Building the data foundation for safer surgery with AI
Cortical Dynamics is a Perth based neurotech company developing AI powered brain monitoring to make surgery under anaesthesia safer.
The problem
Real world patient data is costly, ethically complex and slow to collect, and it is hardest to get for exactly the events that matter most: rare intraoperative complications. The platform also needed to generalise to United States patient populations for global market entry, which meant the training data could not come from one region alone.
On top of that sat a young company's infrastructure question. Standing up and running clinical grade data infrastructure is not where a neurotech team's scarce engineering time should go.
What we did
- Designed a synthetic dataset strategy covering diverse demographics and rare intraoperative events, so model development was not blocked waiting on patient recruitment.
- Stood up an Azure Databricks lakehouse in an Australian region, with the governance and access controls clinical data requires.
- Set a phased AI roadmap sequenced against regulatory milestones rather than against what was technically interesting.
- Supported the grant application through the Biomedical AI Sprints Accelerator, run with ARM Hub and MTPConnect.
Where it landed
- Synthetic clinical datasets covering diverse demographics and rare events
- Scalable, sovereignty compliant cloud storage live and operational
- Phased AI roadmap aligned to regulatory milestones
- A practical pathway toward clinical trial readiness
Our view
In healthcare AI, innovation alone is not enough. Progress depends on the right data foundations, regulatory foresight, and a commercially sustainable path to scale. Getting the sequence right at the start is cheaper than re-architecting after a trial has begun.
Similar problem, different industry?
Scarce data, a regulator to satisfy, and a small team. It is a common combination.