Mathematical analysis of locomotor behavior by mice in a radial maze

Abstract

We investigated the effects of ␤-estradiol on the locomotor behavior of female mice in a radial maze. Data comprising the total distance traveled during each arm entry were obtained from video records of six consecutive daily recording sessions. Distributions of these data were bimodal for both ovariectomized control and ␤-estradiol-treated ovariectomized subjects. Data were fit with the sum of two gamma probability distributions. Three parameters of the analytic fits were useful for quantifying the effect of ␤-estradiol on locomotor behavior: (i) the sampling distance (median of the total distance traveled during each arm entry in the short-distance peak of a bimodal distribution), (ii) the committed distance (median of the total per-arm-entry distance traveled in the long-distance peak), and (iii) the partition distance (distance represented by the minimum between the two peaks). Analysis showed that for sampling-distance arm entries ␤-estradiol typically had little if any significant effect on female locomotor behavior, whereas it significantly increased the total distance traveled during committed-distance arm entries on the first 2 days of exposure to the empty maze. ␤-Estradiol also increased the ability of females to discriminate between empty maze arms and arms that contained intact or castrated male mice and partially prevented loss of this capacity after removal of the males.analytic fit ͉ bimodal distribution ͉ estrogen ͉ gamma distribution ͉ exploratory behavior A long history of endocrine and behavioral research with rodents has supported the concept that estrogens foster exploratory and other locomotor responses in females with the effect of increasing contacts with reproductively competent males (reviewed in ref. 1). For example, ovariectomized female rats treated with estrogens increased their willingness to cross electrified grids and to choose appropriate arms of mazes to approach males (2, 3). Nerve cell groups supporting such behaviors have been found in rats in the lateral subdivision of the medial preoptic area (4) and in hypothalamic efferents to the midbrain central gray (5). To make use of current techniques in functional genomics, mice would be desirable for this type of experimentation. Direct approaches by female mice to males in a seminatural environment have been described (6). Genomic studies with mice have emphasized the importance of gene products coding for estrogen receptors ␣ and ␤, oxytocin, the oxytocin receptor, and the vasopressin 1a receptor in the control of affiliative (as opposed to aggressive) behaviors (7-10). If estrogens heighten CNS arousal, the behavioral choice by any individual mouse in any particular situation hinges on the balance between sexual attraction and fear (11). Our mathematical analyses, which separate shorter sampling-distance entries into a maze arm from longer committed-distance entries, address this behavioral distinction.Our analysis shows that the locomotor behavior of female mice significantly changes in the presence of estrogens. However, th...