Examples gallery

Spikes & 2-d projection & video

In this example, head-direction neurons are being recorded during wake. The video shows the animal’s movement in an open field. The manifold widget shows a projection of the neural activity onto a 2D space using Isomap.

See the example code

import os
import requests
import pynapple as nap
from pynaviz import scope
import pynaviz as viz

nwb_file = "A5044-240404A_wake.nwb"
avi_file = "A5044-240404A_wake.avi"

files = os.listdir(".")
if nwb_file not in files:
    url = "https://osf.io/um4nb/download"
    with requests.get(url, stream=True) as r:
        r.raise_for_status()
        with open(nwb_file, "wb") as f:
            for chunk in r.iter_content(chunk_size=8192):  # Stream in chunks
                f.write(chunk)
if avi_file not in files:
    url = "https://osf.io/gyu2h/download"
    with requests.get(url, stream=True) as r:
        r.raise_for_status()
        with open(avi_file, "wb") as f:
            for chunk in r.iter_content(chunk_size=8192):  # Stream in chunks
                f.write(chunk)

# Load data
data = nap.load_file("A5044-240404A_wake.nwb")
units = data["units"]
manifold = data["manifold"]
video = viz.VideoWidget("A5044-240404A_wake.avi")

# Open the GUI
scope({"units": units, "manifold": manifold, "video": video})

---

Local Field Potential (LFP)

In this example, a Local Field Potential (LFP) signal is being recorded during sleep in mice.

See the example code

import pynapple as nap
import numpy as np
from pynaviz import scope

path_to_dat_file = "lfp.dat"
fs = 20000 # Sampling frequency in Hz
num_channels = 16 # Number of channels in the recording
tsdframe = nap.misc.load_eeg(
    path_to_dat_file,
    n_channels=num_channels,
    frequency=fs
)
tsdframe.channel = np.arange(0, num_channels) # Add channel metadata
tsdframe.group = np.hstack([[0]*10, [1]*6]) # Add group metadata


# Open the GUI
scope({"lfp":tsdframe})

---

DLC pose estimation

This example shows pose estimation data obtained using DeepLabCut (DLC). The data are from the DeepLabCut GitHub repository.

See the example code

import pynapple as nap
import numpy as np
from pynaviz import scope
from pathlib import Path


# Path to the video
video_path = "docs/examples/m3v1mp4.mp4"

# Output of deeplabcut
df = pd.read_hdf("docs/examples/m3v1mp4DLC_Resnet50_openfieldOct30shuffle1_snapshot_best-70.h5")
df.columns = [f"{bodypart}_{coord}" for _, bodypart, coord in df.columns]
df = df[[c for c in df.columns if c.endswith(("_x", "_y"))]]
y_col = [c for c in df.columns if c.endswith("_y")]
df[y_col] = df[y_col]*-1 + 480 # Flipping y axis
skeleton = nap.TsdFrame(t=df.index.values/30, d=df.values, columns=df.columns)


# Open the GUI
scope({"skeleton": skeleton, "video": video_path})

---

Multiple videos

This example shows multiple videos being recorded simultaneously from different angles. The data are from the International Brain Lab (IBL).

See the example code

import pynapple as nap
import numpy as np
from pynaviz import scope
from pathlib import Path
from one.api import ONE

# Load IBL session
one = ONE()
eid = "ebce500b-c530-47de-8cb1-963c552703ea"

# Videos
ibl_path = Path(os.path.expanduser("~/Downloads/ONE/"))
videos = {}
for label in ["left", "body", "right"]:
    video_path = (
            ibl_path
            / f"openalyx.internationalbrainlab.org/churchlandlab_ucla/Subjects/MFD_09/2023-10-19/001/raw_video_data/_iblrig_{label}Camera.raw.mp4"
    )
    if not video_path.exists():
        one.load_dataset(eid, f"*{label}Camera.raw*", collection="raw_video_data")
    times = one.load_object(eid, f"{label}Camera", collection="alf", attribute=["times*"])["times"]
    # The videos seem to start at 5 seconds. Removing artificially 5 seconds for the demo
    times = times - 5
    videos[label] = viz.VideoWidget(video_path, t=times)


# Open the GUI
scope(videos)

---

Video & Event & Trial intervals

This example shows spikes being recorded during a trial-based task. The data are from the International Brain Lab (IBL).

See the example code

import pynapple as nap
import numpy as np
from pynaviz import scope
from pathlib import Path
from one.api import ONE

# Load IBL session
one = ONE()
eid = "ebce500b-c530-47de-8cb1-963c552703ea"

# Videos
ibl_path = Path(os.path.expanduser("~/Downloads/ONE/"))
vars = {}
for label in ["left", "body", "right"]:
    video_path = (
            ibl_path
            / f"openalyx.internationalbrainlab.org/churchlandlab_ucla/Subjects/MFD_09/2023-10-19/001/raw_video_data/_iblrig_{label}Camera.raw.mp4"
    )
    if not video_path.exists():
        one.load_dataset(eid, f"*{label}Camera.raw*", collection="raw_video_data")
    times = one.load_object(eid, f"{label}Camera", collection="alf", attribute=["times*"])["times"]
    # The videos seem to start at 5 seconds. Removing artificially 5 seconds for the demo
    times = times - 5
    vars[label] = viz.VideoWidget(video_path, t=times)

timings = one.load_object(eid, "trials", collection="alf")
licks = nap.Ts(one.load_object(eid, "licks", collection="alf")["times"])
trials = nap.IntervalSet(timings["intervals"])

vars["trials"] = trials
vars["licks"] = licks

# Open the GUI
scope(videos)