Computational Algorithms for Semi-Automatic Image Processing for Targeted Epidural Electrical Stimulation

Semester / Master's Work – Computational Algorithms for Semi-Automatic Image Processing for Targeted Epidural Electrical Stimulation


Central nervous system disorders such as spinal cord injury (SCI) and stroke lead to distinct impairments in motor control and balance. Epidural electrical stimulation (EES) of lumbar and sacral segments of the spinal cord has been proven to restore voluntary and coordinated movements of the lower limbs in various animal models.

Translation of this technology to the treatment of human patients requires the development of dedicated, robust solutions for clinical use. Indeed, therapeutic success of EES entails the definition of dedicated spinal electrode arrays and stimulation strategies that can account for the large variability observed in clinical populations. Personalization of implantation procedures and stimulation protocols will be critical for being able to address the specific motor deficits in large and diversified patient cohorts.

The RESTORE team – a collaboration between the IT’IS Foundation, ZurichMedTech, EPFL, GTX Medical and the UMCU – is currently working on establishing a computational framework for the creation of personalized computational models for the development of targeted EES strategies to promote recovery of motor function after SCI. This framework has already been proven to create a 1-to-1 mapping between simulation and experiment and is currently being advanced to optimize stimulation paradigms. However, the extraction of anatomical features from magnetic resonance imaging (MRI) data is slow and requires a tremendous amount of human effort. We therefore aim to develop image processing algorithms for semi-automatic to automatic extraction of key features.

The aim of this master’s thesis project is the development and implementation of image processing algorithms that can be used to extract the trajectory of the spinal roots from MRI data sets. To this end, the student will be provided with MRI data sets and a computational framework able to extract the center-lines of the spinal roots. The student is required to develop an algorithm based on Kalman filtering to improve the labelling and selection of these spinal roots.

The successful candidate will have:

      • experience with programming in C++ and/or Python
      • experience with image-processing
      • experience and/or interest in machine learning
      • experience and/or interest in 3D computer-aided-design modelling

The workplace will be at the IT’IS Foundation in Zurich.

Please contact Andreas Rowald for more information and further details.


Andreas Rowald (EPFL)

Dr. Bryn Lloyd (IT'IS)

Type of Work:

Computational life sciences


Niels Kuster

Please send applications to Charlotte Roberts at