Open In Colab

Installing requirements¶

The cell below installs every Python package needed to run this notebook, at fully pinned versions, using uv for fast resolution. In Colab the cell is collapsed by default — click the ▶ button to run it.

In [1]:
# install cell skipped during CI (deps preinstalled into system Python)

⚠️ Restart runtime after install

The install may upgrade packages already loaded in the kernel. Go to Runtime → Restart session, then Run all cells below (skip this install cell on re-run).

MICrONS NWB tutorial¶

This tutorial demonstrates how to access the MICrONS functional data using dandi.

The functional data from the MICrONS project contains calcium imaging recorded from multiple cortical visual areas and behavioral measurements while a mouse viewed natural movies and parametric stimuli.

Contents:¶

  • Streaming NWB files
  • Access data and metadata
  • Access the coregistration table from the CAVE database
  • View NWB files with NWBWidgets
  • Downloading NWB files

Streaming NWB files ¶

This section demonstrates how to access the files on DANDI without downloading them.

Based on the Streaming NWB files tutorial from PyNWB.

The DandiAPIClient can be used to get the S3 URL of the NWB file stored in the DANDI Archive.

In [2]:
from dandi.dandiapi import DandiAPIClient

dandiset_id = "000402"
file_path = "sub-17797/sub-17797_ses-9-scan-4_behavior+image+ophys.nwb" # file size ~67GB

# Get the location of the file on DANDI
with DandiAPIClient() as client:
    asset = client.get_dandiset(dandiset_id, 'draft').get_asset_by_path(file_path)
    s3_url = asset.get_content_url(follow_redirects=1, strip_query=True)

Create a virtual filesystem using fsspec which will take care of requesting data from the S3 bucket whenever data is read from the virtual file.

In [3]:
from fsspec.implementations.cached import CachingFileSystem
from fsspec import filesystem
from h5py import File
from pynwb import NWBHDF5IO


# Create a virtual filesystem based on the http protocol and use caching to save accessed data to RAM.
fs = filesystem("http")
file_system = fs.open(s3_url, "rb")
file = File(file_system, mode="r")
# Open the file with NWBHDF5IO
io = NWBHDF5IO(file=file, load_namespaces=True)

nwbfile = io.read()

Access data and metadata ¶

This section demonstrates how to access the data in the NWB files.

Subject¶

The nwbfile.subject field holds information about the experimental subject, such as age (in ISO 8601 Duration format), sex, and species in latin binomial nomenclature.

In [4]:
# Access the subject metadata
nwbfile.subject
Out[4]:

subject (Subject)

age: P75D/P81D
sex: M
species: Mus musculus
subject_id: 17797

Trials¶

Clip¶

This stimulus condition is composed of 10 second clips. The trials for this condition can be accessed as nwbfile.intervals["Clip"].

In [5]:
# View the description for this stimulus type
print(nwbfile.intervals["Clip"].description)
# Load the data into a dataframe object
clip_dataframe = nwbfile.intervals["Clip"].to_dataframe()

Composed of 10 second clips from cinematic releases, Sports-1M dataset, or custom rendered first person POV videos in 3D environment in Unreal Engine.
In [6]:
clip_dataframe
Out[6]:
start_time stop_time stimulus_type condition_hash movie_name short_movie_name duration
id
4 269.464581 279.431084 stimulus.Clip okYfMyn8qc2ClgvpS7WG 2009 Skijoring Event in Whitefish Montana (par... sports1m 10.0
5 279.531081 289.497598 stimulus.Clip NHlF8hOkZ6psFRIVhB6N ancienttomb_107-1-4-70-0_54.165.28.255 Rendered 10.0
6 289.597593 299.564101 stimulus.Clip AxNNo310kAl9uUSr+hXS medievalsewer_009-1-6-70-0_107.23.36.15 Rendered 10.0
7 299.664089 309.630596 stimulus.Clip mjtZRmGqaXCHGB59txoI Powaqqatsi: Life in Transformation (1988) Cinematic 10.0
8 309.730601 319.697093 stimulus.Clip W50xYLSTlUixv1tOfuSJ 2011 Boulder Cup Cyclocross Elite Women sports1m 10.0
... ... ... ... ... ... ... ...
459 5229.733599 5239.700111 stimulus.Clip BmdZ8rRGTB6ueDkz+Kr/ shintonight_095-1-13-70-0_54.89.40.102 Rendered 10.0
460 5239.800111 5249.766615 stimulus.Clip WslKzH0v/OEqVTLiw468 2011 June 14 Midweek MTB Corner Canyon XC Moun... sports1m 10.0
461 5249.866617 5259.833112 stimulus.Clip eNJ7fFpiRbVfX48l845P 2011 August 2 Midweek MTB Solitude XC Mountain... sports1m 10.0
462 5259.933109 5269.899625 stimulus.Clip RmzqOWaAbbEeML35q5Xp 2010 Nova Scotia water ramp camp day 2 sports1m 10.0
463 5269.999622 5279.966128 stimulus.Clip XC+naBDWY+CgiqQGFUTt Koyaanisqatsi: Life Out of Balance (1982) Cinematic 10.0

