Invited Symposium: Nonlinear Dynamical Systems in Psychiatry



Materials & Methods


Discussion & Conclusion



INABIS '98 Home Page Your Session Symposia & Poster Sessions Plenary Sessions Exhibitors' Foyer Personal Itinerary New Search

A Dynamical Analysis of Action Selection in the Laboratory Mouse

Contact Person: Agnes Guillot (Agnes.Guillot@lip6.fr)

Materials and Methods

Although our work dealt with a twelve-hour chronological succession of 10 different general acts, recorded in 20 male mice under day and night conditions (Fig.1), this study only addresses eight behavioral sequences composed of six different acts (sniffing, locomotion, feeding, drinking, nest building and one category of grooming) within the first nocturnal activity bouts - the longest ones - of each sequence.

Fig. 1: Graphic representation of the behavioral sequences of one mouse under day condition (top) and one under night condition (bottom), observed during about 12 hrs. Ten acts are plotted on the Y-axis and are ranked on the basis of their energy costs, from lowest to highest: Rest, two categories of grooming, nest-building, two other categories of grooming, sniffing, feeding, drinking and locomotion.

In order to assign a common currency to each act, each sequence was translated into a succession of metabolic costs associated with the ongoing acts, recorded with a 15-sec time step (Fig.2).

Fig. 2: Experimental time series of metabolic costs (J/sec), plotted against a 15-sec time step, which correspond to the first nocturnal activity bouts of eight mice.

Counter-hypotheses of chaos - i.e. randomness, linearity and long-term predictability - were tested by both qualitative (Sugihara & May, 1990) and quantitative nonlinear methods (Kennel & Isabelle, 1992). The Sugihara & May algorithm tests the long-term predictability of the data by using the first part of a given time series as a model to predict the data of the second part. The 'Noise Versus Chaos' algorithm of Kennel & Isabelle statistically verifies that the prediction errors of the actual data are lower than those of surrogate data (random data with the same length and average power spectral density as the original data sets).

Back to the top.

<= Introduction MATERIALS & METHODS Results =>

| Discussion Board | Next Page | Your Symposium |
Guillot, A.; Meyer, J.A.; (1998). A Dynamical Analysis of Action Selection in the Laboratory Mouse. Presented at INABIS '98 - 5th Internet World Congress on Biomedical Sciences at McMaster University, Canada, Dec 7-16th. Invited Symposium. Available at URL http://www.mcmaster.ca/inabis98/sulis/guillot0208/index.html
© 1998 Author(s) Hold Copyright