SCORHE: a novel and practical approach to video monitoring of laboratory mice housed in vivarium cage racks

Abstract

The System for Continuous Observation of Rodents in Home-cage Environment (SCORHE) was developed to demonstrate the viability of compact and scalable designs for quantifying activity levels and behavior patterns for mice housed within a commercial ventilated cage rack. The SCORHE in-rack design provides day- and night-time monitoring with the consistency and convenience of the home-cage environment. The dual-video camera custom hardware design makes efficient use of space, does not require home-cage modification, and is animal-facility user-friendly. Given the system's low cost and suitability for use in existing vivariums without modification to the animal husbandry procedures or housing setup, SCORHE opens up the potential for the wider use of automated video monitoring in animal facilities. SCORHE's potential uses include day-to-day health monitoring, as well as advanced behavioral screening and ethology experiments, ranging from the assessment of the short- and long-term effects of experimental cancer treatments to the evaluation of mouse models. When used for phenotyping and animal model studies, SCORHE aims to eliminate the concerns often associated with many mouse-monitoring methods, such as circadian rhythm disruption, acclimation periods, lack of night-time measurements, and short monitoring periods. Custom software integrates two video streams to extract several mouse activity and behavior measures. Studies comparing the activity levels of ABCB5 knockout and HMGN1 overexpresser mice with their respective C57BL parental strains demonstrate SCORHE's efficacy in characterizing the activity profiles for singly- and doubly-housed mice. Another study was conducted to demonstrate the ability of SCORHE to detect a change in activity resulting from administering a sedative.

Fig. 1

Thoren Caging Systems Maxi-Miser ventilated double-bay racks found in National Cancer Institute vivariums. The cage features are typical of a mouse home-cage, including a bedding material, b a wire-bar lid, which has both c an integral water bottle basket and d a food basket. The cage is also fitted with e a cage cover, which also serves as an air filter cap

Fig. 2

a SCORHE 3-D computer-aided design model, with component functions indicated by color codes. The unit’s dimensions are given in the Appendix. b A more detailed design drawing of the side assembly, showing placement of the two near-infrared (IR) illumination bars, as well as the assembly covers, which serve three purposes: to secure the two acrylic diffusors, to fasten the side assemblies to the back-plate, and to provide a hinged mount and latch for the front door

Fig. 3

Cage drawing, illustrating the overlapping region visible to both SCORHE cameras (green), as well as the occluded regions: directly beneath the center of the food hopper (pink) and, on both sides of the cage, the partial volume bounded by the cage wall and the side plate of the food hopper (blue)

Fig. 4

a SCORHE installed (yellow arrows) in the cage rack, demonstrating seamless integration while maintaining the full rack cage capacity and b providing easy access to the cage via the hinged door with windows

Fig. 5

SCORHE processing algorithm for a singly-housed mouse. a Read front- and rear-view images from the video files. b Mouse pixels are isolated by the subtraction of fixed background intensities. For each view, the segmentation step retains the single largest binary silhouette (blob). c Features derived from the binary silhouette statistics (e.g., the centroid and ellipse fit parameters, as shown in the middle image) are input to a classifier to automatically determine the placement of the mouse. d If the mouse is found to be on-floor or off-floor, the centroid physical coordinates are estimated via learned mappings (see Fig. 6). e Placement classification and coordinate estimation results are used to calculate and report the mouse activity measures

Fig. 6

Validation system fabricated to obtain precise physical coordinates of the mouse as it moves around the cage, simultaneous to SCORHE video acquisition. a The SCORHE unit is raised by b a clear acrylic box pedestal so that the fields of view of standard lens-equipped cameras c and d encompass the entire cage. The cameras and near-infrared illumination are housed in the e translucent white acrylic base. Views g and i are concatenated, and background subtraction is performed to locate the mouse (annotated with yellow ellipses). The coordinates of the mouse silhouette in the concatenated image are converted to actual physical coordinates via a mapping function obtained through a previous camera calibration procedure. The mouse is simultaneously viewed in the SCORHE images f and h. The blob statistics of the mouse appearing in the SCORHE views are paired with the physical coordinates obtained via the validation system imaging to form a learning set. The learning set is used to generate a mapping function for estimating physical coordinates solely based on SCORHE view blob statistics

