Integrating across neuroimaging modalities boosts prediction accuracy of cognitive ability
This collection is shared privately
This collection contains the following data and results files generated for our paper "Integrating across neuroimaging modalities boosts prediction accuracy of cognitive ability" (https://doi.org/10.1101/2020.09.01.278747). The collection has the following files:
- multimodal_cognition_data: Data used to generate the results of the paper.
- training_set_idxs: Indices for the training sets in the Monte Carlo cross-validation procedure.
- test_set_idxs: Indices for the test sets in the Monte Carlo cross-validation procedure.
- predictions: Single channels and stacking predictions generated during the Monte-Carlo cross-validation.
- scores: Single channels and stacking R2 and MAE scores generated during the Monte-Carlo cross-validation.
- weight_phenotypes: For those cognitive domains in which stacking led to a significant improvement, the different weights from fitting the LASSO-PCR estimator to the entire data in each single channel.
COLLECTION SIZE
5.97 GB