Figure S3.

Sensor Maps, Spatial Correlation, and Classifiers Performance of Trained Lasso Logistic Regression Models, Related to Figures 1 and 4

a, Sensor map for each state decoding model in Study 1 is shown on the left, with correlation matrix between classifiers shown on the right. b, Sensor maps and correlation matrix is shown for Study 2. c, In study 1, leave-one-out cross-validation results for each classifier in functional localizer task is shown on the left. Dotted line indicates the permutation threshold estimated by randomly shuffling the labels and re-doing the decoding process. Classifier performance during applied learning is shown on the right. These plots only use classifiers trained at 200 ms post stimulus onset. The x axis refers to the time point used for testing the classifiers. The curves therefore have a different shape than plots made by varying both the training and testing time. d, Study 2 had a very similar pattern.