatched). Pittendrigh-Pavlidis model parameters: T = L = 2.
In the case of unidirectional coupling from A to B, φA is always phase-locked to zeitgeber L, whereas φB is modulated by both zeitgeber T and by L (via inputs from A; Fig. 5). Therefore, the dynamics of this asymmetrical system is now strongly dependent on the relative interaction strengths between zeitgeber T and CAB, which are pulling the slave B oscillator in opposite phase directions. Having fixed zeitgeber strengths (), the master-slave system was now simulated for different coupling strengths CAB (Fig. 5B,C).
For a representative value of weak coupling (), φB is a smooth function of ΦT until h, where a phase-jump occurs (7th line, Fig.5B), and with the oscillator following the zeitgeber T smoothly thereafter. The sudden increase in φB at h, not at h, corresponds to the shift in the inflection phase of the sigmoidal curve.
The overall picture is similar for stronger coupling values (, 0.15 and 0.18), except that T is not sufficiently strong to phase-lock φB at some ΦT (9 h and 10 h, in this case). Thus, there is no stable entrainment at these ΦT, and oscillator B is in relative coordination  (Fig. 5C, lines 10 and 11, where φB was omitted). As coupling strength increases, the range of ΦLT that yields relative coordination enlarges.
- Asymmetry in zeitgeber strengths
Asymmetry in zeitgeber strength, with symmetric inter-oscillator coupling, yields results similar to asymmetrically coupled oscillator systems; namely, a sigmoidal φAB vs. ΦLT curve, with a shift in the phase of inflection. The inflection phase is greater for greater asymmetry between the two zeitgeber strengths (Fig. 6A). Phase jumps again occur at values that differ from h. For comparison, oscillator phases are shown for a zeitgeber L which is 4 times stronger than T (Fig. 6B) and conversely, for a zeitgeber T which is 4 times stronger than L (Fig. 6C), resulting in a shift of inflection phase at ΦLT smaller or greater than 12 h, respectively.
A) Phase difference between oscillators (φAB) as a function of the phase difference between zeitgebers (ΦLT). Filled circles: asymmetric zeitgeber strengths L = 4,T = 1; Open circles:T = 4, L = 1; Stars: symmetric T = L = 2. B) Oscillator and zeitgeber reference peak phases for an asymmetric zeitgeber configuration. Oscillator phases (φA and φB) are represented by filled squares and circles, respectively; zeitgeber phases (ΦL and ΦT) are represented by open squares and circles, respectively. In each of the 24 successive horizontal bars ΦT was increased by 1 h with the phase of zeitgeber L fixed at ΦL = 12 h. The durations of both zeitgebers were fixed while amplitudes were L = 4 and T = 1. C) Zeitgeber amplitudes T = 4 and L = 1. Pittendrigh-Pavlidis parameters: CAB = CBA = 0.07.
Interpreting Pittendrigh and Brucés (1959) Drosophila eclosion data
The following associations were made, in order to apply our model simulations to the Drosophila data.The phase relationship between zeitgebers was assigned the value when the phase of temperature minima () occurred at dawn (), as represented in the first horizontal bar in Fig. 1A. Other complementary experiments described in  have shown that the phase of eclosion is determined by a temperature dependent oscillator B, which corresponds in our simulations to φB. Furthermore, parallel simulations, not presented in this manuscript, have shown that the main dynamical features of periodic single-pulse zeitgebers are replicated by other cyclic wave forms, if appropriate amplitude adjustments are made.
We now focus on the following main features of the phase dynamics in Pittendrigh and Brucés (1959) data (Fig. 1A):
- The phase shifts of the eclosion rhythm tracked the successive phase shifts of the temperature cycle in the h interval (horizontal bars 1 to 6).
- An abrupt phase jump of the eclosion peak was observed when zeitgebers attained maximum conflicting phase differences; i.e., near h (horizontal bar 7).
- The eclosion phase remained nearly unaltered thereafter, independent of the phase of the temperature cycle along the h interval (horizontal bars 8 to 12).
