The VIKING project Veien Til Kunstig Intelligens I Klinisk Nevrofysiologi
In the VIKING project, our goal is to establish the world’s largest database of clinical neurophysiological data within an open infrastructure designed to facilitate research and AI development. These data will then be used to train AI-based decision support systems for clinical neurophysiological practice. Additionally, we aim to develop a user-focused roadmap for creating and implementing AI-based decision support systems to advance the diagnosis and management of neurological diseases.
Newsletters
We attended the Norwegian Arendalsuka 2025.
You can see the panel discussion on the use of hospital data for AI-research.
Direct link to the video (vimeo.com)
Patient Information and Rights
Patient information and patient rights (English)
Pasientinformasjon og –rettigheter (Norsk)
Background
Norway is in a unique international position due to the high level of standardization between hospitals and laboratories, allowing for the compilation of large volumes of clinical data from neurophysiological examinations. In the VIKING project, we collect data from hospitals all across Norway, representing a diverse cohort of hundreds of thousands of patients.
We will also enrich our data by leveraging Norway’s high-quality health registries. Norway’s health registries are among the most comprehensive in the world, offering a wealth of high-quality data spanning several decades. These registries provide unparalleled insights into disease patterns and treatment outcomes across the population, creating a unique opportunity to drive healthcare advancements through data-driven research and innovation.
Artificial intelligence (AI) is set to transform healthcare, driven by advancements in analytical techniques and computational power. AI has the potential to uncover clinically significant insights hidden within vast amounts of medical data, thereby enhancing clinical decision-making. By harnessing health data and integrating diverse data sources, AI can enable precision medicine, improve patient outcomes, and provide real-time decision support by identifying patient-specific patterns of disease progression. In the VIKING project we will leverage AI to develop machine-learning based models to
- Recognize specific disease patterns in nerve conduction studies
- Classify measurements from electromyography, quantitative sensory testing and evoked potentials into normal and abnormal
- Automatically measure recorded evoked potentials
- Classify EEG measurements for both diagnostic and prognostic purposes
On-going projects related to these aims:
The second part of the VIKING project is split into 3 sections: The first section focuses on clinicians’ and patients’ knowledge, perspectives, attitudes, expectations, thoughts, and trust regarding AI-based decision support systems in clinical neurophysiology. The second section addresses the ethical challenges we face and how these can be addressed. The third section consolidates this new knowledge into a user-oriented roadmap for the development and implementation of AI-based decision support systems for neurological diseases diagnosed through neurophysiological examinations.
Milestones (accomplished per 2026):
- Establish the necessary IT infrastructure for secure, legal, and ethically responsible data processing
- Establish data processing agreements and collect neurophysiological data from patients across Norway.
- Connect to patient registries, preprocess and prepare the data. Establish a protocol for updating the database.
The project aim to answer the questions:
- What knowledge, perspectives, expectations, and attitudes do users (patients, clinicians) have regarding the introduction of an AI-based decision support system for neurophysiological examinations?
- How can we promote trust in AI-based decision support systems among users (patients, clinicians)?
- What ethical challenges exist for the development, testing, and implementation of AI-based decision support tools for neurophysiological examinations, and how can they be addressed?