Researchers recently revamped an open-source neuroimaging database in the hopes of expanding algorithm development in the field of stroke care. 

With accurate lesion segmentation in mind, the USC-led expansion built upon the Anatomical Tracings of Lesion After Stroke (ATLAS), version 1.2. ATLAS 1.2 contained hundreds of stroke T1w MRIs and manually segmented lesion masks, and though the intentions were good the results were less so. Many algorithms developed using that dataset did not yield acceptable accuracy or were improperly invalidated, limiting the usefulness of the data in stroke research. 

To compensate for this, researchers expanded upon the original data, curating a larger dataset of T1w MRIs and manually segmented lesion masks—ATLAS 2.0. 

Citing the time-consuming nature of manual lesion segmentation, lead author of the study Sook-Lei Liew, PhD, of the Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine at the University of Southern California, Los Angeles, and colleagues elaborated on how the data expansion can improve stroke research: 

“Algorithm development using this larger sample should lead to more robust solutions; the hidden datasets allow for unbiased performance evaluation via segmentation challenges. We anticipate that ATLAS v2.0 will lead to improved algorithms, facilitating large-scale stroke research.”