384 rows × 7 columns

Monet2¶

This stimulus condition is generated from smoothened Gaussian noise and a global orientation and direction component. The trials for this condition can be accessed as nwbfile.intervals["Monet2"].

In [7]:
# View the description for this stimulus type
print(nwbfile.intervals["Monet2"].description)
# Load the data into a dataframe object
monet2_dataframe = nwbfile.intervals["Monet2"].to_dataframe()

Generated from smoothened Gaussian noise and a global orientation and direction component.
In [8]:
monet2_dataframe.head()
Out[8]:
start_time stop_time stimulus_type condition_hash rng_seed duration blue_green_saturation pattern_width pattern_aspect temp_kernel temp_bandwidth ori_coherence ori_fraction ori_mix num_directions
id
0 209.198904 224.181996 stimulus.Monet2 WD0uyxvusJmVxfeKu9mZ 1.0 15.0 0 72 1.7 hamming 4.0 2.5 1.0 1.0 16
1 224.281990 239.265071 stimulus.Monet2 kEq0YeYOSSBtM1ZkWA1a 4.0 15.0 0 72 1.7 hamming 4.0 2.5 1.0 1.0 16
2 239.348411 254.331497 stimulus.Monet2 clN4ica4q7huSnWVFTFi 3.0 15.0 0 72 1.7 hamming 4.0 2.5 1.0 1.0 16
3 254.414823 269.397922 stimulus.Monet2 NnOFXtdjpcasByEfeJf9 2467.0 15.0 0 72 1.7 hamming 4.0 2.5 1.0 1.0 16
50 752.506706 767.489783 stimulus.Monet2 yv/S61fYnDSQcA9J6Nk8 7.0 15.0 0 72 1.7 hamming 4.0 2.5 1.0 1.0 16

Trippy¶

The stimulus table for the cosine of a smoothened noise phase movie can be accessed as nwbfile.intervals["Trippy"].

In [9]:
# View the description for this stimulus type
print(nwbfile.intervals["Trippy"].description)
# Load the data into a dataframe object
trippy_dataframe = nwbfile.intervals["Trippy"].to_dataframe()

The stimulus table for the cosine of a smoothened noise phase movie.
In [10]:
trippy_dataframe.head()
Out[10]:
start_time stop_time stimulus_type condition_hash rng_seed texture_height texture_width duration xnodes ynodes up_factor temp_freq temp_kernel_length spatial_freq
id
10 329.863606 344.846688 stimulus.Trippy e9eVr8Agj8TsRJR3XEBg 1007.0 90 160 15.0 12 6 24 4.0 61 0.08
11 344.930015 359.913113 stimulus.Trippy tRgm529SPpQspr220FRm 2.0 90 160 15.0 12 6 24 4.0 61 0.08
12 359.996445 374.979541 stimulus.Trippy hhWRsNy4ljt4M1PtvZ3u 6.0 90 160 15.0 12 6 24 4.0 61 0.08
13 375.062862 390.045950 stimulus.Trippy ZJvulWBItHtLlDVkIntm 8.0 90 160 15.0 12 6 24 4.0 61 0.08
94 1235.298793 1250.281879 stimulus.Trippy EhwIk1vM51LHZPE6xFZX 5.0 90 160 15.0 12 6 24 4.0 61 0.08

Stimulus movie data¶

The visual stimulus composed of parametric stimuli and natural movie clips.

The movie data is stored in an ImageSeries object.

The frames are stored in a 4D array where the first dimension is time (frame), the second and third dimension represents the size of the image and the last dimension are the RGB channels.

