regarding time-frequency plots and spectrogram plots

[kshtjkumar]

Hi, I was wondering if it is possible to make spectrograms and time-frequency plots using spikeinterface.

I am currently able to make the time-series plots using this command :
w = sw.plot_timeseries({"raw": recording_rhs, 'filt': recording_f, "common_referenced": recording_cmr}, clim = (-50, 50), time_range=[10, 10.1], order_channel_by_depth= True, backend='ipywidgets')

Any suggestions ?

[alejoe91]

Hi @kshtjkumar

Currently, we have no built-in methods to visualize spectra or spectrograms.

What you can do is to get traces of the different recordings and use matplotlob functions for plotting! Of course, this will need to be done separately for each channel. It should be relatively straightforward to do it, but let us know if you need additional help!

[kshtjkumar]

Yes @alejoe91 can you please help with this? How do I get the traces for individual of each channel ? Like an array that for each channel that can be used to plot the spectrogram ?

import spikeinterface as si
file = '1st_DEC_LKO _MICE_Merged.rhs'
reader = si.extractors.neoextractors.intan.IntanRecordingExtractor(file,stream_name ='RHD2000 amplifier channel')
recording_rhs.annotate(is_filtered = False)

How do i access the values for every channel from this ?

[kshtjkumar]

hi @alejoe91 , could you guide further with this please?

[alejoe91]

Sure!
you should do something like this:

import matplotlib

# we use 5 seconds between 30 and 35 (you can change this)
start_time = 30
end_time = 35
fs = recording_rhs.sampling_frequency

traces = recording_rhs.get_traces(start_frame=int(start_time * fs), end_frame=int(end_time * fs))

# plot spectrum for channel index 3
channel_index = 3
traces_channel = traces[:, channel_index]

_ = plt.psd(trace_channel, Fs=fs) # more info here: https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.psd.html
plt.xlabel("feqeuency (Hz)")

# plot spectrogram
_ = plt.specgram(trace_channel, Fs=fs) # more info here: https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.specgram.html#matplotlib.pyplot.specgram

[kshtjkumar]

@alejoe91 Thank you so much, you have been of great help.