These depictions are produced from the approach described in my paper [PubMed], Locally Linear Embedding and fMRI feature selection in psychiatric classification.
The data produced and used for these experiments is available on the ABIDE Preprocessed Data Repository*

Visualisations for other published figures:


We explicitly state that the new figures are not in our original article, but are supported by similar metrics (save PPMI, see note on its page) that will be revealed upon publication.
statistically-significantly better than the chance threshold, defined as the majority class proportion for a given dataset.
§ statistically-significantly better than the reduction from Principal Component Analysis (PCA).
ǁ statistically-significantly better than the original, preprocessed, fMRI data.

* We want to express our deepest gratitude to Dr. Michael Milham for allocating resources and staff to host our experimental data.
Special thanks to Don McCurdy, and Dr. Duncan Murdoch (for his rGL and upcoming rgl2GLTF packages) who made these easy-to-consume depictions possible
P.S. you are probably (still) best-served if you view these pages on a real computer, as the renderings may not load on certain mobile device/browser configurations!

You are currently consuming the content in the latest webGL format. To view the legacy site, please click here

About Me