First, our simulations support the two-oscillator model of Pittendrigh and Bruce because an alternative model, comprising a single oscillator with temperature and light inputs is equivalent to a master-slave configuration system (Fig. 5) that does not replicate the above features.
Our simulations of the two-oscillator symmetric system with intermediate coupling relative to zeitgeber strength (Fig. 3D), qualitatively reproduces the findings in Drosophila. During the interval h, the phase-shifts of oscillator B track the phase shifts of the temperature zeitgeber. When h, a “phase jump” of φB occurs; thereupon φB tracks the phase of the light zeitgeber for h. Our simulations have thus shown that the switch at h of the preferential phase association of eclosion rhythm, first to the temperature and then to the light cycle can arise even in the most symmetric configuration, with identical oscillators coupled to identical zeitgebers (Fig. 3). By using a simple symmetric configuration, we show that additional complexity (unequal parameter values), even if it exists, is unnecessary for generating the observed complex phenomena. In other words, the results of the Pittendrigh and Bruce experiment do not indicate any zeitgeber dominance in Drosophila, neither of light/dark nor of temperature.
Our general study of conflicting zeitgeber experiments involving two differentially entrainable oscillators is applicable for a wide range of systems, such as: light/dark and temperature entrainable oscillators controlling eclosion rhythms in other insect species ; neuronal groups within the Drosophila brain, assessed by molecular concentrations ,  or by emergent activity levels ; gene expression in the plant Arabidopsis thaliana ; dawn and dusk entrainable neuronal groups within the suprachiasmatic nucleus ; and finally, the light- and food- entrainable circadian oscillators in several species , , –, , .
Most of the experimental studies that altered the natural phase relationship between photic and non-photic zeitgebers have restricted it to two ΦLT values, usually 0 and 12 h. Furthermore, they have also assumed linear relationships between the phases of zeitgebers and oscillators –, , . Such assumptions lead to the proposition that the phase of an output rhythm tracks the phase displacement of the stronger zeitgeber and thus two ΦLT conditions are sufficient for revealing the zeitgeber hierarchy. While this is correct when there is no coupling between oscillators (Fig. 3A), inter-oscillator coupling effects (Fig. 3B–E) add more complexity to the conflicting zeitgeber experiment, requiring a systematic change in ΦLT conditions to avoid ambiguous models.
The complex picture that emerged from Pittendrigh and Brucés experiment, comprising 12 ΦLT relations (Fig. 1A) is best replicated by the features of Fig. 3D, where light is as strong as temperature in the entrainment of the eclosion oscillator. However, if their experiment were restricted to only two phase relations between zeitgebers, for instance, and h (horizontal bars 1 and 7 of Fig. 1), the more limited set of phase results could be replicated by several other configurations.Furthermore, the problem persists if a different ΦLT pair is chosen for a conflicting zeitgeber experiment.
Some predictions and guidelines for a complete conflicting zeitgeber experiment arise from our simulations. The hypothesized general system is composed of two oscillators, differentially entrained by two zeitgebers. If the phase of the output rhythm (or of the oscillator itself) is plotted against the phase difference between zeitgebers Z1 and Z2, ΦZ1Z2 a curve with different characteristics is expected.
- If the resulting graph is linear, instead of sigmoidal, there is no coupling between oscillators. Alternatively, the strength of one zeitgeber may not have been sufficiently strong to entrain its corresponding oscillator or the observed output is controlled by a master oscillator without feed-back from the unobserved slave (as it would occur in Fig. 5B, if the phase of zeitgeber L were shifted, instead of T).
- For a sigmoidal graph, the steeper the curve, with more pronounced phase jumps, the stronger is the inter-oscillator coupling with respect to the zeitgeber strengths (e.g., Fig. 3).Phase-jumps occurring at h are indicative of either an asymmetric system, with unequal coupling/zeitgeber strength between the oscillators, or a bistable system, with two potential entrained oscillator phases for an interval of ΦZ1Z2.
