Issue link: https://read.uberflip.com/i/1545817
JAX ® Envision™ A Platform Built for Home Cage Research Figure 3: Percent time climbing determined by algorithm validated against "ideal curated data" (left) vs. Envision (right), shown for homogenous cages (all saline, all LPS) and mixed cages (saline and LPS treated housed together). Dosing at dotted line. Figure 5: Percent time huddling in LPS model. LPS-treated mice show increased huddling relative to saline controls, consistent with inflammatory response and known thermoregulatory disruption. Figure 4: Full suite of Envision 5.0 Measures Learn More JAX.ORG/ENVISION TS0290 2026.07 1.800.422.6423 1.207.288.5845 sales@JAX.org PHYSIOLOGICAL MEASURE CAGE & RESOURCE MO NITORING Activity Suite BEHAVIOR CLASSIFICATIONS MOVEMENT Locomoting Huddle Active Inactive Drinking Feeding Inferred Sleep Total Distance Traveled Climbing Activity Undetected Social Distance Respiratory Rate Motion Estimation Water Level Food Level SIGNAL-TO-NOISE MATTERS IN HOME CAGE RESEARCH Training and validating on the full range of real conditions – not a curated set – yields measurably cleaner data. Better data quality means more than aesthetics: lower measure variance translates directly into greater statistical power to detect effects – and the ability to distinguish a biological signal from measurement noise. That's the difference between an ambiguous signal that can require a repeat study versus a clear endpoint to confidently support decisions. To illustrate this point, the figures below show percent time spent climbing from the exact same video run on a pipeline validated using "ideal curated data" compared to Envision validated using LENS. With Envision, the single LPS-treated animal (gold line) in a mixed cage is clearly distinguishable from its saline control cage mates – a separation the less rigorously trained algorithm obscures. This doesn't just complicate study interpretation, it can cause a cascade of misleading study conclusions, unnecessary study repeats, and incorrect go/no go decisions. A GROWING SUITE OF VALIDATED MEASURES Robust algorithms provide the foundation for digital measures that deliver interpretable and reproducible performance across diverse study conditions. Digital measures allow researchers to move beyond manual observation and intermittent testing by continuously quantifying individual animal behavior in the home cage. Envision is designed to provide a range of digital measure offerings for researchers starting from core capabilities applicable across any study type to advanced measures necessary for specific contexts of use. Envision 5.0 supports a validated suite of 10+ behavioral and physiological endpoints – spanning measures of activity, physiology, and social behavior – continuously and non-invasively. Activity with Best-Case Validation Data Activity with Envision™ LENS Validation