Bundle profiles and regression coefficients for the WH dataset MD.
figureposted on 28.06.2021, 17:35 authored by Adam Richie-Halford, Jason D. Yeatman, Noah Simon, Ariel Rokem
This and S3 Fig demonstrate that SGL behaves much more like the lasso than the group lasso, as discussed in the main text. The parsimony pitfall is most evident in the IFOL and IFOR bundles.
Read the peer-reviewed publication
brain functionStatistical inferenceaccount tissue propertiesmultivariate relationshipsphenotypic predictionsbrain regionsMultidimensional analysistissue propertiesgroup lassocorticospinal tractregression modelsmatter changefiber bundlebundle-level sparsityALSdiffusion propertiesSGL leveragesresonance imaging