- Dependence on initial conditions, with two possible entrained phases, is predicted to occur around the phase jump region if the inter-oscillator coupling is strong compared to the zeitgebers. This can be tested by pre-entraining with either of two ΦZ1Z2: in the natural phase relation or in ΦZ1Z2 greater than the phase of inflection.
- Relative coordination is expected to occur for some ΦZ1Z2 values, when coupling is strong with respect to zeitgeber strength.
Some other experiments have reported our predicted shift in the inflection phase, typical of asymmetric systems. This was observed by Bruce  and Pittendrigh , when they assayed, respectively, the circadian phototaxis rhythm in Euglena viridis and activity in cockroaches under light/dark and temperature cycles. Recently, Watari and Tanaka  verified that phase jumps occur at a h in their conflicting zeitgeber experiments on the eclosion of onion fly Delia antiqua. Since only fixed initial conditions were used in these experiments, phase-jumps at h could alternatively reflect a single branch of a bistable system (Fig. 3E), which would be testable by manipulating initial entrainment conditions.
There is much to be understood about the roles and interactions between non-photic and photic daily cues in the circadian organization of different species, in the context of different environments , –. Theoretical studies of multiple oscillator models explained phenomena that could not be accounted for by a single oscillator , – and widened opportunities for creative biological experiments , , , –. We are now facing the challenge of including multiple zeitgebers in this scenario, but in these early steps, our modeling study is appropriately limited to two general cases: symmetrical systems and those with simple asymmetries. Despite its simplicity, our system sufficed to give a glimpse of the rich dynamics behind multiple zeitgeber phenomena.
Dr. Andreas Bohn for suggesting theoretical studies on two zeitgeber circadian systems; Daniel Damineli for discussions; Dr. Didier Gonze for the critical review of this manuscript.
Conceived and designed the experiments: GAO WOF. Performed the experiments: GAO. Wrote the paper: GAO WOF. Developed computer program: WOF.
- 1. Gwinner E (1966) Periodicity of a circadian rhythm in birds by species-specific song cycles (Aves, Fringillidae: Carduelis spinus, Serinus serinus). Experientia 22: 765–766.
- 2. Menaker M, Eskin A (1966) Entrainment of circadian rhythms by sound in Passer domesticus. Science 154: 1579–1581.
- 3. Aschoff J (1981) Free running and entrained circadian rhythms. In: Aschoff J, editor. In Behavioral neurobiology, vol 4 – Biological rhythms,. New York: Plenum Press. pp. 81–93.
- 4. Mrosovsky N (1988) Phase response curves for social entrainment. J Comp Physiol A 162: 35–46.
- 5. Mistlberger RE (1994) Circadian food-anticipatory activity: formal models and physiological mechanisms. Neurosci Behav Rev 18(2): 171–195.
- 6. Stephan FK (2002) The “other” circadian system: food as a zeitgeber. J Biol Rhythms 17(4): 284–292.
- 7. Sweeney BM, Hastings JW (1960) Effects of temperature upon diurnal rhythms. Cold Spring Harbor Symposia on Quantitative Biology 25: 87–104.
- 8. Rensing L, Ruoff P (2002) Temperature effect on entrainment, phase shifting, and amplitude of circadian clocks and its molecular bases. Chronobiol Intern 19(5): 807–864.
- 9. Brown SA, Zumbrunn G, Fleury-Olela F, Preitner N, Schibler U (2002) Rhythms of mammalian body temperature can sustain peripheral circadian clocks. Curr Biol 12(18): 1574–83.
- 10. Refinetti R (2010) Entrainment of circadian rhythm by ambient temperature cycles in mice. J Biol Rhythms 25(4): 247–56.
- 11. Buhr ED, Yoo SH, Takahashi JS (2010) Temperature as a universal resetting cue for mammalian circadian oscillators. Science 15: 379–385.
