When industry tried to reproduce published preclinical studies, Amgen could confirm only 6 of 53 landmark studies, and Bayer found inconsistencies in about two thirds of the projects it checked (Begley & Ellis, Nature 2012; Prinz et al., Nat Rev Drug Discov 2011). Mouse Data tackles this by replacing manual, observer-dependent evaluations with AI-driven Digital Researchers. By using standardized hardware and version-controlled Digital Researchers in Mouse Data Lab, we ensure that laboratories on different continents apply identical measurement criteria to their animals.

A Digital Researcher is an AI module based on deep learning and computer vision. Unlike human observers, it does not tire and observes every video in exactly the same way, transforming facial expressions into quantitative, auditable metrics without the need for manual labeling.

The patented Face Analyzer is a specialized hardware system consisting of cubicles designed to encourage mice to face the camera naturally. Its Digital Researchers automatically identify facial patterns that discriminate between states, such as pain versus no pain, without human intervention. Each unit can evaluate up to eight mice simultaneously, providing a continuous index of pain expression.

Mouse Data Lab analyzes every recording with the same version-controlled and calibrated method, whatever the local computing power of each laboratory. Results may differ between labs, but the method of observation is always the same, so data from different sites can be compared.

Yes. Mouse Data provides official Digital Researchers for CD-1 (albino) and C57BL/6 (pigmented) mice. For other strains, you can train your own: the AI identifies patterns by comparing control versus affected groups without requiring the user to manually label thousands of frames, effectively removing the subjective "bottleneck" of human pre-labeling.

Users manage physical infrastructure through the 'Device' tab. Here, you can link your hardware units, manage recording settings, and monitor the connection of local units to the cloud infrastructure for high-performance processing.

Work is organized into 'Projects'. With DR Systems, researchers can share their videos and the exact Digital Researcher version with collaborators, so every site applies the same method of observation, which is key for multi-site preclinical studies.

Credits work as follows:

• Pre-purchase: Users buy analysis credits in advance, as one-off packs valid for four years.
• New packs: When your credits run out, you buy a new pack. Unused credits always add up and only expire after four years.
• Consumption: Credits are deducted only when a Digital Researcher analyzes a video. Only DR Systems use credits; Essential Systems don't.

Data is transmitted to centralized, high-performance computing (HPC) servers. This architecture removes the need for local IT maintenance or expensive high-end GPUs, making AI-driven analysis accessible to any laboratory.

Mouse Data ensures a fully traceable and auditable research trail through:

• Access Control: Role-based access.
• Auditability: Immutable audit logs that link every reported metric back to its original source recording.

Our systems support two of the 3Rs, Refinement and Reduction:

• Refinement: With Face Analyzer, pain is measured with no stimulus and no observer in the room, which avoids the stress caused by human presence.
• Reduction: With DR Systems, videos stay stored in Mouse Data Lab, so they can be analyzed again with new Digital Researchers to test new hypotheses without new animal experiments.

The Mouse Data Community connects the labs that work with our systems. In Mouse Data Lab, you can share DR System videos and Digital Researchers with your collaborators, so they can analyze them with exactly the same method.