Feature extraction GUI

The NeuroFeatureExtract tool allows users to select data from an HBP dataset and/or upload their own, and extract electrophysiological features of interest.

The application leverages the Python Electrophys Feature Extract Library (eFEL) and provides a friendly interface to select both individual voltage traces (based on the stimulus current applied) and features to be extracted.

Please find below a short tutorial on how to use the application, or refer to the complete guide at this link.

Select the data from a dataset, based on cell properties to be chosen from the filter dropdown menus. Additionally and/or alternatively upload your own data for processing (see Fig. 1). The only file extension allowed at present is .abf (axon binary file) and the uploaded files must contain both the voltage and the stimulus signals. This is a mandatory condition for several features to be computed and for the feature extraction process to be completed successfully. The analog voltage and stimulus signals are extracted through the neo python library, designed for representing different formats of intracellular and extracellular electrophysiological data (see neo for a complete reference).

efel_selection — Fig 1. - Traces selection

efel_traces — Fig 2. - Voltage traces selection

Step 2 of 4 [skip introduction]

Once the filtering/upload done, individual traces are plotted and can be checked for selection. Traces represent the voltage membrane responses of the selected cell to the stimulus displayed in the legend. Highlight individual traces by hovering on the corresponding stimulus and select them by clicking on the same legend or through the selection buttons (see Fig. 2). When all the traces of interest have been selected, go the next page for feature selection.

Step 3 of 4 [skip introduction]

The feature selection page allows you to select the features to extract from the electrophysiological chosen signals. Features are grouped by type: 1) Spike event features - 2) Spike shape features and 3) Voltage features, and are selected by clicking on the corresponding box. When hovering on feature names a brief description is provided (see Fig. 3). Once the selection is completed, click on the Next button to launch the extraction process.

efel_features — Fig 3. - Feature selection

Step 4 of 4 [skip introduction]

The feature extraction process computes the means and standard deviations of the selected features for individual cells and for individual stimuli. Once a stimulus is chosen for a single file, it will be taken into account for all the selected files corresponding to the same cell. Finally, a grand mean is computed among different cells.
A features.json and protocols.json files are generated for both individual cells and the entire ensemble. These files contain the above mentioned grand-means and the protocols adopted for the experimental recordings. They are intended to be used for the data-driven optimization step of the Hodgkin-Huxley Neuron Builder workflow (via the BluePyOpt software library), made available to the user at this link.
A link to a .zip file containing the entire set of results is provided in the results page (see Fig. 4).