M.R. Kruk

Department of Medical Pharmacology, University of Leiden, Leiden, The Netherlands

Ethopharmacological research is based on the implicit assumption that the detailed structure of behavior in so-called natural settings' reflects the organization of mechanisms in the brain which control such behaviour. Following this reasoning, detailed description of the effects of drugs in such settings should reveal functions and mechanisms of action of specific neurotransmitter and neuromodulator systems in the brain. The potential merits of this kind of approach are handicapped by practical and theoretical problems.

An important problem is the time- and labour-consuming nature of data acquisition. This problem is being solved for simple behavioural measures, and further automation may solve this problem for more complex behavioural patterns in the future. One advantage of such automation is that differences in judgement between raters will have less effect on behavioural analysis. Automation of the acquisition of complex behavioural patterns will also force the ethologist to reconsider the validity of established ethograms, which are based on largely intuitive interpretations of behaviour.

A more profound problem is how to deal analytically with the wealth of data obtained by automated data acquisition. Eliminating the personal bias of the observer by automation does not necessarily eliminate the arbitrariness of the behavioural classification system used. Different items scored by any acquisition system may be statistically dependent, either by the configuration used as 'natural' setting, or because such items share control mechanisms within the central nervous system. Many analytical methods used do not test for such dependencies, and the analytical methods that systematically do test for such dependencies such as Markov methods, are - incorrectly - seen as difficult, but they can be time-consuming. However, that problem can be remedied in part by advanced software packages. Moreover, these methods provide powerful ways to investigate the validity of our behavioural classifications statistically. In our laboratory, they have been extremely useful to develop concepts on behavioural brain mechanisms:

These results suggest that the basic assumptions of ethological brain research on the usefulness of detailed behavioural data acquisition and description followed by appropriate detailed analyses, is indeed a powerful tool to study brain mechanisms involved in the control of behaviour. This methodology also may provide with the conceptual tools for the next step in behavioral analysis, which is - in my view - to relate behavioral items to non-behavioral sequential events such as electrical or chemical brain activity.

Paper presented at Measuring Behavior '96, International Workshop on Methods and Techniques in Behavioral Research, 16-18 October 1996, Utrecht, The Netherlands