Here are the models and codes for the paper of The representation of observed actions at the subordinate, basic and superordinate level.
sub.mat: the model of the subordinate level,
basic.mat: the model of the basic level,
super.mat: the model of the superordinate level,
humanRating_mds: the model obtained from human-rating experiment.
places_last2nd.mat: the place RDM obtained from the last second layer of ResNet50 pretrained with places
ROI-based analysis:
- Fig4_ROI_analysis.m
Searchlight-based analysis:
- Step 1. RSA searchlight analysis - searchlight1_xx.m/searchlight2_xx.m/searchlight3_xx.m
- Step 2. Searchlight4_smoothdata.sh - smoothing data
- Step 3. Searchlight5_createStatMapRSA.m - to create maps for statistics
- Step 4. Searchlight6_TFCE_group.m - to make the statistic
- Step 5. Fig7_ConjunctionAnalysis.m - to compute the conjunction analysis
Additionally,
-
vif.m: VIF analysis;
-
Fig5_plot_rdm_mds.m: to visualize the regressed LOTC/V1 rdm and the corresponding mds;
-
Fig2_plot_rdm_mds_histogram_of_human_rating_model.m: to visualze the rdm, mds, and histogram of human-rating results.
-
RN_reshape_to_fMRI.m: to change the order of stimuli with the neural conditions.Due to the order of the stimuli of RN50-based RDM is different with the the neural RDM, we used this function to keep the order of stimuli consistently.
-
rotateXLabels.m: to rotate x axis labels.