G. Bernroider, K. Tritscher and M. Grubinger

Institute of Zoology, University of Salzburg, Salzburg, Austria

Computational neuroethology seeks to study neural control algorithms that govern the motoric expression of behaviour. One particularly challenging approach is to simulate time-advanced neural coding principles and relate the result to the dynamics of spatial locations during active behaviour. Similar procedures from invasive physiological registrations have been provided by place field maps derived from hippocampal place cell recordings during signal guided spatial tasks in rats (Burgess et al., Neural Networks, 7, 1065-1081, 1994; Wilson & McNaughton, Science, 265, 676, 1994). In contrast to these studies, the present approach replaces single or mutliple cell recordings in the behaving animal with computer simulated, dynamic population codes derived from a senso-motoric integration model (Bernroider et al., Forma, KTK Scientific Publ., Tokyo, 1996). The model yields a phase-time code with high biological plausibility that precedes the orientation sensitive execution of movement.

The code, expressed as a single, complex number, becomes implanted into allocentric space coordinates representing the subject's spatial location. The dynamic evolution of the code reflects a time-dependent process depending on the polarization of the animal's environment through spatially distributed visual signals. During the subject's active behaviour, e.g. selective approaching or active avoidance of spatial cues, a 2-dim spatial map is composed, with intensities coding the length of the underlying population vector and phase coding the direction at a given instance of time. The method helps to discern spatial context as it applies to the subject's egocentric space representation and offers an intriguing way of combining neural modelling techniques with real behavioural observations. In addition, it allows a multitude of experimental manipulations from interactive computer-controlled access, changing spatial context in dependence of the subject's spatial location in real time. Further, invasive experimental restrictions affecting the behaving animal are completely avoided. Results are demonstrated by spatial maps registered during visual exposure learning in naive quail chicks (Bernroider, Nervous Systems and Behaviour, 443, Thieme Verl. 1995).

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