Load Data

The following python script can be used to load and plot the data produced by the NeuropixelsV2e Headstage example workflow.

# Import necessary packages
import os
import numpy as np
import matplotlib.pyplot as plt
import spikeinterface.extractors as se
import probeinterface
import probeinterface.plotting

#%% Set parameters for loading data

suffix = 0                                                 # Change to match filenames' suffix
data_directory = 'C:/Users/open-ephys/Documents/data/np2e' # Change to match files' directory
plot_num_channels = 10                                     # Number of channels to plot
start_t = 3.0                                              # Plot start time (seconds)
dur = 2.0                                                  # Plot time duration (seconds)

# Neuropixels 2.0 constants
fs_hz = 30e3
gain_to_uV = 3.05176
offset_to_uV = -2048 * gain_to_uV
num_channels = 384

#%%  Load acquisition session data

dt = {'names': ('time', 'acq_clk_hz', 'block_read_sz', 'block_write_sz'),
      'formats': ('datetime64[us]', 'u4', 'u4', 'u4')}
meta = np.genfromtxt(os.path.join(data_directory, f'start-time_{suffix}.csv'), delimiter=',', dtype=dt, skip_header=1)
print(f'Recording was started at {meta["time"]} GMT')
print(f'Acquisition clock rate was {meta["acq_clk_hz"] / 1e6 } MHz')

#%% Load Neuropixels 2.0 Probe A data

np2_a = {}

# Load Neuropixels 2.0 probe A time data and convert clock cycles to seconds
np2_a['time'] = np.fromfile(os.path.join(data_directory, f'np2-a-clock_{suffix}.raw'), dtype=np.uint64).astype(np.double) / meta['acq_clk_hz']

# Load and scale Neuropixels 2.0 probe A ephys data
rec_a = se.read_binary(os.path.join(data_directory, f'np2-a-ephys_{suffix}.raw'),
                     sampling_frequency=fs_hz,
                     dtype=np.uint16,
                     num_channels=num_channels,
                     gain_to_uV=gain_to_uV,
                     offset_to_uV=offset_to_uV)
np2_a['ephys_uV'] = rec_a.get_traces(return_scaled=True, channel_ids=np.arange(plot_num_channels))

np2_a_time_mask = np.bitwise_and(np2_a['time'] >= start_t, np2_a['time'] < start_t + dur)

#%% Load Neuropixels 2.0 Probe B data

# np2_b = {}

# Load Neuropixels 2.0 probe B time data and convert clock cycles to seconds
# np2_b['time'] = np.fromfile(os.path.join(data_directory, f'np2-b-clock_{suffix}.raw'), dtype=np.uint64).astype(np.double) / meta['acq_clk_hz']

# # Load and scale Neuropixels 2.0 probe B ephys data
# rec_b = se.read_binary(os.path.join(data_directory, f'np2-b-ephys_{suffix}.raw'),
#                      sampling_frequency=fs_hz,
#                      dtype=np.uint16,
#                      num_channels=num_channels,
#                      gain_to_uV=gain_to_uV,
#                      offset_to_uV=offset_to_uV)
# np2_b['data_uV'] = rec_b.get_traces(return_scaled=True, channel_ids=np.arange(plot_num_channels))

# np2_b_time_mask = np.bitwise_and(np2_b['time'] >= start_t, np2_b['time'] < start_t + dur)

#%% Load BNO055 data

dt = {'names': ('clock', 'euler', 'quat', 'is_quat_id', 'accel', 'grav', 'temp'),
      'formats': ('u8', '(1,3)f8', '(1,4)f8', '?', '(1,3)f8', '(1,3)f8', 'f8')}
bno055 = np.genfromtxt(os.path.join(data_directory, f'bno055_{suffix}.csv'), delimiter=',', dtype=dt)

# Convert clock cycles to seconds
bno055_time = bno055['clock'] / meta['acq_clk_hz']

bno055_time_mask = np.bitwise_and(bno055_time >= start_t, bno055_time < start_t + dur)

#%% Plot time series

fig = plt.figure(figsize=(12, 8))

# Plot scaled Neuropixels 2.0 probe A data
plt.subplot(611)
plt.plot(np2_a['time'][np2_a_time_mask], np2_a['ephys_uV'][np2_a_time_mask])
plt.xlabel('Time (seconds)')
plt.ylabel('µV')
plt.title('Neuropixels 2.0  Probe A')

# Plot scaled Neuropixels 2.0 probe B data
plt.subplot(612)
# plt.plot(np2_b['time'][np2_b_time_mask], np2_b['data_uV'][np2_b_time_mask])
plt.xlabel('Time (seconds)')
plt.ylabel('µV')
plt.title('Neuropixels 2.0 Probe B')

# Plot BNO055 data
plt.subplot(613)
plt.plot(bno055_time[bno055_time_mask], bno055['euler'].squeeze()[bno055_time_mask])
plt.xlabel('Time (seconds)')
plt.ylabel('degrees')
plt.title('Euler Angles')
plt.legend(['Yaw', 'Pitch', 'Roll'])

plt.subplot(614)
plt.plot(bno055_time[bno055_time_mask], bno055['quat'].squeeze()[bno055_time_mask])
plt.xlabel('Time (seconds)')
plt.title('Quaternion')
plt.legend(['X', 'Y', 'Z', 'W'])

plt.subplot(615)
plt.plot(bno055_time[bno055_time_mask], bno055['accel'].squeeze()[bno055_time_mask])
plt.xlabel('Time (seconds)')
plt.ylabel('m/s\u00b2')
plt.title('Linear Acceleration')
plt.legend(['X', 'Y', 'Z'])

plt.subplot(616)
plt.plot(bno055_time[bno055_time_mask], bno055['grav'].squeeze()[bno055_time_mask])
plt.xlabel('Time (seconds)')
plt.ylabel('m/s\u00b2')
plt.title('Gravity Vector')
plt.legend(['X', 'Y', 'Z'])

fig.tight_layout()

#%% Load and plot Neuropixels 2.0 ProbeInterface probe group

np2_config_a = probeinterface.io.read_probeinterface(os.path.join(data_directory, 'np2-config-a.json'))
# np2_config_b = probeinterface.io.read_probeinterface(os.path.join(data_directory, 'np2-config-b.json'))

fig = plt.figure()

plt.subplot(121)
ax_a = plt.gca()
probeinterface.plotting.plot_probegroup(np2_config_a, show_channel_on_click=True, ax=ax_a)
plt.title('Neuropixels 2.0 Probe A')

plt.subplot(122)
ax_b = plt.gca()
# probeinterface.plotting.plot_probegroup(np2_config_b, show_channel_on_click=True, ax=ax_b)
plt.title('Neuropixels 2.0 Probe B')

plt.show()

Note

To plot probeinterface data, save the probe configuration file into the same directory of your data.
By default, this script only populates plots for the Neuropixels 2.0 probe A and leaves plots for Neuropixels 2.0 probe B empty. If you are recording from both probes and want to plot probe B ephys and probe data, uncomment the relevant lines in the script.
This script will attempt to load entire files into arrays. For long recordings, data will need to be split into more manageable chunks by:
- Modifying this script to partially load files
- Modifying the workflow to cyclically create new files after a certain duration

Table of Contents

Load Data

Note