GitHub - zhuang0409/RSA

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 15 Commits
images72		images72
models/labels		models/labels
outputs		outputs
A01_runRSA_glmRN50.m		A01_runRSA_glmRN50.m
A01_runRSA_glmRNactionvsobject.m		A01_runRSA_glmRNactionvsobject.m
A03_runRSA_DNNs.m		A03_runRSA_DNNs.m
A04_RSA_Smooth5.sh		A04_RSA_Smooth5.sh
A05_run_smooth.sh		A05_run_smooth.sh
A06_createStatMapRSA.m		A06_createStatMapRSA.m
A07_TFCE_group.m		A07_TFCE_group.m
A09_TFCE_groupGLMDNNs.m		A09_TFCE_groupGLMDNNs.m
MNI152_T1_2mm_brain_mask.nii.gz		MNI152_T1_2mm_brain_mask.nii.gz
ReadMe.rtf		ReadMe.rtf
Resnet50_actions.ipynb		Resnet50_actions.ipynb
Resnet50_objects.ipynb		Resnet50_objects.ipynb
VGG_reshape_to_fMRI.m		VGG_reshape_to_fMRI.m
Workflow Travel_RSA.docx		Workflow Travel_RSA.docx
run_all.m		run_all.m
utils.py		utils.py

Repository files navigation

Step:
Download npy-matlab

1. RSA
2. Smooth
3. Group statistics


Codes:
1. Python code: to create layer-based RDMs of ResNet 50-action, and ResNet 50-object.
2. Matlab code: run RSA analysis. 
	A03_runRSA_DNNs: standard RSA
	A01_runRSA_glmRN50: regressing out 1st layer
	A01_runRSA_glmRNactionvsobject: regressing out 1st layer and another factor.

Folders:
1. Images 72: stimuli
2. Models: 
	labels: names of stimuli
	weights: pre-trained model weights of ResNet 50-action. (The weights of ResNet50-object is the default setting in codes)\
3. output: 
	layers name
	RDMs results
4. Plots: results of plots.