analysis for updates to Deutsch & Clemens et al 2019

paper: Shared Song Detector Neurons in Drosophila Male and Female Brains Drive Sex-Specific Behaviors

data collection : Yiqin Gao
analysis: Kyle Thieringer

order of operations

1. exportslp.py
   • exports .slp files to .h5 files
2. createfeatures.py
   • calculates kinematic features and saves them to .h5 files
3. createMetatable.ipynb
   • uses exported googledrive csv file and creates a metadata sheet to be used by the createAllData.m script
4. exptSync.ipynb
   • uses the vfaas (video frame at audio sample) to interpolate the audio sample at each video frame which is needed by createAllData.m
5. checkFtrs.ipynb
   • visualizes basic features as a sanity check to assess quality of data
6. createAllData.m
   • creates a mat file containing kinematic features, audio data, and sync data for each fly
   • too large for github upload
   • stored here google drive link
7. Analyze_Behavior_Jans.m
   • uses output of createAllData to bin behavior and audio data for later analysis
   • this code was sent to me by Dudi. I updated it slightly to work with the new data.
8. rankCorrelation.ipynb
   • make plots using the output of Analyze_Behavior_Jans.m

---

some notable distinctions in processing the data

• I smoothed data during the createfeatures.py step using a kalman filter.
  • in the original analysis, behavior features were median filtered
• female speed was zscored for each fly then binned rather than using the raw speed

update : I took out the smoothing and zscoring to match the original data. The ouputs are stored in output_2.mat and allData2.ma in the google drive folder.