Entrainment Is NOT Synchronization: An Important Distinction and Its Implications

“You keep using that word. I do not think it means what you think it means.”

—Inigo Montoya

It is often stated in lectures, textbooks, and even refereed scientific publications that entrainment is equivalent to synchronization. This is inaccurate and misleading. While both terms indicate temporal coordination of two or more events, to say that entrainment is synchronization is to gloss over critical distinctions and to overlook the most important adaptive values of having a circadian clock: to achieve appropriate phase control with respect to the environment and to insure that the relative timing of internal periodic events optimizes function.

Entrainment is a term used in physics to apply to a situation in which two or more oscillators match period. The oscillators may be organized hierarchically or may have a more equivalent coupling relationship whereby they influence one another mutually to set period and (inevitably) phase. In contrast, the Oxford English Dictionary defines synchronize as “to occur at the same time; to coincide in point of time; to be contemporary or simultaneous.” Literally, the term synchronous is drawn from “syn”—from the Greek for “with” or “together”—and “chronos”—time. It is important to distinguish between these terms.

Unlike entrainment, synchronization does not require even a single oscillator. If I see someone shooting at me, I duck—hopefully simultaneously or nearly so. Although neither the act of shooting nor the act of ducking is controlled by an oscillator, they occur at the same time, that is, they are synchronized. Of course, chronobiologists are typically interested in situations in which an oscillator dictates the timing of an event, but even in such a case synchronization need not result from entrainment. To state that entrainment is synchronization is to miss the critical distinction between masking and entrainment (Mrosovsky, 1999). In masking, periodicity of a behavior or a physiological event is a passive response to a recurring environmental stimulus. For example, many streetlights turn off when the sun rises and on again at sunset. This is achieved not by an oscillator in the light but by a photosensor that has no capacity to oscillate in constant conditions. The light operates periodically and synchronously with the light:dark cycle, but it is not entrained. To examine a biological instance, masking may impose periodicity upon an organism that lacks a circadian oscillator, perhaps because of an anatomical or molecular lesion. It may also apply to behavioral or physiological events in an intact organism that are responsive to the environment but not controlled by a circadian oscillator. In the former case, release into constant conditions will reveal arrhythmicity (Figure 1a). In the latter, the free run will not depart from the phase set by the environmental cycle. Unlike synchronized events, circadian rhythms show limits of entrainment to a range of periods (T; Figure 2a). Upon release into constant conditions, the onset of the free run can be extrapolated back to the entrained phase (Figures 1b and 2b and 2c).

Figure 1.

Locomotor activity rhythms can be synchronized with the dark phase of the L:D cycle in animals lacking a functional circadian clock. (a) Conditional deletion of Bmal1 in cells expressing the vesicular inhibitory amino acid transporter Vgat, including γ-aminobutyric acid-containing neurons of the suprachiasmatic nuclei, results in mice that avoid light, showing a masked activity rhythm. (b) A control mouse is also nocturnal in the L:D cycle, but free runs from the entrained phase when transferred to constant darkness. Modified from Weaver et al. (2018).

Figure 2.

Entrained oscillations differ from rhythms that are merely synchronized with the environmental cycle. (a) Unlike rhythmic activity patterns that are synchronized by the light:dark cycle, entrained locomotor activity varies in phase as the period of the zeitgeber cycle (T) changes and breaks entrainment when the difference between T and the endogenous period (τ) exceeds a limit. Shaded regions of this double-plotted actogram indicate darkness; the duration of the light and dark phases remained constant as T was gradually shortened from 24.1 h to 20 h. Running onset in this wild-type hamster was approximately synchronous with the onset of darkness for the first week of the experiment, but the phase angle (ψ) became increasingly negative as shortening of T required progressively greater phase advances. On day 26, the difference between T and τ exceeded the maximum possible phase shift as predicted by the amplitude of the phase response curve. From this point on, the hamster exhibited relative coordination. (b) Wheel running activity is plotted modulo τ in a short-period double mutant (super duper) hamster whose free running period is about 18 h and which shows a high-amplitude (Type 0) phase response curve. Activity onset is approximately synchronous with dark onset on a 3L:16.5D cycle (T = 19.5, well below the limit of entrainment for a wild-type). This animal must adopt such a phase angle to entrain by phase delays. When released into constant darkness (DD) on cycle 19, the free run departs from the entrained phase, indicating that the light cycle has, in Aschoff’s phrase, “caught the clock.” (c) When the same super duper hamster was exposed to a 3L:12.6D (T = 15.6) cycle, activity offset was approximately synchronous with lights on. T is now shorter than τ, so the hamster adopts a phase angle ψ that enables it to entrain by phase advances. When released into DD, the free run again departs from the entrained phase. In contrast, masking may lead to synchrony but entails neither a change of ψ with T nor control of phase of the endogenous clock by the environmental cycle. Figure modified from Bittman (2014).

