What is mtrk?
mtrk is a novel concept for developing and working with pulse sequences in Magnetic Resonance Imaging (MRI). Traditionally, MRI pulse sequences are developed as dynamically loadable binary libraries, programmed using vendor-provided proprietary SDKs. These monolithic SDKs are complex, barely documented, and only in part accessible to external developers and members of the MRI research community. This makes MRI development very challenging - on the one hand because only few researchers manage to work with the SDKs, on the other hand because frequent invasive changes to the vendor SDKs create significant maintenance effort. Furthermore, strict license agreements make dissemination of research developments impossible and contend with the idea of reproducible research.
mtrk is an attempt to change the paradigm in MRI sequence development towards a modern Open-Source-driven model.
Flexibility
mtrk uses a highly modular design in which sequences are formulated using a Sequence Description Language (SDL) based on the common JSON syntax. The SDL is un-opinionated, meaning that multiple ways for formulating one sequence exist. Moreover, the SDL makes no assumption on how the SDL files have to be generated, giving researchers the freedom of choice regarding which programming language they want to use for their sequence development projects. Moreover, because the SDL is a common text-based format, the sequence calculation can be done on any operating system and on any computer, ranging from the MRI scanner’s console computer over local research servers to cloud instances.
Reproducibility
When executing an mtrk sequence, the MRI scanner plays out the instructions and pulses contained in the mtrk file. This means that the pulse sequence is entirely defined by the mtrk file and not dependent on the specific MRI software version used during data acquisition. Hence, researchers can reproduce scientific results at a later time or at a different location because the mtrk file can be easily archived or distributed. Because the SDL format is decoupled from the vendor SDKs, the mtrk files can be shared in Git repositories attached to publications without violating license agreements, and they do not require refactoring for every software update installed on the MRI scanners.
Transparency
With the traditional development model, sequence binaries are black-boxes in the sense that it is unclear how the sequence settings translate into DSP instructions. Moreover, because part of the functions are hard-coded into the MRI software platform and can change with every release, it is impossible to say which exact waveforms and timings were used to acquire the data. With mtrk on the other hand, all instructions can transparently be read form the mtrk file. This enables exact Bloch simulations of the sequence, so that automated testing and quality control of sequences is possible. Furthermore, part of the development work can be shifted from sitting in front of an MRI scanner towards simulation-driven development.
To learn more, please take a look at the mtrk Documentation. Please note that mtrk is still under active development, so that the code and documentation may not be complete yet and evolve over time. Please also don’t hesitate to get in touch with us if you like to contribute to the development.