Multi-oscillatory control of circadian rhythms in human performance

Abstract
Circadian rhythms are known to exist in many measures of human performance efficiency as well as in physiological processes. The demands of a task, and in particular its 'working memory' load, play a large part in determining the time of day at which it is best performed. Furthermore, task demands may affect the speed with which performance rhythms adjust to the altered sleep/wake schedules occasioned by shift-work and rapid time-zone transitions. These differences in rate of adjustment may be explained by a similar multi-oscillatory model to those proposed for physiological rhythms. These assume any given circadian rhythm to be jointly controlled by two endogenous oscillators. The first is thought to be relatively immune to exogenous factors and to control the temperature rhythm, while the second is thought to be more influenced by exogenous factors and to have the major role in governing the sleep/wake cycle. Normally, the pronounced 24-h time cues, or 'zeitgebers', in our environment result in both oscillators, and hence all circadian rhythms, running with a period of 24 h. However, under altered sleep/wake schedules, and in conditions of temporal isolation, the temperature rhythm and sleep/wake cycle may separate from one another and run with distinctly different periods. When such 'internal desynchronization' occurs, other physiological rhythms have been found to run in synchrony with one or other of these two functions. This finding forms the basis of current multi-oscillatory models. However, studies of abnormal sleep/wake schedules suggest that the rhythm in working memory performance may sometimes separate from both the sleep/wake cycle and temperature rhythm by running with a period of less than 24 h. We have investigated this possibility here and our results indicate control of working memory performance rhythms by a previously unidentified oscillator with an autonomous period of about 21 h.

This publication has 10 references indexed in Scilit: