Name

db_decode - Decode DNN activation from brain response.

The db_decode decodes the response of a voxel in the human brain to the representation from each layer in DNN.

Synopsis

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

Arguments

Required Arguments

Argument	Discription
anal	choices=(uv, mv) uv: Do univariate mapping mv: Do multivariate mapping
act	Path of a .act.h5 file which contains activation information.
resp	Path of a .roi.h5/.nii file which contains brain response information. If it is .nii file, -roi will be ignored. If it is .roi.h5 file, -bmask will be ignored.
model	choices=(glm, lasso) Select a model to predict dnn activation by brain responses. glm = general linear model lasso = lasso regression
out	an output directory

Optional Arguments

Argument	Discription
layer	Name of the target layer(s). Default using all layers in .act.h5 file
chn	Index of target channel(s). Default using all channels of each layer specified by -layer.
dmask	a .dmask.csv file in which layers of interest are listed with their own channels, rows and columns of interest.
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 for decoding. Default is using all ROIs in .roi.h5 file.
cv	cross validation fold number Default is 3.
scoring	model evaluation rules: correlation or sklearn scoring parameters Default is explained_variance.

Outputs

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

Examples

Decode brain response(test.nii.gz) to DNN activation(test.act.h5) using GLM model with 10 cross validation fold numbers in multivariate analysis.

db_decode -anal mv -act test.act.h5 -resp test.nii.gz -model glm -cv 10 -out test_glm_cv-10