peer package¶
Submodules¶
peer.peer_func module¶
Functions to predict eye movements
- Authors:
- Jake Son, 2017-2018 (jake.son@childmind.org) http://jakeson.me
-
peer.peer_func.
estimate_em
(_x_fix, _y_fix, _fix_xname, _fix_yname, _output_dir)[source]¶ Parameters: - _x_fix (float) – List of predicted fixations in the x-direction
- _y_fix (float) – List of predicted fixations in the y-direction
- _fix_xname (string) – Filename of the CSV containing fixation predictions in the x-direction
- _fix_yname (string) – Filename of the CSV containing fixation predictions in the y-direction
- _output_dir (string) – Pathname of the output directory
-
peer.peer_func.
global_signal_regression
(_data, _eye_mask_path)[source]¶ Performs global signal regression
Parameters: - _data (float) – Data from an fMRI scan as a 4D numpy array
- _eye_mask_path – Pathname for the eye mask NIfTI file (the standard MNI152 2mm FSL template is used for the linked preprint)
Returns: 4D numpy array containing fMRI data after global signal regression
Return type: _data
-
peer.peer_func.
load_config
()[source]¶ Loads configuration parameters as a dictionary from config.json
Returns: Dictionary containing configuration options from the config file (config.json) Return type: _configs
-
peer.peer_func.
load_data
(_filepath)[source]¶ Loads fMRI data
Parameters: _filepath (string) – Pathname of the NIfTI file used to train a model or predict eye movements Returns: _data – 4D numpy array containing fMRI data Return type: float
-
peer.peer_func.
load_model
(_output_dir)[source]¶ Loads the SVR models used to estimate eye movements
Parameters: _output_dir (string) – Pathname of the output directory Returns: - _xmodel – SVR model to estimate eye movements in the x-direction
- _ymodel – SVR model to estimate eye movements in the y-direction
- _xname (string) – Filename of the model used to estimate eye movements in the x-direction
- _yname (stringf) – Filename of the model used to estimate eye movements in the y-direction
-
peer.peer_func.
motion_scrub
(_ms_filename, _data_dir, _motion_threshold)[source]¶ Determines volumes with high motion artifact
Parameters: - _ms_filename (string) – Pathname of the CSV file containing the framewise displacement per time point for a given fMRI scan
- _data_dir (string) – Pathname of the directory containing data
- _motion_threshold (float) – Threshold for high motion (framewise displacement, defined by Power et al. 2012)
Returns: _removed_indices – List of volumes to remove for motion scrubbing
Return type: int
-
peer.peer_func.
predict_fixations
(_xmodel, _ymodel, _data)[source]¶ Predict fixations
Parameters: - _xmodel – SVR model to estimate eye movements in the x-direction
- _ymodel – SVR model to estimate eye movements in the y-direction
- _data – 4D numpy array containing fMRI data used to predict eye movements (e.g., movie data)
Returns: - _x_fix (float) – List of predicted fixations in the x-direction
- _y_fix (float) – List of predicted fixations in the y-direction
-
peer.peer_func.
prepare_data_for_svr
(_data, _removed_time_points)[source]¶ Preprocess fMRI data prior to SVR model generation
Parameters: - _data (float) – 4D numpy array containing fMRI data after global signal regression
- _removed_time_points (int) – List of volumes to remove for motion scrubbing
Returns: - _processed_data (float) – List of numpy arrays, where each array contains the averaged intensity values for each calibration point
- _calibration_points_removed (int) – List of calibration points removed if all volumes for a given calibration point were high motion
-
peer.peer_func.
save_fixations
(_x_fix, _y_fix, _xname, _yname, _output_dir)[source]¶ Save predicted fixations
Parameters: - _x_fix (float) – List of predicted fixations in the x-direction
- _y_fix (float) – List of predicted fixations in the y-direction
- _xname (string) – Filename of the model used to estimate eye movements in the x-direction
- _yname (string) – Filename of the model used to estimate eye movements in the y-direction
- _output_dir (string) – Pathname of the output directory
Returns: - _fix_xname (string) – Filename of the CSV containing fixation predictions in the x-direction
- _fix_yname (string) – Filename of the CSV containing fixation predictions in the y-direction
-
peer.peer_func.
save_model
(_xmodel, _ymodel, _train_file, _ms, _gsr, _output_dir)[source]¶ Saves the SVR models used in the PEER method
Parameters: - _xmodel – SVR model to estimate eye movements in the x-direction
- _ymodel – SVR model to estimate eye movements in the y-direction
- _train_file (string) – Pathname of the NIfTI file used to train the SVR model
- _ms (bool) – Whether or not to use motion scrubbing
- _gsr (bool) – Whether or not to use global signal regression
- _output_dir – Pathname of the output directory
-
peer.peer_func.
scaffolding
()[source]¶ Creates the project folder and file hierarchy and returns pathnames
Returns: - _project_dir (string) – Pathname of the highest-level project directory
- _data_dir (string) – Pathname of the directory containing data
- _output_dir (string) – Pathname of the output directory
- _stimulus_path (string) – Pathname of the PEER calibration scan stimuli
-
peer.peer_func.
set_parameters
(_configs, new=False)[source]¶ Sets configuration parameters
Parameters: - _configs – Dictionary containing configuration options from the config file (config.json)
- new (bool) – Do you want to start from a new file?
Returns: Updated dictionary containing configuration options from the config file (config.json)
Return type: _configs
-
peer.peer_func.
train_model
(_data, _calibration_points_removed, _stimulus_path)[source]¶ Trains the SVR model used in the PEER method
Parameters: - _data (float) – List of numpy arrays, where each array contains the averaged intensity values for each calibration point
- _calibration_points_removed (int) – List of calibration points removed if all volumes for a given calibration point were high motion
- _stimulus_path (string) – Pathname of the PEER calibration scan stimuli
Returns: - _xmodel – SVR model to estimate eye movements in the x-direction
- _ymodel – SVR model to estimate eye movements in the y-direction
peer.create_peer module¶
peer.estimate_eyemove module¶
peer.reset_config module¶
Script used on the command line to reset configuration file
- Authors:
- Jake Son, 2017-2018 (jake.son@childmind.org) http://jakeson.me