Automated behavioral analysis of mice using INTELLICAGE: Inter-laboratory comparisons and validation with exploratory behavior and spatial learning

H.-P. Lipp¹, O. Litvin², M. Galsworthy¹, D.L. Vyssotski², A.L. Vyssotski¹, A.E. Rau¹, F. Neuhäusser- Wespy¹, H. Würbel³, R. Nitsch² and D.P. Wolfer¹

¹Institute of Anatomy and ²Dept. of Psychiatry Research, University of Zürich, Zürich, Switzerland
³Institute of Veterinary Physiology, University of Giessen, Giessen, Germany

INTELLICAGE™ is a large home cage containing four complete operant conditioning units placed in the corners. A central computer continuously controls and monitors activity and learning of up to 16 transponder-tagged mice per cage without human interference.

In a recent standardized multi-lab study, we noted identical strain rank order in the open-field, elevated Null-maze, water maze and object exploration, but comparison of absolute values revealed significant differences between laboratories, or strain-by-laboratory interactions. A subset of 46 female mice was also tested in 4 INTELLICAGES, located in different animal facilities of the University of Zurich (2 cages per site), each unit housing 11- 12 transponder-tagged mice from different strains. The system measured continually visits of drinking sides, development of place preference and activity parameters. INTELLICAGE revealed significant strain differences in initial exploratory activity and baseline activity during the first day. Most importantly, the behavioral scores of the strains as observed in both laboratories were statistically indistinguishable.

We then compared the first 10 min after introduction to INTELLICAGE with standard tests of exploration lasting about the same time span. The measure in INTELLICAGE was the number of visits to the yet unfamiliar test corners. It was positively correlated with visits to the center in the open-field. For the elevated Null-maze, the number of head dips correlated strongly with the number of corner visits in INTELLICAGE, while measures of anxiety such as time in the protected area did not correlate significantly. In the water maze (WM), we found a significant positive correlation between average escape latencies and place preference learning in INTELLICAGE. No correlations were found with WM probe trial scores. These findings indicate that either test measures the acquisition of spatial learning, but the WM testing took four man-weeks, while INTELLICAGE showed the same findings after 24 hours without presence of an experimenter.

We conclude that optimal standardization and comparability is best achieved by the use of automated procedures, and that such automation measures the same behavioral dimensions as standard manual tests, yet with much less stress for animals and experimenters.

Supported by SNF, the NCCR “Neural Plasticity and Repair” and the Foundation 3R.

Paper presented at Measuring Behavior 2005 , 5^th International Conference on Methods and Techniques in Behavioral Research, 30 August - 2 September 2005, Wageningen, The Netherlands.

© 2005 Noldus Information Technology bv