Name

db_encode - Use DNN activation to encode brain.

The db_encode uses the representation from each layer in DNN to predict the response of a voxel in the human brain by using voxel-wise encoding models (EM).

Synopsis

db_encode  [-h] -anal Analysis -act Activation
           [-layer Layer [Layer ...]] [-chn Channel [Channel ...]]
           [-dmask DnnMask] [-iteraxis Axis] -resp Response
           [-bmask BrainMask] [-roi RoiName [RoiName ...]] -model Model
           [-scoring Scoring] [-cv CrossValidationFoldNumber] -out Output

Arguments

Required Arguments

Argument	Discription
anal	Choices is (‘uv’, ‘mv’). ‘uv’ means doing univariate analysis; ‘mv’ means doing multivariate analysis.
act	Path of a .act.h5 file which contains activation information.
resp	Path of a .roi.h5/.nii file which contains brain response information. Note: if it is .nii file, -roi will be ignored. “All voxels’ activation will be regarded as the “ground truth of a regression task.”
model	Select a model to predict brain responses by dnn activation. Choices is (‘glm’, ‘lasso’). You can use glm (general linear model) for regression or use lasso (lasso regression) for regression.
out	An output directory.

Optional Arguments

Argument	Discription
layer	Name of the target layer(s).Default is all.
chn	Index of target channel(s).Default is all.Channel index starts from 1.
dmask	Path of a .dmask.csv file in which detailed information of neuron(s) of interest in DNN is specified. Argument layer/chn and dmask are mutually exclusive. Provide only one of them if needed.
iteraxis	Iterate along the specified axis. If -anal is uv: channel: Summarize the maximal prediction score for each channel. row_col: Summarize the maximal prediction score for each location (row_idx, col_idx). default: Summarize the maximal prediction score for the whole layer. If -anal is mva: channel: Do mva using all units in each channel. row_col: Do mva using all units in each location (row_idx, col_idx). default: Do mva using all units.
bmask	Brain mask is used to extract activation locally. Only used when the response file is .nii file. If not given, the whole brain response will be used.
roi	Specify ROI names as the ground truth. Default is using all ROIs in .roi.h5 file.
Scoring	Model evaluation rules: correlation or sklearn scoring parameters. Default is explained_variance.
cv	Cross validation fold number. Default is 3.

Outputs

Arrays containing the prediction score of each layer. Note：Different layers’ output is stored in different folders.

Examples

DNN activation(test.act.h5) was used to predict brain response(test.nii.gz) using GLM model in multivariate analysis. The example uses the scoring of correlation with 10 cross validation fold numbers.

db_encode -anal mv -act test.act.h5 -resp test.nii.gz -model glm -scoring correlation -cv 10
-out test_glm-corr_cv-10