Articles

By Pablo Velasco, Senior Scientific Solutions Engineer, Flywheel

One of the best ways you can accelerate neurology research is by running algorithms that automate some of the most time-intensive imaging processing tasks and workflows. The Flywheel medical imaging platform does this with built-in containerized algorithms called Gears that are available to all Flywheel users. These Gears have either been custom-built by Flywheel’s engineers or they’re trusted open-source algorithms we’ve added to the platform to automate common imaging processing and workflows.  

Below, we explore five of the most-used gears by neurology researchers and discuss why they're so invaluable for advancing neuroimaging research. 

1. dcm2niix: DICOM to NIfTI conversion 

The DICOM to NIfTI conversion gear (dcm2niix) is a cornerstone tool for neuroimaging research. It converts DICOM files, ubiquitous in medical imaging, into NIfTI format, making them more accessible for analysis. While DICOM is the standard format for clinical imaging, it’s a more complicated format that isn’t compatible with many neuroimaging analytic tools. By facilitating conversion to the versatile NIFTI format, this gear enables researchers to employ a broader range of applications for in-depth analysis. 

Why Neuroimaging Professionals Should Care: 

DICOM to NIfTI conversion simplifies data preparation—and does so at scale. This allows researchers to focus on exploration and understanding rather than file formatting. Its automated conversion process within Flywheel enables quicker transitions between scanning and analysis, boosting productivity. 

2. File Metadata Importer and File Classifier 

These two gears extract metadata from incoming files and classify them according to various criteria, such as imaging modalities and contrast. The creation of gear rules automates this process, ensuring that each new file is systematically processed and organized upon arrival. These gears not only support DICOM and DICOM Zip archives but also PTD (Siemens PT format), NIfTI, ParaVision (Bruker format) and PAR/REC (Philips format). 

Why Neuroimaging Professionals Should Care:  

Metadata extraction and classification optimize data organization, making it easier for researchers to locate, access and utilize specific datasets for targeted analyses. This systematic approach reduces manual errors and enhances data integrity. You can also chain this gear with others to create a full file import workflow. 

3. BIDS Curation 

This gear uses the structure and filenames from incoming NIfTI files and converts them into a BIDS-compliant structure. This lets you then run other BIDS-supporting gears on that data. 

Why Neuroimaging Professionals Should Care: 

It’s important to organize data by the Brain Imaging Data Structure (BIDS) so you’re organizing and sharing data obtained in neuroimaging experiments in the same way as everyone else. Not using this standard can lead to time wasted rearranging data or rewriting scripts. This gear saves time in the conversion process, as the only manual intervention typically required happens at the project level, so you don’t have to convert data file by file. 

4. FreeSurfer 

Offered by the Laboratory for Computational Neuroimaging at the Athinoula A. Martinos Center for Biomedical Imaging, FreeSurfer delivers a full processing stream for anatomical MR imaging data, including skull-stripping, bias field correction, registration and anatomical segmentation. 

Why Neuroimaging Professionals Should Care: 

FreeSurfer is a computing-intensive gear, but it doesn't require BIDS conversion and can run complex analyses directly within the platform. Running it through Flywheel helps researchers scale their computing power as needed through the cloud. Using this gear helps researchers dig deeper into their imaging data and differentiate white matter, gray matter and ventricles, providing researchers with detailed insights into brain structures. 

5. BIDS fMRIPrep & BIDS QSIPrep 

Both of these gears contribute to preparing neuroimaging data for comprehensive analysis by addressing distortions and aligning images with anatomical references. BIDS fMRIPrep enhances functional magnetic resonance imaging (fMRI), while BIDS QSIPrep configures pipelines for processing diffusion-weighted MRI (dMRI) data, ensuring clearer, noise-reduced images.