- 12. Abraham U, Granada AE, Westermark PO, Heine M, Kramer A, et al. (2010) Coupling governs entrainment range of circadian clocks. Mol Sys Biol 6: 1–13.
- 13. Pittendrigh CS, Bruce VG (1959) Daily rhythms as coupled oscillator systems and their relation to thermoperiodism and photoperiodism. In: Withrow RB, editor. In Photoperiodism and related phenomena in plants and animals. Washington, DC: Am Assoc Adv Sci. pp. 475–505.
- 14. Bruce VG (1960) Environmental entrainment of circadian rhythms. Cold Spring Harbor Symp Quant Biology 25: 29–48.
- 15. Reierth E, Stokkan KA (1998) Dual entrainment by light and food in the Svalbard Ptarmigan (Lagopus mutus hyperboreus). J Biol Rhythms 13(5): 393–402.
- 16. Challet E, Solberg LC, Turek FW (1998) Entrainment in calorie restricted mice: conflicting zeitgebers and free-running conditions. Am J Physiol 274: 1751–1761.
- 17. Lague M, Reebs SG (2000) Phase-shifting the light-dark cycle influences food-anticipatory activity in golden shiners. Physiol Behav 70: 55–59.
- 18. Damiola F, Le Minh N, Preitner N, Kornmann B, Fleury-Olela F, et al. (2000) Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus. Genes Dev 14: 2950–2961.
- 19. Hara R, Wan K, Wakamatsu H, Aida R, Moriya T, Akiyama M, et al. (2001) Restricted feeding entrains liver clock without participation of the suprachiasmatic nucleus. Genes Cells 6(3): 269–278.
- 20. Stokkan KA, Yamazaki S, Tei H, Sakaki Y, Menaker M (2001) Entrainment of the circadian clock in the liver by feeding. Science 291: 490–493.
- 21. Michael TP, Salomé PA, McClung CR (2003) Two Arabidopsis circadian oscillators can be distinguished by differential temperature sensitivity. Proc Natl Acad Sci USA 100(11): 6878–6883.
- 22. Miyasako Y, Umezaki Y, Tomioka k (2007) Separate sets of cerebral clock neurons are responsible for light and temperature entrainment of Drosophila circadian locomotor rhythms. J Biol Rhythms 22: 115–126.
- 23. Currie J, Goda T, Wijnen H (2009) Selective entrainment of the Drosophila circadian clock to daily gradients in environmental temperature. BMC Biology 7: 492–507.
- 24. Watari Y, Tanaka K (2010) Interacting effect of thermoperiod and photoperiod on the eclosion rhythm in the onion fly, Delia antiqua supports the two-oscillator model. J Insect Physiol 56: 1192–1197.
- 25. Yoshii T, Hermann C, Helfrich-Förster C (2010) Cryptochrome-positive and negative clock neurons in Drosophila entrain differentially to light and temperature. J Biol Rhythms 25(6): 387–398.
- 26. Zimmerman WF, Pittendrigh CS, Pavlidis T (1968) Temperature compensation of the circadian oscillation in Drosophila pseudoobscura and its entrainment by temperature cycles. J Insect Physiol 14: 669–684.
- 27. Pavlidis T (1967) A mathematical model for the light affected system in the Drosophila eclosion rhythm. Bull Math Biophys 29: 291–310.
- 28. Pittendrigh CS, Bruce V, Kaus P (1958) On the significance of transients in daily rhythms. Proc Natl Acad Sci 44: 965–973.
- 29. Oda GA, Friesen WO (2002) A model for splitting of the running-wheel activity in hamsters. J Biol Rhythms 17(1): 76–88.
- 30. Oda GA, Menaker M, Friesen WO (2000a) Modeling the dual pacemaker system of the tau mutant hamster. J Biol Rhythms 15(2): 246–264.
- 31. Schwartz MD, Wotus C, Liu T, Friesen WO, Borjigin J, et al. (2009) Dissociation of circadian and light inhibition of melatonin release through forced desynchronization in the rat. Proc Natl Acad Sci USA 106(41): 17540–17545.
