Changing the shape of circadian rhythms with light no brighter than moonlight

Day and night are often simulated for nocturnal rodents in the laboratory with 24 h lighting regimes alternating between moderate indoor light levels and complete darkness. Sometimes, complete darkness is replaced by dim illumination to aid human vision, with the assumption that the circadian clock of the rodent is blind to light of extremely low irradiance. However, completely dark nights are not functionally equivalent to dimly lit nights, even when nighttime illumination is below putative thresholds for the circadian visual system. Under a variety of analytical paradigms, dim nighttime illumination increases the plasticity of circadian entrainment in the nocturnal hamster. The present studies are designed to investigate the mechanisms through which dim light operates under entrained conditions. In my first series of experiments, I demonstrate that canonical effects of dim light on circadian period and phase are insufficient to account for the potency of dim illumination under entrained conditions. Dim illumination is not a stronger zeitgeber than predicted based on previous research, nor does it directly potentiate stronger zeitgebers. Constant dim illumination increased the period of the daily activity rhythm and increased the duration of the active phase by ̃3 h relative to that displayed under complete darkness. Specific effects of dim light on circadian waveform suggest a novel action of dim light on the coupling between oscillators regulating circadian waveform. To test for effects of dim light on circadian coupling, I employed a novel behavioral assay for studying oscillator interactions. 24 h light:dark:light:dark (LDLD) cycles induce bimodal rhythms, with two daily activity bouts programmed by "split" oscillator groups cycling in antiphase. Nearly 100% of Syrian hamsters split under LDLD with dimly lit nights, compared to only 33% of animals with completely dark nights. Dimly lit nights do not facilitate splitting through mere increases in nonphotic feedback. Instead, dim illumination alters circadian responses to photic and nonphotic stimuli, with specific effects present after entrainment to short night conditions simulating the unsplit state under LDLD. The photoperiod-dependent nature of these effects indicates that dim light alters circadian coupling between oscillators entrained to short night conditions. Interactions between oscillators mediate the fusion of split rhythms after release into constant conditions and the rejoining of split activity components under LDLD. Fusion of LDLD-induced split rhythms under free-running and entrained conditions is likewise modulated by dim illumination. Specific effects on the induction, maintenance, and fusion of LDLD-induced split rhythms are consistent with the hypothesis that dim illumination alters the nature of coupling within the mammalian pacemake