Loading NWB files using Neo¶
This notebook demonstrates how to load electrophysiology data using Neo for eFEL eFeatures extraction.
import efel
import numpy
%matplotlib notebook
%matplotlib inline
from matplotlib import pyplot as plt
plt.rcParams['figure.figsize'] = 10, 10
We will use the voltage trace recordings obtained from a mouse thalamic cell classified as a bursting accommodating (bAC) etype. The data is stored in an NWB (Neurodata Without Borders) file, and we will use the Python Neo library to load it.
test_data = "../../tests/testdata/JY180308_A_1.nwb"
stim_start = 250
stim_end = 600
blocks = efel.io.load_neo_file(test_data, stim_start, stim_end)
traces = blocks[0][0]
Let’s get a trace that includes a burst
trace = traces[5]
time = trace['T']
voltage = trace['V']
plt.rcParams['figure.figsize'] = 10, 10
fig1, ax1 = plt.subplots(1)
ax1.plot(time, voltage)
ax1.set_xlabel('Time (ms)')
ax1.set_ylabel('Membrane voltage (mV)');
We can now use the eFEL to extract eFeature values from the trace shown above.
We will use the get_feature_values() function, which accepts a list
of trace and the requested eFeature names as input.
Let’s extract some burst-related features
efel.set_setting("ignore_first_ISI", False) # Don't ignore the first spike
efel.set_setting("strict_burst_factor", 4.0) # The burst detection can be fine-tuned by changing the setting strict_burst_factor. Default value is 2.0.
feature_values = efel.get_feature_values([trace], ['spikes_per_burst', 'strict_burst_number', 'strict_burst_mean_freq', 'peak_time', 'AP_height', 'peak_indices', 'burst_begin_indices', 'burst_end_indices'])[0]
feature_values
{'spikes_per_burst': array([4]),
'strict_burst_number': array([1]),
'strict_burst_mean_freq': array([28.96451846]),
'peak_time': array([330.4, 343.2, 382.7, 468.5]),
'AP_height': array([11.40000057, 11.71249962, 12.23750019, 12.19999981]),
'peak_indices': array([3304, 3432, 3827, 4685]),
'burst_begin_indices': array([0]),
'burst_end_indices': array([3])}
burst_begin_indices = feature_values['burst_begin_indices'][0]
burst_end_indices = feature_values['burst_end_indices'][0]
peak_times = feature_values['peak_time']
ap_heights = feature_values['AP_height']
burst_mean_freq = feature_values['strict_burst_mean_freq']
time_spike_indices = numpy.where((time > stim_start) & (time < stim_end))
time_spike = time[time_spike_indices]
voltage_spike = voltage[time_spike_indices]
plt.figure(figsize=(10, 6))
plt.plot(time_spike, voltage_spike, label='Voltage Trace')
burst_start = peak_times[burst_begin_indices]
burst_end = peak_times[burst_end_indices]
mean_frequency = burst_mean_freq[0]
plt.axvspan(burst_start, burst_end, color='yellow', alpha=0.3, label='Burst Interval')
for spike_time in peak_times[burst_begin_indices:burst_end_indices+1]:
plt.axvline(x=spike_time, color='red', linestyle='--', label='Spike' if 'Spike' not in plt.gca().get_legend_handles_labels()[1] else "")
plt.text((burst_start + burst_end) / 2, max(voltage_spike), f'Burst Mean Freq: {mean_frequency:.2f} Hz', horizontalalignment='center', color='black')
plt.legend()
plt.xlabel('Time (ms)')
plt.ylabel('Voltage (mV)')
plt.show()
We can save the feature values obtained from get_feature_values for
later use. Two functions, save_feature_values_to_json and
save_feature_values_to_csv, are provided for this purpose. They save
the feature values as a JSON file and a CSV file, respectively.
efel.io.save_feature_to_json(feature_values, 'output.json')
efel.io.save_feature_to_csv(feature_values, 'output.csv')