In [11]:
# Access the fields for movie data
movie = nwbfile.acquisition["Video: stimulus_17797_9_4_v4"]
movie
Out[11]:

Video: stimulus_17797_9_4_v4 (ImageSeries)

resolution: -1.0
comments: no comments
description: The visual stimulus is composed of natural movie clips ~60 fps.
conversion: 1.0
offset: 0.0
unit: Frames
data
HDF5 dataset
Data typeuint8
Shape(348291, 144, 256, 3)
Array size35.87 GiB
Chunk shape(1, 144, 256, 3)
Compressiongzip
Compression opts4
Uncompressed size (bytes)38518198272
Compressed size (bytes)12228627068
Compression ratio3.14983833081269
timestamps
HDF5 dataset
Data typefloat64
Shape(348291,)
Array size2.66 MiB
Chunk shapeNone
CompressionNone
Compression optsNone
Uncompressed size (bytes)2786328
Compressed size (bytes)2786328
Compression ratio1.0
timestamps_unit: seconds
interval: 1

Data arrays are read passively from the NWB file. Accessing the data attribute of the ImageSeries object does not read the data values, but presents an HDF5 object that can be indexed to read data.

You can use the [:] operator to read the entire data array into memory. (e.g. movie.data[:])

To read only a portion of data at a time, index or slice into the data attribute just like if you were indexing or slicing a numpy array.

In [12]:
# Read the first 500 frames from the movie data
movie_data = movie.data[:500]
In [13]:
# Show frames from the movie.
from matplotlib import pyplot as plt

plt.imshow(movie.data[11900], aspect="auto")
plt.show()

plt.imshow(movie.data[15680], aspect="auto")
plt.show()

plt.imshow(movie.data[110600], aspect="auto")
plt.show()
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image

Two Photon Imaging¶

The two photon data contains 50000 imaging volumes collected at ~8.6 Hz, with six fields per volume. The center of the volume was placed at the junction of primary visual cortex (VISp) and three higher visual areas, lateromedial area (VISlm), rostrolateral area (VISrl) and anterolateral area (VISal).

The imaging data for each field is stored in a TwoPhotonSeries object which can be accessed from nwbfile.acquisition.

The information about the imaging plane can accessed as nwbfile.acquisition["TwoPhotonSeries1"].imaging_plane or from nwbfile.imaging_planes["ImagingPlane1"].

In [14]:
# The two photon imaging data for field 1
two_photon_series1 = nwbfile.acquisition["TwoPhotonSeries1"]
# The timestamps for the imaging data
two_photon_series1_timestamps = nwbfile.acquisition["TwoPhotonSeries1"].timestamps[:]     
print(two_photon_series1.data.shape)
            
# The two photon imaging data for field 2
two_photon_series2 = nwbfile.acquisition["TwoPhotonSeries2"]
            
# The two photon imaging data for field 3
two_photon_series3 = nwbfile.acquisition["TwoPhotonSeries3"]

(50000, 248, 440)
In [15]:
# Visualize the imaging data.

from matplotlib import pyplot as plt

plt.imshow(two_photon_series1.data[2200], aspect="auto")
plt.show()

plt.imshow(two_photon_series2.data[2200], aspect="auto")
plt.show()

plt.imshow(two_photon_series3.data[2200], aspect="auto")
plt.show()
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image

Fluorescence traces¶

The "ophys" processing module contains the fluorescence traces, and image segmentations which can be accessed as nwbfile.processing["ophys"].

The fluorescence traces for each field are stored in RoiResponseSeries objects in a Fluorescence container within the "ophys" processing module as nwbfile.processing["ophys"].data_interfaces.

In [16]:
fluorescence = nwbfile.processing["ophys"].data_interfaces["Fluorescence"]
# The fluorescence traces for field 1 
fluorescence.roi_response_series['RoiResponseSeries1']

roi_response_series = fluorescence.roi_response_series['RoiResponseSeries1']

The traces are stored in a 2D array where the first dimension is time, the second dimension is the number of ROIs for this field.

In [17]:
roi_response_series.data.shape
Out[17]:
(50000, 1410)
In [18]:
# get first 1000 samples of the fluorescence traces of first 100 ROIs
traces = fluorescence.roi_response_series['RoiResponseSeries1'].data[:1000, :100]

# get times of first 1000 samples
timestamps = fluorescence.roi_response_series['RoiResponseSeries1'].timestamps[:1000]

# plot fluorescence data
import numpy as np
import matplotlib.pyplot as plt

offsets = np.cumsum(np.hstack((0, np.max(traces, axis=0)[:-1])))

fix, ax = plt.subplots(figsize=(6, 4))
plt.plot(timestamps, traces + offsets)
ax.set_xlim(min(timestamps), max(timestamps))
ax.set_ylim(0, None)
ax.set_xlabel("time (s)")
_ = ax.set_yticks([])
No description has been provided for this image

The image masks for each field are stored in PlaneSegmentation objects in an ImageSegmentation container within the "ophys" processing module.

