ajcastany/spike_helper

Neuronal Spikes Sorting

Spikes are the action potential as measure outside of the brain cell. With code from: Zoe

Load recording.

The format of the recording is in `.dat`. The recordign file contains N channels x M samples in 16 bit intergers. Naming convention:

ch[n]_samp[m]

import spikeinterface.full as si
import spikeinterface.extractors as se
import spikeinterface.sorters as ss
import spikeinterface.comparison as sc
import spikeinterface.widgets as sw

file_path = './data/continous.dat'
sampling_frequency = float(25000)
num_chan = 128
data_type = int
#gain_to_uV = 0.0000001907
recording = si.BinaryRecordingExtractor(file_paths,
                            sampling_frequency,
                            num_chan,
                            dtype,
                            t_starts=None,
                            channel_ids=None,
                            time_axis=0,
                            file_offset=0,
                            gain_to_uV=0.0000001907,
                            offset_to_uV=None,
                            is_filtered=None)

print(recording) #check it works

Preprocessing steps

Can be processed instead on other software.

Hi-Pass filtering

recording_band_filter = si.bandpass_filter(recording,
                                           freq_min=300,
                                           freq_max=6000)

Common reference

recording_cmr = si.common_reference(recording_f,
                                    reference='global',
                                    operator='median')

Save to disk

recording_preprocessed = recoding_cmr.save(format='binary',
                                              n_jobs=1)

Recreate and load a Probe Map

List of x, y coordinates for recording sites for 128 channels.

# imports
from probeinterface import Probe, ProbeGroup
from probeinterface import generate_linear_probe, generate_multi_shank
from probeinterface import combine_probes
from probeinterface.plotting import plot_probe

probe = generate_multi_shank(num_shank=32,
                                   num_columns=2,
                                   num_contact_per_column=2)

plot_probe(probe, with_contact_id=True)

Attach Probe Map to recording object

probe.set_device_channel_indices(np.arange(128))
recording = recording.set_probe(probe)

Spike Sorting.

Set parameters

Run the sorting algorithms

Extract Waveforms

Export to Phy