The relative Zeitgeber strength of lights-on and lights-off is changed in old mice.

The daily activity pattern of old mice is characterized by a decreased amplitude, a phase advance, and less stable relationship between lights-off and the onset of the main activity maximum. When analyzing the possible causes of these changes, it must be remembered that the activity rhythm of laboratory mice is bimodal, with a main peak in the first half of the dark time and a secondary one shortly after lights-on. Thus it seems to be controlled by at least two circadian oscillators--an "evening oscillator" coupled more strongly to lights-off and a "morning oscillator" coupled to lights-on--though both oscillators are also coupled to each other. The objective of the present paper was to investigate the putative changes in the strength of these couplings in HaZ:ICR mice of different ages (adult animals of 20 weeks, n = 12; old mice of 72 and 91 weeks of age, n = 6 each) and kept in a 24 h LD-cycle with a gradually reduced light:dark ratio. In adult mice, lengthening the dark time caused the onset of the main maximum of activity to be delayed in relation to the time of lights-off, while the morning maximum of activity was advanced in relation to lights-on. On average, the sizes of the advance and the delay were equal. As a consequence, the activity pattern did not shift in relation to the middle of the dark time. Lengthening the dark time resulted in a bigger (on average, 1.5 h) difference between the evening and morning activity onsets. Under short photoperiods (< or = 2 h of light) the activity rhythm started to free run, and the difference between evening and morning activity onsets decreased again. The changes obtained in senile mice were similar. However, the limits of entrainment were reached with longer photoperiods compared to adult animals. Also, the phase delay of the activity onset in the evening was much less, nearly zero. As a consequence, the activity pattern as whole phase-advanced in relation to the middle of the dark time. A model was proposed in which lights-off triggers advances of the "evening oscillator," lights-on delays the "morning oscillator," and the two oscillators are coupled with each other. Though it was probably the case, decreased coupling strengths could not be shown with the present experimental approach. However, it was clearly evident that, with increasing age, the advancing effect of lights-off exceeded the delaying effect of lights-on.