In [19]:
image_segmentation = nwbfile.processing["ophys"].data_interfaces["ImageSegmentation"]

# The plane segmentation for field 1 
plane_segmentation1 = image_segmentation.plane_segmentations["PlaneSegmentation1"]

# view plane segmentation table
plane_segmentation1.to_dataframe()
Out[19]:
image_mask mask_type
id
1 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
2 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... artifact
3 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
4 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
5 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
... ... ...
1406 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
1407 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
1408 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
1409 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma
1410 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... artifact

1410 rows × 2 columns

In [20]:
# view example image mask for ROI #5
plt.imshow(plane_segmentation1["image_mask"][5], aspect="auto")
Out[20]:
<matplotlib.image.AxesImage at 0x7f7e5425e210>
No description has been provided for this image

Accessing the coregistration table from the CAVE database ¶

The functional data were co-registered with electron microscopy (EM) data. The structural identifiers of the matched cells are added as plane segmentation columns extracted from the CAVE database.

To access the latest revision see the notebook this notebook. The structural ids might not be present for all plane segmentations.

The "cave_ids" are the identifiers from the CAVE database (can be more than one for a given cell), "pt_position_x", "pt_position_y" and "pt_position_z" corresponds to the location in 4,4,40 nm voxels at a cell body for the cell, "pt_supervoxel_id" is the ID of the supervoxel from the watershed segmentation, "pt_root_id" is the ID of the segment/root_id from the Proofread Segmentation (v117).

In [21]:
image_segmentation = nwbfile.processing["ophys"].data_interfaces["ImageSegmentation"]
# The plane segmentation for field 4
plane_segmentation4 = image_segmentation.plane_segmentations["PlaneSegmentation4"][:]
plane_segmentation4
Out[21]:
image_mask mask_type cave_ids pt_supervoxel_id pt_root_id pt_x_position pt_y_position pt_z_position
id
1 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... artifact [nan] NaN NaN NaN NaN NaN
2 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... artifact [nan] NaN NaN NaN NaN NaN
3 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [4310.0] 1.120332e+17 8.646911e+17 343888.0 107008.0 15649.0
4 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [nan] NaN NaN NaN NaN NaN
5 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [nan] NaN NaN NaN NaN NaN
... ... ... ... ... ... ... ... ...
1400 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [nan] NaN NaN NaN NaN NaN
1401 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [6985.0] 9.275343e+16 8.646911e+17 203737.0 116074.0 23591.0
1402 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... artifact [nan] NaN NaN NaN NaN NaN
1403 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... artifact [nan] NaN NaN NaN NaN NaN
1404 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [nan] NaN NaN NaN NaN NaN

1404 rows × 8 columns

In [22]:
# The structural ids for field 4 
plane_segmentation4.dropna()
Out[22]:
image_mask mask_type cave_ids pt_supervoxel_id pt_root_id pt_x_position pt_y_position pt_z_position
id
3 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [4310.0] 1.120332e+17 8.646911e+17 343888.0 107008.0 15649.0
8 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [2243.0] 1.132301e+17 8.646911e+17 352640.0 111392.0 19327.0
12 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [2244.0] 1.136521e+17 8.646911e+17 355616.0 109728.0 20087.0
16 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [8945.0] 1.138626e+17 8.646911e+17 356899.0 105465.0 22206.0
26 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [2245.0] 1.116119e+17 8.646911e+17 340608.0 113856.0 18439.0
... ... ... ... ... ... ... ... ...
1381 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [2296.0] 1.047158e+17 8.646911e+17 290656.0 113904.0 21738.0
1390 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [9303.0] 9.233204e+16 8.646911e+17 200481.0 122465.0 17207.0
1394 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [4360.0] 1.048560e+17 8.646911e+17 291504.0 110000.0 15556.0
1397 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [4361.0] 1.120325e+17 8.646911e+17 344048.0 101872.0 23999.0
1401 [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,... soma [6985.0] 9.275343e+16 8.646911e+17 203737.0 116074.0 23591.0

219 rows × 8 columns

The average and correlation images for each field are stored in GrayscaleImage objects in an Images container within the "ophys" processing module.

