huthlab / semantic-decoding Goto Github PK

Thanks for your sharing. But I have some problems in running train_EM.py, the sess_to_story.json is missing.
I found the sess_to_story.json in another repository here, but it seems doesn't fit with the default sessions in train_EM.py.

Thank you for your exciting research results.I have a question.

Is my understanding correct that you used "derivative/preprocessed_data/*.hf5" as resp data? I wonder how to make these .hf5 files. A website（https://openneuro.org/datasets/ds004510/versions/1.1.0）stated that "preprocessed data: fully preprocessed data as described in previous works." but I don't know where to refer.

I would appreciate it if you could give me some guidance.

Wanted to follow-up on the previous issue. Could the ROI jsons be released? We would like to be able to train the word-rate models from scratch for our own individual analysis. I can't identify any similar file in the OpenNeuro release as well.
Thank you!

Thanks fot your fascinating and solid works. I wonder if you can provide your corpus comprising Reddit comments and other stories. Because I want to verify if changing to another LLM model affects the decoding performance.
It's OK if the dataset is not available due to competing interests or patent pretection. I would appreciate it if you could help me with this. Thanks a lot.

Hello!

When I open the h5 files, the data of voxels have the shape(N, 81126).
But through the pycortex lib, I find that **.nii.gz have the shape(54, 84, 84).
The different shapes between them make me confused.

Hello, thank you for your share. Could you upload the data_train/ROIs file in which codes need [SUBJECT_ID].json?. It showed abscence of it when I estimated the word rate model following README.md. And it seems that there aren't any files like it in OpenNeuro.

Why is the spatial dimension of the signal 81126 and how to get it?