In a collaborative project led by Lars Kasper from the Translational Neuromodeling Unit (TNU) and the Institute for Biomedical Engineering at the University of Zurich and ETH Zurich together with National Imaging Facility (NIF) Fellow Steffen Bollmann from the Centre for Advanced Imaging (CAI), The University of Queensland node, a new open source toolbox for improving fMRI data analysis has been developed.
The physIO simplifies the workflow of incorporating physiological noise correction in fMRI analyses by providing robust pre-processing, fully automated modelling and a user-friendly performance assessment for group studies. All of this is tightly integrated in the open source toolbox SPM, the most widely used software for analysing fMRI data.
The project was started by Lars Kasper to access physiological data, i.e. breathing and heartrate from a Philips MRI scanner. Soon, he extended the toolbox to perform analysis of the physiological data and implemented various noise modelling methods. Steffen Bollmann joined the project in 2013 when he wanted to apply the algorithms to data collected during his PhD on a GE MRI scanner. The toolbox was not compatible with the GE data format and Steffen contributed routines for reading in data from this vendor and developed an algorithm for detecting ECG peaks more robustly. This was required due to poor data quality at times, as Steffen worked with children with Attention Deficit Hyperactivity Disorder (ADHD), who had a hard time lying still in the MRI scanner.
“New features for the toolbox were often developed on weekends when we met and discussed problems and possible solutions. The toolbox and our understanding of the problems we had solved grew during this time and we had more and more people testing it on their data and reporting problems they faced. Now, more than five years after the first lines of code, we have read-in routines for all major MRI vendors and a large repertoire of noise correction methods has been tested and implemented”, Steffen said.
Encouraged by its successful application in several studies including hundreds of volunteers, the team hopes that this open source toolbox will find useful application in future neuroimaging studies of health and disease, particularly in areas strongly affected by physiological noise such as the brainstem.
Translational Neuromodeling Unit (TNU), Institute for Biomedical Engineering, University of Zurich and ETH Zurich, Switzerland
Institute for Biomedical Engineering, ETH Zurich and University of Zurich, Switzerland
Centre for Advanced Imaging, University of Queensland, Brisbane, Australia
Wellcome Trust Centre for Neuroimaging, University College London, London, UK
Max Planck University College London Centre for Computational Psychiatry and Ageing Research, London, UK
Department of Psychiatry and Psychotherapy, Campus Mitte, Charité Universitätsmedizin, Berlin, Germany
Department of Neurology, Schulthess Clinic, Zurich, Switzerland
Max Planck Institute for Metabolism Research, Cologne, Germany