Neurophysiology Analysis Pipeline
When you recorded your neurophysiology data in the SCNI lab, it was saved to the solid state drive of the local PC (e.g. SCNI-Red-Neuro) and optionally also the TDT RS4 streamer (if you were using Synapse or OpenEX software). To analyse the data, you must first transfer it to the network as follows:
-
Transfer .SEV files to Nifstorage. If you used TDT, you should copy the 'Tanks' (directories) for each day's recording session from the RS4. The RS4 should be mapped to the TDT PC in your rig as a network drive. If not, you will need to map it by selecting the 'Map network drive' option from the File tab when 'This PC' is selected. Manually enter the RS4's IP address (accessible from the touchscreen on the front panel of the RS4) in the format suggested by the popup window. Find the directories created for your session(s) and copy them (e.g. drag and drop) to your
Rawdatadirectory on Nifstorage. Do not delete the original folders from the RS4 until you have confirmed that the data transferred without file corruption. -
Transfer TDT files to Nifstorage. In addition to the raw neural data that is saved to .sev binary files on the RS4, TDT saves downsampled LFP, online-sorted spikes, analog and digital inputs (including event codes) to proprietary files on the local PC running Synapse. To convert these data into a Matlab-readable format that is of use, we need a Windows PC with Matlab installed. The NIF server Nifsort1 is the best option for this purpose. Connect to it on Mac OSX using Finder: Go > Connect to server > and then enter the following server address:
vnc://nifsort1.nimh.nih.gov
Once connected, open a Matlab session. Navigate to a local copy of the SCNI_Toolbar repository on Nifstorage (e.g. P:/murphya/SCNI_Toolbox/SCNI_Toolbar/SCNI_Subfunctions/TDT_Matlab/) and run SCNI_TDTRead_Batch.m. This function will ask you to select folders of TDT files to convert to .mat format and where to save them to. A progress bar will appear while the task is in progress and updates will be printed to the command line.
We use the spike-sorting algorithm WaveClus to sort data from chronically implanted microwire brush array electrodes. We have implemented a version of this code that can run on Biowulf - the NIH's high-performance computer cluster- hosted in a repository called SortSpikes. See the Wiki on that page for full instructions.
