Validation and implementation of a novel high-throughput behavioral phenotyping instrument for mice

Abstract

Background: Behavioral assessment of mutant mouse models and novel candidate drugs is a slow and labor intensive process. This limitation produces a significant impediment to CNS drug discovery.

New method: By combining video and vibration analysis we created an automated system that provides the most detailed description of mouse behavior available. Our system (The Behavioral Spectrometer) allowed for the rapid assessment of behavioral abnormalities in the BTBR model of Autism, the restraint model of stress and the irritant model of inflammatory pain.

Results: We found that each model produced a unique alteration of the spectrum of behavior emitted by the mice. BTBR mice engaged in more grooming and less rearing behaviors. Prior restraint stress produced dramatic increases in grooming activity at the expense of locomotor behavior. Pain produced profound decreases in emitted behavior that were reversible with analgesic treatment.

Comparison with existing method(s): We evaluated our system through a direct comparison on the same subjects with the current "gold standard" of human observation of video recordings. Using the same mice evaluated over the same range of behaviors, the Behavioral Spectrometer produced a quantitative categorization of behavior that was highly correlated with the scores produced by trained human observers (r=0.97).

Conclusions: Our results show that this new system is a highly valid and sensitive method to characterize behavioral effects in mice. As a fully automated and easily scalable instrument the Behavioral Spectrometer represents a high-throughput behavioral tool that reduces the time and labor involved in behavioral research.

Keywords: Automated detection; Behavioral screening; Ethological analysis; Grooming; Neurobehavioral assessment; Video analysis.

Fig. 1

Diagram of the Behavioral Spectrometer. The box contains a camera, an instrumented floor to measure vibrations and a row of photobeams to capture rearing behavior. One mouse is run at a time and behavior identification takes place automatically in real-time.

Fig. 2

Analysis of human verses computer scoring shows excellent correspondence. Human scores represent the average score of two observers for each mouse expressed as a mean (+S.E.M.) of all mice over 10min. Computer score is the score for the same mice expressed as the mean (+S.E.M.). Inset graph shows the relationship between the human average score and the computer score across all behaviors measured. R value computed from Pearson’s Product Moment Correlation test. N = 10 CD1 mice for frequency graph, N=20 behaviors for inset graph.

Fig. 3

Automated measures are stable within mice and consistent between groups. (A) Behavioral Spectrometer data of the 23 categories is plotted for the same mice run on two consecutive days for 20 min. Each point represents the mean score of a behavior for the group plotted against the score the following day. N =6 CD1 mice for each of 23 behaviors. (B) Data obtained from two different cohorts of mice are plotted against each other. N=6 (cohort 1) and 12 (cohort 2) for each of 23 behaviors over 8 min. R value computed from Pearson’s Product Moment Correlation test.

Fig. 4

Wet mice displayed more grooming. Data obtained from the Behavioral Spectrometer are shown as number of seconds of behavior scored by behavioral category for control (dry) and water sprayed (wet) mice (mean +S.E.M.) over 20 min. Wet mice showed more grooming of the nose, head, face, leg, back and tummy categories. Statistical comparisons were made between wet and dry mice for each category. N = 12 CD1 mice per treatment, *p < 0.05.

Fig. 5

Stressed mice displayed much more grooming and less locomotor behavior. Data obtained from the Behavioral Spectrometer are shown as number of seconds of behavior scored by behavioral category for control (unrestrained) and stressed (2 h of prior restraint) mice (mean +S.E.M.) over 20 min. Stressed mice showed more grooming in all categories and less locomotor activity. Small decreases were also evident in the Rear_Bob, Orient_Sniff and Orient_Look categories. Statistical comparisons were made between control and restrained mice for each category. N = 12 C57 mice per treatment, *p < 0.05.

Fig. 6

BTBR mice displayed more grooming and less rearing. Data obtained from the Behavioral Spectrometer are shown as number of seconds of behavior scored by behavioral category for control (C57) and BTBR mice (mean +S.E.M.) over 10 min. BTBR mice showed more grooming in all categories except cheek. BTBR mice also displayed dramatically less rearing behavior. Statistical comparisons were made between control and BTBR mice for each category. N = 12 mice per treatment, *p < 0.05.

Fig. 7

Representative ethograms of the BTBR mouse. These two panels show the second-by-second categorization of behavior produced by the Behavioral Spectrometer in real time. Upper panel shows an ethogram of a control (C57) mouse and the lower panel shows a BTBR mouse. Behavioral (y-axis) events are marked by time (x-axis) in a color coded fashion. Locomotor activity is marked in red, rearing activity in yellow, orienting in green, vigorous grooming in blue, fine grooming in purple and still in black. Comparing ethograms shows less rearing and more grooming in the BTBR (lower panel) compared to the control mouse (upper panel). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 8

Inflammatory pain produces a profound and reversible effect on behavior. Data obtained from the Behavioral Spectrometer are shown as number of seconds of behavior scored by behavioral category for control (saline i.pl. + saline i.p.), inflamed (Carrageenan i.pl. + Saline i.p.) and treated (λ-Carrageenan i.pl. + naproxen i.p.) mice (mean +S.E.M.) over 20 min. Mice in pain showed a dramatic inhibition of all behavior except sitting still. When naproxen was co-administered with the irritant, behavior was returned to normal levels (with small increases in walk and trot). Statistical comparisons were made between groups within each behavioral category with control mice serving as the control comparison. N = 12 CD1 mice per treatment, *p < 0.05.

Fig. 9

Behavioral spectrograms. The effect of each treatment was computed as the change in behavior relative to control expressed as a multiple of the standard deviation (SD) of the mean of the control values. The depiction is color coded to correspond to degree of change in the occurrence of the behavior, with dark blue representing an increase of more than 2 SDs, light blue between 1.5 and 2 SDs, green between 1 and 1.5 SDs and red representing as decrease of more than 2 SDs, orange between 1.5 and 2 SDs and yellow between 1 and 1.5 SDs. Results that were not statistically significantly different from control values or where the difference was less than 1 SD are plotted in white. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)