NWB Workshops and Hackathons

Back to the projects list

Work on integrating lab ECoG, behavioral, and imaging data using PyNWB

Key Investigators

• Jessie R. Liu (Chang Lab, UCSF)

Project Description

The Chang Lab records from a TDT acquisition system and has an in-house preprocessing pipeline for converting the raw TDT files into raw and preprocessed HTK files. The goal is to update this pipeline to be completely in Python and to utilize the NWB format.

Objectives

• Have experience integrating the most common types of lab data (TDT/htk, behavioral logs, imaging) into .nwb files
• Be able to help lab members convert their data/be familiar with existing efforts to convert lab data to NWB format

Approach and Plan

• Code has been/is currently being developed to help piece together a complete pipeline (Matlab script for TDT tanks to NWB, the preprocessing pipeline in Python that can read NWB files, visualization GUI that can read in NWB files)
• Become familiar with each of these pieces
• Make sure the parts all work together
• Utilize an example TDT block to go through the entire process

Progress and Next Steps

• TDT recently released a Python package to read in data from their raw files
  • Converted Matlab script to Python to go from TDT tanks to NWB files with the raw data in an ElectricalSeries
  • Utilize NWB ECoG extension to add subject’s cortical surfaces information.
• Edit Python preprocessing pipeline to save the frequency decomposition of the raw data into a DecompositionSeries
• Used visualization GUI to read in NWB files with the raw data
• Verified that each generated file was able to be used by the next step
• Next steps:
  • Continue working with others to test/expand the visualization GUI
    • Display high gamma data
    • Visualize ERPs aligned to a stimulus onset
  • Work with lab members to introduce them to the NWB format and get them familiar with reading and writing to these files

Materials

Background and References

• The TDT package: https://pypi.org/project/tdt/