Identification of Neuronal Activity from Extracelullar Recordings with Ground-truth Patch Clamp

Automated algorithm for the identification of neuronal activity from multichannel electrode arrays available on an online database¹

Decomposition of an initial segment of the recording:
main code: NeuroNella_DecomposeGroundTruth

How to run (in MATLAB):

1: Download the main code (NeuroNella_DecomposeGroundTruth) and zip folder containing the subcodes
2: Download recording files: Download one or more files from the database and save in one folder
3: Set Up the Main Directory: Locate the MainDir variable in the code and edit it to include the path to the folder containing the code files.

4: Set Up the Recording Directory: Locate the RecordingDir variable in the code and edit it to include the path to the folder containing the downloaded recordings.
5: Set Up Decomposition parameters
6: Run the code
7: If needed, run section "IMPROVE DECOMPOSITION" and adjust parameters

Extend Decomposition to the entire length of the recording:
main code: NeuroNella_ExtendGroundTruth

How to run (in MATLAB):

1: Decompose recording: first, decompose recording using the code 'NeuroNella_DecomposeGroundTruth'
2: Set up the Main Directory: Locate the 'MainDir' variable in the 'NeuroNella_ExtendGroundTruth' code and edit it to include the full path to the folder containing the code files.
3: Ensure the decomposed recordings are saved in the same directory as specified in the 'MainDir' path
4: Select Recording: find the 'f' variable and edit it to the index of the decomposed recording you wish to extend
5: Set up the EXTEND DECOMPOSITION PARAMETERS
6: Run the code
7: Optional: Improve Decomposition: If needed, run section "IMPROVE DECOMPOSITION" within the code and adjust parameters as necessary to refine the classification of the discharge instances

¹ Dataset Authors: Giulia LB Spampinato, Elric Esposito, Pierre Yger, Jens Duebel, Serge Picaud, Olivier Marre