Careers at UMC Utrecht
Vacancy number: 2021/01242
Apply to: September 30 2021
Education: University
Employment: Temporary position
Full-time/Part-time: 36 - 36 hours

brainXplain: unravelling the heterogeneity in causes and consequences of cerebral small vessel disease with explainable AI

Cerebral small vessel disease (SVD) is a vascular health care problem that represents a huge health burden in stroke and dementia. Approximately one quarter of all ischemic strokes and most primary intracerebral haemorrhages are caused by SVD. Furthermore, SVD is the most common cause of vascular dementia and often co-occurs with Alzheimer’s disease (with additional cognitive impact), and thus contributes to about 50% of dementias worldwide.

Key information to better support personalised diagnosis and treatment is embedded in MR images and should be extracted from big data sets with machine learning techniques. Lesion-symptom mapping technology is used to quantify the association between SVD lesion burden and potential causes or consequences on a group level. Some important technical challenges need to be addressed to apply this on an individual level: taking into account the full spectrum of SVD lesions and multi-variate outputs. In this project, these challenges will be addressed with innovative explainable machine learning techniques.

Project description

Machine learning and artificial intelligence (AI) techniques are often considered ‘black box’. It can be difficult to understand their internal workings, which affects how medical decisions based on such algorithms are received. Modern technology, also known as "explainable AI”, can be used to gain insights into the decision making of AI. In SVD, it can be used to assess multiple lesion types, have multi-variate output, and simultaneously explain which lesion locations are most strongly associated with the output.

Your role in this project will be to develop and evaluate explainable AI technology. You will implement a number of explainable AI techniques that pinpoint which lesion locations are most strongly associated with the output (e.g. causative factors or clinical outcomes). Next, these methods will be validated on international datasets of thousands of patients of the Meta-VCI-Map consortium. Finally, a number of pilot studies will demonstrate potential applications assessing the relation between burden/distribution of SVD lesions and aetiology, cognitive outcomes, and prognosis.


You will be working at the internationally renowned Image Sciences Institute at the UMC Utrecht, which is housed within the Imaging Division. You will work in a team of PhD candidates and post-docs in the field of medical image analysis and machine learning. In this interdisciplinary project, there will be close collaborations with the Departments of Radiology and Neurology of the UMC Utrecht; and international partners of the Meta-VCI-Map consortium.


You are an excellent candidate with an MSc in computer sciences, medical imaging, physics, mathematics, biomedical engineering or a similar field. You have a strong interest in medical image analysis and machine learning. You have a good scientific background, are highly motivated and independent, able to work in an interdisciplinary team of engineers and medical doctors. Programming experience is required and knowledge of machine learning is an advantage.

We believe in the power of a diverse team in which there is room for different skills, expertise, and social and cultural backgrounds. We invite you to respond to this vacancy.

The maximum salary for this position (36 - 36 hours) is € 3.196,00 gross per month based on full-time employment.

In addition, we offer an annual benefit of 8.3%, holiday allowance, travel expenses and career opportunities. The terms of employment are in accordance with the Cao University Medical Centers (UMC).

More information

If you have any questions about this vacancy, please contact Hugo Kuijf, Assistant Professor, phone number: 088 75 577 72, e‑mail adress:

Acquisition based on this jobopening is not appreciated.


No suitable vacancies?