Evaluation of two automated metrics for analyzing partner preference tests

Abstract

The partner preference test (PPT) is commonly used to examine sexual and social preferences in rodents. The test offers experimental subjects a choice between two stimulus animals, and time spent with each is used to calculate a preference score. In monogamous prairie voles (Microtus ochrogaster), the PPT has been paramount to the study of pair bonding. Although powerful, use of the PPT in voles has depended primarily on human manual scoring. Manual scoring is time-consuming and is susceptible to bias and fatigue, limiting the use of the PPT in high-throughput studies. Here we compared manual scoring (real-time and 16x) and two automated scoring metrics: "social proximity" and "immobile social contact". We hypothesized that "immobile social contact" would provide data most comparable to manually scored "huddling", and thus be the most sensitive measure of partner preference in prairie voles. Each automated metric produced data that highly correlated with manual scoring (R>0.90); however, "immobile social contact" more closely reflected manually scored huddling (R=0.99; P<0.001). "Social proximity" and "immobile social contact" were then used to detect group partner preferences in four data sets that varied by cohabitation length and sex. "Immobile social contact" revealed a significant partner preference in each data set; "social proximity" detected partner preferences in only three of the four. Our results demonstrate the utility of automated systems in high-throughput PPTs, and further confirm that automated systems capable of scoring "immobile social contact" yield results indistinguishable from manual scoring.

Figure 1

Partner preference tests (PPTs) were conducted in a three-chambered arena (A). Stimulus animals (one partner and one stranger) were loosely tethered at each end; the test subject roamed freely for 3 hours. 12 PPT boxes were monitored by four cameras (B). The four video streams are compressed to a single output by a QUAD box and sent to both a DVD recorder and PC.

Figure 2

An early version of EthoVision (version 3.0) and the newest version of SocialScan (version 2.0) were used to automatically process vole behavior in the partner preference test (PPT). Virtual arenas (area in which voles can be tracked) and zones (regions of interest) were outlined for EthoVision (A) and SocialScan (B). For EthoVision 3.0, arenas were truncated at a point beyond which the tethered animals could not reach; zones included LeftSocial, RightSocial, two transitions zones, CenterChamber, and NonSocial (an accumulation of all non-hidden zones); the social zones were designated as hidden, allowing the program to track the amount of time the test subject spent in “social proximity” to the left and right tethered animals. For SocialScan 2.0, arenas included the entire test box floor; zones included LeftSocial, RightSocial, Center, and NonSocial; the NonSocial zone served to correct occasional ID swaps. During analysis, both EthoVision 3.0 and SocialScan 2.0 tracked the animals’ location according to center-of-gravity. EthoVision 3.0 logged time in “social proximity” (C, see middle test box); SocialScan 2.0 logged time in “immobile social contact” (D, see middle test box).

Figure 3

Dot-plots and linear correlations were generated for key scoring method comparisons to illustrate variation. Plots demonstrate that there is high inter-rater reliability in real-time (A) and intra-rater reliability across scoring speeds (B). The metric of “immobile social contact” (D) approximated manually scored “huddling” better than the metric of “social proximity” (B). Equations (y = mx + b) and Pearson’s Rs illustrate how closely the linear models approximate a 1:1 ratio with no scale shift and the degree of concordance, respectively.

Figure 4

Males and females underwent two separate 3-hour PPTs, once after 24-hours of cohabitation, and once after a full week of cohabitation. All tests were processed post hoc and assessed for “social proximity” time (A) and “immobile social contact” time (B). The “immobile social contact” metric detected a significant partner preference in all four data sets; the “social proximity” failed to detect a partner preference in males after 1 week of cohabitation. Bar-graphs represent mean + SEM; (*) represents a statistically significant preference for the partner over the stranger, P < 0.05. “ns” = not significant.

Figure 5

Males were assessed for aggression in twenty-four 15-minute video segments, half during the post-24 hour cohabitation PPT test and half 1-week PPT. Males spent more time in aggressive scuttles with stranger females than with their female partners. Bar-graphs represent mean + SEM. (*) P < 0.05.