- Actual Analytics
- 7 days ago
- 3 min read
When it comes to behavioural safety,
What you see depends on when you look.
That principle sits at the heart of a new multi-company study — just published in Frontiers in Toxicology — led by Actual Analytics in collaboration with the NC3Rs, AstraZeneca, GSK, Janssen, ICON, Charles River Laboratories, and the University of Edinburgh.
The findings matter not only for behavioural science, but for the economics of drug development —Because in drug development, missed CNS safety signals aren’t just gaps in knowledge — they are multimillion-dollar failures.
The wider problem: CNS safety attrition is extremely expensive
Behind every approved medicine is a long pipeline of candidates that didn’t make it. Industry-wide:
Around 90% of drug candidates entering clinical trials ultimately fail.
Safety and toxicity account for a substantial proportion of those failures — often around a third in some analyses.
The effective cost of each approved drug exceeds US $1–2 billion once the cost of failures is included.
This means a single late-stage safety surprise can wipe out hundreds of millions of dollars of sunk investment and delay promising medicines reaching patients. Even small improvements in the sensitivity, timing, or translational relevance of preclinical safety assessments can materially reduce that risk. And that is exactly the gap this new study addresses.
The behavioural blind spot
Traditional neurofunctional tests such as Irwin and the Functional Observational Battery (FOB) rely on brief, observer-led sessions performed in bright artificial light — usually during the day, when nocturnal rodents are least active.
This creates a behavioural blind spot where:
dark-phase activity goes unmonitored
multi-day patterns cannot be observed
subtle CNS effects may never be captured at all
The core question behind the study:
Can continuous, automated home-cage monitoring (HCA) reveal CNS-relevant changes that episodic tests systematically miss?
The Study
In this blinded, multi-site validation, three historical compounds with known clinical outcomes were re-tested using the Actual Analytics Home Cage Analyser (HCA).
Across all three compounds, HCA detected behavioural changes that Irwin/FOB assessments did not — including:
robust dark-phase effects (suppression or hyperactivity)
signals that emerged or persisted over several days after dosing
clinically relevant changes at doses previously described as “well tolerated"
These data show that CNS-relevant signals often sit in the dark phase or unfold over days — precisely where Irwin and FOB are blind
The Results
Dark-phase matters—a lot
The biggest signals often live in the night-time window that episodic tests don’t capture. Robust dark-phase effects (e.g., suppression or hyperactivity) were captured that standard Irwin/FOB windows simply missed.
Effects unfold over days, not minutes
Several behavioural changes persisted for multiple nights post-dose. Snapshot tests cannot capture these trajectories.
Continuous monitoring provides different data — not just more of it.
Sensitivity at clinically relevant doses
HCA revealed behavioural changes at doses deemed safe by Irwin/FOB — one of which aligned with a human adverse event. Earlier insight = reduced likelihood of costly surprises downstream.
Integrates smoothly into existing repeat-dose toxicology workflows
HCA is operationally compatible with repeat-dose environments: standard husbandry, social housing, daily dosing, and in-life procedures. You get both non-evoked behaviour and how animals respond around evoked events (e.g., handling/blood sampling). You don’t need a bespoke experiment to get better behavioural safety data—you can add HCA to what you already do.
Better science with better welfare (3Rs)
Minimal handling and group housing reduce stress artefacts and improve reproducibility.
Richer continuous data supports refinement and reduction.
These results highlight a practical way forward: integrate automated home-cage monitoring alongside existing neurofunctional assessments, build internal evidence, and ultimately replace subjective snapshot tests with richer, reproducible data.
What this means for drug developers
Every behavioural signal you fail to see in rodents increases the likelihood of encountering it later — where consequences are much more expensive. This study demonstrates a practical, evidence-backed way to:
detect subtle, early CNS safety effects
improve the reliability of translational decision-making
reduce the probability of safety-driven Phase I/II failures
obtain richer data without adding burden to in-life operations
improve animal welfare and study reproducibility
For R&D organisations navigating billion-dollar development cycles, early visibility of CNS risk is not a luxury — it’s a risk-reduction strategy.
What’s next
With continued validation, HCA can complement — and in some cases replace — subjective snapshot assessments like Irwin/FOB within GLP-style toxicology programmes.
It offers a more sensitive, more humane and more economically efficient route to understanding CNS safety risk.
Acknowledgements
Huge thanks to our collaborators at NC3Rs, AstraZeneca, GSK, Janssen, ICON, Charles River, and the University of Edinburgh.
About Actual Analytics
Actual develops advanced home-cage behavioural analytics for socially housed rodents.
The Home Cage Analyser (HCA) delivers 24/7, automated, objective data streams that enhance safety pharmacology, toxicology and neuroscience research — improving translational reliability and animal welfare.