Still more substantively, to equate entrainment with synchronization is to miss its adaptive significance. Synchronization indicates only a single-phase angle: zero. While some biological events within or between organisms that are entrained occur simultaneously, most do not—and this is a good thing. For example, animals of different social rank may differ in the time of day or night at which they approach a resource. As another example, incompatible physiological events are scheduled (through internal entrainment) to occur at different times. This is illustrated in cyanobacteria, which photosynthesize during the day and fix nitrogen at night (Welkie et al., 2014). Indeed, distinct phase clusters are evident in the expression of multiple clock-controlled genes in fungi (Dong et al., 2008), plants (Harmer et al., 2010), and animals (Koike et al., 2012). In each instance, events are entrained, but they are not synchronous. Natural selection acts to the disadvantage of organisms whose circadian period and entrainment mechanism deviate from whatever is necessary to optimize the timing of incompatible processes or to achieve a compromise in their schedule so that both needs are satisfied. Organisms that attempt to synchronize incompatible processes are less likely to survive. By insuring a variety of phase angles, entrainment maximizes their chances of living long enough to reproduce. Finally, differing phase angles of entrainment can carry important information and even specify physiological state, as in internal coincidence models of photoperiodism. This would not be possible if events were merely synchronized.

The term “desynchronization” has gained favor among some chronobiologists who manipulate light:dark cycles to examine the effects of “circadian disruption.” The latter term conveys a departure from the normal temporal program, but is neutral with respect to the entrainment of, and phase relationships between, internal oscillators. Vetter (2020) considers “circadian desynchronization” to refer specifically to a state in which two or more oscillators differ in period and traces this usage to Aschoff (1965). This implies that rhythmic events that do not occur simultaneously are in some sense synchronous, as long as their period is the same. In my opinion, this is at odds with the definition (see above): commonality of period is insufficient to define synchrony. Lack of internal synchrony does not necessarily indicate instability, nor would it inevitably be considered pathological. Perhaps it is most important to distinguish between asynchronous states in which phase relationships remain stable and those in which they change progressively and over how long an interval. Aschoff (1965) used “internal desynchrony” to refer not to a transient state but to a condition in which the different periods of sleep and body temperature rhythms persist indefinitely in constant environmental conditions and show no sign of recoupling over many weeks. This is not typical of jet lag, shift work, or other common manipulations: shifts of the light:dark cycle induce transient disruption, and deleterious effects that may result are expected to recede as normal entrainment and internal phase relationships are reestablished. Removal of the zeitgeber cycle may cause internal oscillators to split under particular conditions, such as constant light, but not others, such as constant darkness, and thus to be more or less disruptive, respectively (Pittendrigh and Daan, 1976). Other treatments, such as entrainment to T cycles that differ considerably from the endogenous period but remain within the range of entrainment, may result in a stable state that is not “desynchronized” in that the periods of the oscillations in question remain the same (West et al., 2017). Exposure to unpredictable lighting conditions or to light cycles outside the range of entrainment presents yet other types of disruptive manipulations in which instability may be the culprit, but effects of deviations from normal temporal order are more difficult to analyze or interpret. The differences between these situations are interesting and merit exploration: instability per se might cause damage, perhaps depending on its time course, but altered phase relationships might endanger health even if they do not change over the long term. Given that “alignment” refers to “arrangement into appropriate relative positions,” the use of “misalignment” to indicate a temporal relationship which departs from normal ordering has considerable advantage over “internal desynchrony.”

Why is it important to clarify the meaning of “entrainment” and to use it only when appropriate criteria have been met? One answer is that the concept lies at the core of our field: at various points, Pittendrigh referred to entrainability of circadian rhythms as “an empirical generalization” (1960) and “a characteristic, almost defining feature” (1981). Entrainability is central to the adaptive function of circadian clocks: nature selects most directly for phase and only indirectly for period. Even a circadian clock that is stripped down to its minimal components can retain the capacity for entrainment (Yoshida et al., 2009). But if inputs to a circadian oscillator are broken so that it can no longer be entrained, it remains able to free run and thus is still a circadian mechanism. Some may question whether it still qualifies as a “clock” even if it might still be usable as an interval timer. We may conceivably find organisms living in aperiodic environments, such as in deep caves or in the benthic abyss of the Pacific rift, that have circadian oscillators that cannot be entrained by external cues. However, an understanding of these relationships, and the consideration of whether entrainability is a defining characteristic of circadian rhythms, is obscured when the distinction between entrainment and synchronization is overlooked.

Victor Bruce (1960) wrote that synchronization can be equated with entrainment only if “speaking loosely.” Ravignani (2017) pointed out that “synchrony” is sometimes taken to mean loose coordination patterns, but can also refer to precise coincidence of events in time. She suggests that if life scientists were to “adopt the most restrictive definition of synchrony already used in mathematics and physics”—a precise coincidence of events in time—we would avail ourselves of better quantitative tools. I agree and urge chronobiologists to give this idea careful consideration. I understand that to equate entrainment with synchronization is to use shorthand, perhaps to convey the idea of temporal control to people unfamiliar with our field. I suggest, however, that any such statement should be immediately followed by an explanation of the distinction between the terms so that the audience understands that synchronization is at most a narrow example of entrainment, if it is even that. Respect for accuracy in terminology, and dedication to achieving a detailed understanding of critical biological processes, demands nothing less.