Some spatial characteristics of behavior: A new technology.

Abstract

A methodology consisting of an experimental chamber, operational definitions of behavior categories, and methods for quantitative analysis of the use of space by small rodents is presented. The chamber consists of a large open box in which the location of the animal was determined by a grid of 24 x 24 infrared beams. Recording was achieved by sampling every 5 seconds, and storing the status of the 48 photosensors in a disk file for later analysis. Each sample produced a binary matrix with 576 cells identifying the location of the animal at that time. From this sequence of matrices a number of behaviors were defined and described. Exploration was measured as the cumulative percentage of unique cells visited by the subject during the observation period. Monotonically decelerating curves were found, with exploration during the first day in the chamber significantly different from that on the following three days, according to the Kolmogorov-Smirnov test. Time allocation was measured as the relative frequency of visits to individual cells computed for various time intervals. Three dimensional charts with spatial location along the X and Y axis, and relative frequency in the Z axis were used to represent time allocation to space or "spatial preference". These surfaces were transformed into vectors and compared with the same goodness-of-fit test. The operational definition of time allocation as well as the methods for comparing the surfaces were found adequate to reliably describe nesting and patrolling patterns of Ratus Norvegicus. Finally, activity was defined as the amount of space visited as a function of time. This measure of rat of location change was computed in 15 min. intervals for 15 consecutive nights. A consistent 3-hour cycle of activity was detected using ARIMA methods. The cycle was accurately described and predicted with a model containing one periodic and one nonperiodic components. The usefulness of this nonintrusive technology for digitizing behavior in space is discussed in terms of possible applications to studies on learning, behavioral pharmacology, and ethology.