Reliability of two activity boxes commonly used to assess drug induced behavioural changes.

Rats were given different drug treatments known to affect central catecholamine neurotransmission and to induce different types of behavioural changes which were recorded simultaneously by two different types of conventional activity boxes: an Animex activity meter and a photocell cage. All animals were also visually observed simultaneously with the automatic recordings. It was found that the two activity boxes reflected the behavioural changes differently and that the results from the two boxes were not correlated. When comparing the observations with the automatic recordings it was found that some clearly observable changes in behaviour were not reflected as changes in the automatic recordings and conversely that increases or decreases in the recorded motor activity were not always related to any particular changes in behaviour. The results show that motor activity can not be regarded as a simple or homogeneous behaviour that is reliably measured in conventional activity boxes. It is an undescriptive measure consisting of an artificial summation of those components of behaviour that affect the movement-detecting device in the particular box which is used. Practical and theoretical implications of this are discussed.