Physical services
Overview
At the frontier between breakthrough innovation and patient safety, validation is the decisive threshold between concept and clinical credibility.
Within the TEF-Health network, the collaboration between Fraunhofer-Institut für Integrierte Schaltungen IIS and Sonovum GmbH illustrates how structured physical testing accelerates the safe deployment of AI-enabled medical technologies.
Sonovum set out to develop a non-invasive device for measuring intracranial pressure in patients suffering from severe brain trauma. The system builds upon CE-marked medical technology (TRL 10) capable of monitoring brain pulse curves, while integrating a novel proprietary biosignal and AI-based classification models.
Yet, despite its technological maturity, the innovation required rigorous validation in a clinical setting to substantiate performance, clarify regulatory positioning, and strengthen market credibility.
Through TEF-Health, Fraunhofer IIS provided access to its Multimodal Biosignal Laboratory, enabling controlled physical testing and feasibility studies that moved the technology decisively forward on its path to clinical adoption.
Impact
The collaboration delivered measurable and strategic progress.
Through structured testing and methodological refinement, Sonovum’s AI-based classifier advanced from TRL 6 to TRL 8, demonstrating system validation in a relevant clinical environment. The company is now conducting a clinical study aimed at completing the final progression toward TRL 10.
Beyond the TRL advancement, the impact unfolded across multiple dimensions:
• Performance evidence generation under controlled clinical conditions
• Refinement of machine learning algorithms through expert-supported testing
• Clarification of regulatory pathways in the absence of predicate devices
• Strengthening of investor and stakeholder confidence
• Acceleration toward real-world deployment.
For an SME operating in a highly regulated medical domain, these steps are absolutely foundational to market entry.
By anchoring innovation in evidence, Sonovum transformed technological promise into clinically grounded credibility.
The Challenge
Measuring intracranial pressure non-invasively represents a significant unmet medical need. However, Sonovum’s approach relied on a proprietary biosignal for which no existing predicate medical device was available.
This absence created a dual complexity:
1. Scientific validation challenges, the need to prove that the novel biosignal reliably correlates with intracranial pressure indicators.
2. Regulatory navigation challenges, the requirement to interpret and align with concurrent regulatory frameworks without a direct comparator device.
Additionally, Sonovum sought to move beyond purely expert-based signal curation toward a data-driven approach, integrating AI models capable of classifying and interpreting complex biosignal patterns.
In such a context, laboratory simulation alone would not suffice. The company required structured physical testing in a clinical environment capable of generating defensible, regulator-ready evidence.
The Solution
Through TEF-Health’s service catalogue, Fraunhofer IIS provided access to its Multimodal Biosignal Laboratory, a controlled clinical testing environment designed for advanced physiological signal analysis.
The collaboration combined:
• Controlled physical testing in clinical conditions
• Feasibility studies to assess performance robustness
• Methodological consulting to refine software architecture
• Optimization of machine learning algorithms
• Scientific communication support to strengthen evidence positioning.
This integrated validation framework allowed Sonovum to systematically test, iterate, and improve its AI-based classifier solution.
By grounding development in structured evidence generation, the partnership reduced uncertainty, clarified compliance pathways, and accelerated technological maturity.
The case exemplifies TEF-Health’s broader mission: enabling European medical innovators to bridge the gap between scientific ambition and clinical reality. Because in healthcare AI, trust is built through testing.
