Coherent activity at three major lateral hypothalamic neural outputs controls the onset of motivated behavior responses
This repository includes scripts and source data that were used in the following preprint: Martianova, E., Pageau, A., Pausik, N., Doucet, T., Leblanc, D, Proulx, C.D. Coherent activity at three major lateral hypothalamic neural outputs controls the onset of motivated behavior responses. bioRxiv, doi: 10.1101/2021.04.28.441785v1, 2021.
FiberPhotmetryDataAnalysis.ipynb notebook includes functions necessary to process, analyze, and visualize fiber photometry data. Using these functions, jupyter notebooks in the rawData folder show all the steps of analysis from raw data to the final source data and plots used in the figures of the preprint. The sourceData folder also includes jupyter notebooks showing stastical analysis done for the preprint.
Using functions from FiberPhotmetryDataAnalysis.ipynb, you can initialize FiberPhotometryRecording object containing your recordings:
# Create dictionaries with your recordings, events, measurements
signals = {'Aneurons': signal_from_A, # The keys are names of neuronal populations, values are vectors of intensity.
'Bneurons': signal_from_B} # You can add as many recordings as you have from the same animal.
references = {'Aneurons': reference_from_A, # The keys have to match the ones in the signals.
'Bneurons': reference_from_B}
time_ = yourTimeVector # All your recordings and time vectors have to be the same length.
events = {'event1': event1array, # You can add as many events as you have. Each event is a 2D array of timestamps,
'event2': event2array, # with 2 columns if it has onset and offset (e.g. consumption, immobility bouts),
'event3': event3array} # with 1 column if it has only onset (e.g. air puff, foot shock).
measurements = {'measure1': 'time': measure1time, # The measurements can be for example speed or freezing score.
'values': measure1values} # The recording time of the measurements can be not the same as fiber photometry
'measure2': 'time': measure1time, # You can add as many measures as you have.
'values': measure1values}
# Create FiberPhotometryRecording object
r = FiberPhotometryRecording(signals,references,time_,events,
measurements,'mouse1name','test1name','trialNumber') After initializition, one can calculate dF/F signal:
# Calculate dF/F for all recordings
r.getDFF()
# Get a dF/F trace from one neuronal population
r.dFFs['Aneurons']Then one can calculate average signal around events (perievents):
# Calculate perievent arrays for each neural population and each event
r.getPerievents()
# Get perievent array recorded from neurons A at the offset of event 3
r.Perievents['Aneurons']['event3']['offset']
# Save data to hdf file
r.saveRecording('HDFname.h5') # In the same HDF file, you can save several recordings from different mice, tests, and trials
If you set parameters plot and save to True, the functions create a folder 'figures', which contains plots of the steps of the data processing.
The pipeline of processing of fiber photometry data is explained in our JoVE paper: Martianova, E., Aronson, S., Proulx, C.D. Multi-Fiber Photometry to Record Neural Activity in Freely Moving Animal.. J. Vis. Exp. (152), e60278, doi: 10.3791/60278 (2019). An example can be found in our other github repo.
Using FiberPhotometryTest object, you can calculate mean perievent traces per mouse.
# Create FiberPhotometryTest object
test = FiberPhotometryTest('HDFname.h5','testName')
# Calculate means and area under curve (AUC)
test.getMeans()
# Or you can precise the frames for AUC (the default frames are -2s - -1s and +1s - +2s from an event)
test.getMeans(auc_frames={'event1-onset': [[-2,-1],[0,2]],
'event2-onset': [[-2,-1],[-0.5,1],[2,4]],
'event2-offset': [[-2, 0],[0, 2]],
'event3-offset': [[-3,-1],[-1,0],[0,1],[1,3]]})
# Plot means for neurons A at onset of event 1
test.plotMeans('neuronsA','event1','onset')
# Get a data frame of AUCs for all neural populations at onset of event 1
df = getDataFrameAUC('event1','onset',['before','after']) # The last argument sets names to AUC time frames
# You can use this data frame to do statistical analysis
You can also do correlation analysis between dF/F signal and measurement or dF/F signals of two neural populations. It can be total correlation across the whole recordings or at specific events. To plot the results across several tests, FiberPhotometryExperiment object can be used.
# Total coorelation between dF/F and measure traces
# Create FiberPhotometryTest objects corresponding to one of the tests saved in HDF file.
test1 = FiberPhotometryTest('HDFname.h5','test1name')
test2 = FiberPhotometryTest('HDFname.h5','test2name')
# Calculate overall Pearson correlation between dF/F traces and measure1 traces.
test1.getMeasureCorrelation('measure1')
test2.getMeasureCorrelation('measure2')
# Create FiberPhotometryExperiment object
exp = FiberPhotometryExperiment('HDFname.h5')
# Plot distribution of R values for neurons A and measure 1 in tests 1 and 2
exp.plotRoverTests('Aneurons',measure='measure1',tests=['test1name','test2name']
# Get data frame of R values
df = exp.getDataFrameRmeasure('measure1',[test1name','test2name'])The same can be done for correlation between dF/F traces of two neural populations
# Total coorelation between dF/F traces of two neural populations
# Create FiberPhotometryTest objects corresponding to one of the tests saved in HDF file.
test1 = FiberPhotometryTest('HDFname.h5','test1name')
test2 = FiberPhotometryTest('HDFname.h5','test2name')
# Calculate overall Pearson correlation between dF/F traces of A and B neurons.
test1.getOutputCorrelation('Aneurons','Bneurons')
test2.getOutputCorrelation('aneurons','Bneurons')
# Create FiberPhotometryExperiment object
exp = FiberPhotometryExperiment('HDFname.h5')
# Plot distribution of R values for neurons A and measure 1 in tests 1 and 2
exp.plotRoverTests('Aneurons',output1='Bneurons',tests=['test1name','test2name'])
# Get data frame of R values for tests 1 and 2
df = exp.getDataFrameRoutputs([test1name','test2name'])
Correlation analysis can be done at recorded events and in between them. For more information check our preprint.
# Create FiberPhotometryTest objects corresponding to one of the tests saved in HDF file.
test1 = FiberPhotometryTest('HDFname.h5','test1name')
# Calculate perievent correlation between dF/F and measure 1 at event 1 and between them
test1.getMeasurePerieventCorrelation('event1','measure1')
# Plot a number of positively, negatively, not correlated events
test1.plotMeasureCorrelationCounts('measure1','Aneurons','event1')
# Create FiberPhotometryExperiment object
exp = FiberPhotometryExperiment('HDFname.h5')
# Get data frame of the number of positively, negatively, not correlated events
df = exp.getDataFrameMeasureCorrCounts('measure1',['test1name','test2name'],['Aneurons','Bneurons'])
From the results of the previous perievent correlation analysis, you can plot cross correaltion.
# Plot cross correlation between dF/F and measure 1 traces
test1.plotMeasureCrossCorrelation('measure1','Aneurons','event1','offset')
# Get dataframe of cross correlation at offset of event 1
df = test1.getDataFrameCrossCorr('measure1','event1','offset')
The same analysis can be done between dF/F traces of two neural populations.
This explanation covers all the functions that were used to analyze data for the preprint.
This project is licensed under GNU GPLv3.