- 32. Pittendrigh CS, Daan S (1976) Functional analysis of circadian pacemakers in nocturnal rodents IV. Entrainment - pacemaker as clock. J Comp Physiol 106(3): 291–331.
- 33. Friesen WO, Friesen JA (2009) Neurodynamix II. Concepts of neurophysiology illustrated by computer simulations. New York: Oxford University Press.
- 34. Friesen WO, Friesen JA (2009) Neurodynamix II. Neuronal simulation models with preprogrammed lesson. available on-line at www.neurodynamix.net.
- 35. Glass L, Mackey MC (1988) From clocks to chaos: the rhythms of life. Princeton, USA: Princeton University Press.
- 36. Wever R (1972) Virtual synchronization towards the limits of the range of entrainment. J Theor Biol 36: 119–132.
- 37. Inagaki N, Honma S, Ono D, Tanahashi Y, Honma K (2007) Separate oscillating cell groups in mouse suprachiasmatic nucleus couple photoperiodically to the onset and end of daily activity. Proc Natl Acad Sci 104(18): 7664–7669.
- 38. Sánchez-Vásquez F, Zamora S, Madrid JA (1995) Light-dark and food restriction cycles in sea bass: effect of conflicting zeitgebers on demand-feeding rhythms. Physiol Behav 58(4): 705–714.
- 39. Mendoza J, Angeles-Castellanos M, Escobar C (2005) A daily palatable meal without food deprivation entrains the suprachiasmatic nucleus of rats. Eur J Neurosci 22: 2855–2862.
- 40. Pittendrigh CS (1960) Circadian rhythms and the circadian organization of living systems. Cold Spring Harbor Symp Quant Biology 25: 159–184.
- 41. Neumann D, Heimbach F (1985) Circadian range of entrainment in the semilunar eclosion rhythm of the marine insect Clunio marinus. J Ins Physiol 31(7): 549–557.
- 42. Roces F, Núñez JA (1996) A circadian rhythm of thermal preference in the ant Camponotus mus: masking and entrainment by temperature cycles. Physiol Entomol 21: 138–142.
- 43. Tanaka K, Watari Y (2003) Adult eclosion timing of the onion fly, Delia antique, in response to daily cycles of temperature at different soil depths. Naturwissenschaften 90: 76–79.
- 44. Van der Veen DR, Le Minh N, Gos P, Arneric M, Gerkema MP, et al. (2006) Impact of behavior o central and peripheral circadian clocks in the common vole Microtus arvalis, a mammal with ultradian rhythms. Proc Natl Acad Sci 103: 3393–3398.
- 45. Menaker M (2006) Circadian organization in the real world. Proc Natl Acad Sci 103: 3015–3016.
- 46. Winfree AT (1980) The geometry of biological time. New York, USA: Springer.
- 47. Kronauer RE, Czeisler CA, Pilato SF, Moore-Ede MC, Weitzman ED (1982) Mathematical model of the human circadian system with two interacting oscillators. Am J Physiol 242(1): R3–17.
- 48. Page TL (1985) Clocks and circadian rhythms. In: Kerkut GA, Gilbert LI, editors. In Comprehensive Insect Physiology, Biochemistry and Pharmacology,. vol6. Oxford: Pergamon Press. pp. 632–635.
- 49. Takahashi JS, Menaker M (1982) Entrainment of the circadian system of the house sparrow: a population of oscillators in pinealectomized birds. J Comp Physiol 146: 245–253.
- 50. Huang G, Wang L, Liu Y (2006) Molecular mechanism of suppression of a circadian rhythm by a critical stimulus. EMBO J 25: 5349–5357.
- 51. Nagano M, Adachi A, Nakahama K, Nakamura T, Tamada M, et al. (2003) An abrupt shift in the day/night cycle causes desynchrony in the mammalian circadian center. J Neurosci 23(14): 6141–6151.