A cost-benefit analysis of behavioural sequence organization in laboratory mice

Abstract

To test the hypothesis that the actual organization of diurnal and nocturnal behavioural sequences observed in C3H mice tend to maximize a classical functional criterion — the net energy gain — artificial behavioural sequences were defined by reorganization of the acts of actual ones, through application of three Monte-Carlo-type processes (ALT, SEQ and RND). ALT sequences correspond to a random choice of rest/activity bouts, and SEQ sequences to a random reorganization of acts in each activity bout. RND sequences correspond to a random choice of rest/activity bouts together with a random choice of acts in each activity bout. The net energy gain associated with each of these artificial behavioural sequences was derived from a computation of the corresponding energy inputs — by means of a dynamic feeding model — and energy outputs, on the basis of the energy cost of each constituent act. This net gain was then compared with the net gain corresponding to the actual behavioural sequence of the animal. The results of RND and ALT sequences substantiate the maximization hypothesis, while those of SEQ sequences reveal individual differences. This suggests the existence of two independent behavioural strategies — choice of rest/activity alternations and choice of acts during an activity bout — as they can lead to antagonist effects on the functional criterion.