Nilearnhttp://mloss.orgUpdates and additions to NilearnenWed, 29 Apr 2015 16:16:25 -0000Nilearn 0.1.2<html><p>Neuroimaging datasets are constantly growing in resolution, sample size, and complexity. This entails an always bigger interest in data-driven statistical analysis methods. Nilearn is a new scientific computing package that has been designed to address these challenges in contemporary data analysis in imaging neuroscience. It facilitates data pre-processing (i.e., "feature engineering"), state-of-the-art statistical learning algorithms (i.e., learning patterns from data), and visualization of various types of neuroimaging results (i.e., experimental fMRI, VBM, and resting-state correlations). </p> <p>Novice and expert users can, for instance, readily compute brain parcellations and extract signals from those. Brain signals could then be fed into sparse inverse covariance estimation to compute "functional connectomes". More generally, Nilearn can leverage a diverse set of unsupervised and supervised data analysis scenarios by integration with a well-mainted and growing general-purpose machine-learning library (i.e., scikit-learn). </p> <p>The successful extraction of structured knowledge/insight from current and future large-scale neuroimaging datasets will be a critical prerequisite for our understanding of human brain architecture in healthy populations and psychiatric/neurological disease. It is in this aim that Nilearn is conceived and developed. Given hosting on the coding plattform Github, Nilearn further encourages inter-laboratory collaboration towards software quality and data-analysis standards for the scientific community. </p></html>Gael Varoquaux,Alexandre Abraham,Loic Esteve,Danilo Bzdok,Chris Filo Gorgolewski,Michael Eickenberg,Ben Cipollini,Virgile Fritsch,Philippe Gervais,Alexandre Gramfort,Bertrand ThirionWed, 29 Apr 2015 16:16:25 -0000 learningstatistical learningneuroimagingscikit learn