Fig. 7

Confusion matrix for SCORHE’s processing of the placement class of a singly-housed mouse. The matrix is based on the automated classifier output for 88,811 manually scored frames

Fig. 8

SCORHE processing algorithm for doubly-housed mice. a Read front- and rear-view images from the video files. The dashed red circles enclose a mouse that is visible to both cameras (i.e., overlapping region between camera views). b Segmentation is performed to isolate fore-ground objects (mouse blobs colored in red and blue) in both views. The blobs are assigned identities on the basis of the presence in overlapping (highlighted in green) and nonoverlapping (highlighted in yellow) regions. Blobs appearing in the overlapping region are matched and assigned the same identity (indicated by the red color), whereas blobs appearing at least partially in nonoverlapping regions are assigned another identity (indicated by the blue color). c If the number of blobs with unique identities is two, then the status is one of no occlusion. If, however, only one unique identity has been assigned, then the status is two occluding mice (i.e., each mouse does not result in a binary silhouette isolated from that of the other mouse). For a status of occlusion, a rough assessment of the occlusion’s severity is obtained by comparing the size of the convex hull of the occluded blob to the overall size of the blob. Occluded blobs in which the size ratio does not exceed an empirically chosen threshold (e.g., the green-colored blob in the middle image) are deemed to be severely occluded, and there is no subsequent attempt to resolve the occlusion. For other occluded blobs with a size ratio exceeding the threshold (e.g., the segmentation output of the left set of frames, shown in red and blue), the perpendicular (solid pink line) to the (dashed pink) line connecting the extreme points (green X annotations) having the longest distance to the blob perimeter is used to identify the occlusion junction (dashed yellow circle). The occluded blob is split along a line passing through the occlusion junction point and parallel to the radial line of a circle centered at the centroid of the occluded blob. All pixels on one side of the dividing line are considered to belong to one mouse, whereas pixels on the other side of the line are considered to belong to the other mouse. d The minimum Euclidean distance between blob centroids in successive frames is used to maintain the identities of the mice in cases of no occlusion or resolved occlusion. If, however, the mice are determined to be severely occluded, the mouse identities are lost. e The blob identities, combined with the blob properties, are used to compute activity measures. One measure of interest is the number of times that a mouse exceeds a vertical threshold (yellow annotation)

Fig. 9

Mouse activity profile comparison for five sets of singly-housed wild-type C57BL/6J strain (WT) and ABCB5 knockout strain (KO). Plot a shows the percentages of time for off-floor and on-hind legs placement. Plot b is the distance traveled measure. The start time for both plots corresponds to the start of the vivarium dark cycle, as indicated by the gray shading

Fig. 10

Three-day average of mouse circadian activity profile for the doubly-housed C57BL/6J strain. The gray shading in the plots indicates the dark cycle. The utility of SCORHE to monitor circadian rhythm is demonstrated by four measures: a rearing counts, b percentages of time inactive, c front-to-rear crossing counts, and d side-to-side crossing counts

Fig. 11

Effect of a 0.667-mg/kg dosage of Acepromazine (a tranquilizer) on wild-type C57BL/6J strain mice: a rearing counts, b percentages of time inactive, c front-to-rear crossing counts, and d side-to-side crossing counts

Fig. 12

Two-day average for doubly housed HMGN1 overexpresser mice (OE) and C57BL/6J wild-type strain mice (WT): a front-to-back crossing counts, and b side-to-side crossing counts

Fig. 13

Dimension drawings of the SCORHE unit. a SCORHE unit in rack, with components color-coded for clarity. b Top view of the SCORHE unit out of the rack. c Front view of the SCORHE unit in rack. The shelf rail on which cages are suspended while in the rack is shown in more detail. d Side view of the SCORHE unit in the rack. Measurements of all color-coded elements are also included.