Precision of circadian wake and activity onset timing in the mouse.

In each circadian cycle, a mouse begins its major activity period with discrete wake onset and activity onset events. The precision with which these events are timed in constant darkness was analyzed using the approach outlined by Pittendrigh and Daan (1976). Negative serial correlations of observed circadian period values (mean r1 = -0.471 for wake data, -0.409 for activity data) imply that deviations in period tend to be compensated by opposite deviations in the following cycle. As a result, precision of the circadian pacemaker must be better than that of observed rhythms. Standard deviation of the pacemaker period sigma(tau) was estimated at 5.1 min. Some individual data series had estimates s(tau) = 0, implying a nearly perfect pacemaker. Previous speculation was that wake onset would be under more direct pacemaker control than activity onset, and would therefore be timed more precisely (Pittendrigh and Daan 1976; Richardson et al. 1985). Contrary to this prediction, intervals between successive wake onsets exhibited significantly greater variance than intervals between successive activity onsets. Two possible interpretations of this finding were proposed.