In [23]:
# The average and correlation images for field 1
segmentation_images1 = nwbfile.processing["ophys"].data_interfaces["SegmentationImages1"]
# Load the average image data
average_image = segmentation_images1.images["average"][:]
# Load the correlation image data
correlation_image = segmentation_images1.images["correlation"][:]

segmentation_images1
Out[23]:

SegmentationImages1 (Images)

description: Correlation and average images for field 1.
average (GrayscaleImage)
correlation (GrayscaleImage)
In [24]:
# Show the average and correlation images for field 1
plt.imshow(average_image)
plt.show()

plt.imshow(correlation_image)
plt.show()
No description has been provided for this image
No description has been provided for this image

Behavior¶

The raw behavior data is composed of the velocity of the treadmill, eye movements and the pupil diameter data.

The velocity of the treadmill is stored in a TimeSeries object.

In [25]:
# Access the treadmill velocity
treadmill_velocity = nwbfile.acquisition["treadmill_velocity"]
# Access the timestamps for the treadmill velocity
treadmill_timestamps = treadmill_velocity.timestamps[:]
            
treadmill_velocity
Out[25]:

treadmill_velocity (TimeSeries)

resolution: -1.0
comments: no comments
description: Cylindrical treadmill rostral-caudal position extracted at ~60-100 Hz and converted into velocity.
conversion: 0.01
offset: 0.0
unit: m/s
data
HDF5 dataset
Data typefloat64
Shape(374122,)
Array size2.85 MiB
Chunk shape(2923,)
Compressiongzip
Compression opts4
Uncompressed size (bytes)2992976
Compressed size (bytes)457971
Compression ratio6.535295903015693
timestamps
HDF5 dataset
Data typefloat64
Shape(374122,)
Array size2.85 MiB
Chunk shape(2923,)
Compressiongzip
Compression opts4
Uncompressed size (bytes)2992976
Compressed size (bytes)2207444
Compression ratio1.3558559129925833
timestamps_unit: seconds
interval: 1

The position of the eye is stored in a SpatialSeries object which is stored in an EyeTracking container.

In [26]:
nwbfile.acquisition["EyeTracking"]
Out[26]:

EyeTracking

eye_position (SpatialSeries)
resolution: -1.0
comments: no comments
description: The x,y position of the pupil.The values are estimated in the relative pixel units.
conversion: 1.0
offset: 0.0
unit: px
data
HDF5 dataset
Data typefloat64
Shape(101678, 2)
Array size1.55 MiB
Chunk shape(3178, 1)
Compressiongzip
Compression opts4
Uncompressed size (bytes)1626848
Compressed size (bytes)590626
Compression ratio2.7544469765977118
timestamps (link to acquisition/PupilTracking/pupil_minor_radius/timestamps)
HDF5 dataset
Data typefloat64
Shape(101678,)
Array size794.36 KiB
Chunk shape(1589,)
Compressiongzip
Compression opts4
Uncompressed size (bytes)813424
Compressed size (bytes)604896
Compression ratio1.344733640162937
timestamps_unit: seconds
interval: 1
reference_frame: unknown
In [27]:
eye_position = nwbfile.acquisition["EyeTracking"].spatial_series["eye_position"]

eye_position
Out[27]:

eye_position (SpatialSeries)

resolution: -1.0
comments: no comments
description: The x,y position of the pupil.The values are estimated in the relative pixel units.
conversion: 1.0
offset: 0.0
unit: px
data
HDF5 dataset
Data typefloat64
Shape(101678, 2)
Array size1.55 MiB
Chunk shape(3178, 1)
Compressiongzip
Compression opts4
Uncompressed size (bytes)1626848
Compressed size (bytes)590626
Compression ratio2.7544469765977118
timestamps (link to acquisition/PupilTracking/pupil_minor_radius/timestamps)
HDF5 dataset
Data typefloat64
Shape(101678,)
Array size794.36 KiB
Chunk shape(1589,)
Compressiongzip
Compression opts4
Uncompressed size (bytes)813424
Compressed size (bytes)604896
Compression ratio1.344733640162937
timestamps_unit: seconds
interval: 1
reference_frame: unknown

The pupil major and minor radius are stored as TimeSeries objects and they can be accessed from an EyeTracking container.

In [28]:
pupil_major_radius = nwbfile.acquisition['PupilTracking'].time_series["pupil_major_radius"]
pupil_minor_radius = nwbfile.acquisition['PupilTracking'].time_series["pupil_minor_radius"]
            
pupil_major_radius
Out[28]:

pupil_major_radius (TimeSeries)

resolution: -1.0
comments: no comments
description: Major radius extracted from the pupil tracking ellipse.The values are estimated in the relative pixel units.
conversion: 1.0
offset: 0.0
unit: px
data
HDF5 dataset
Data typefloat64
Shape(101678,)
Array size794.36 KiB
Chunk shape(1589,)
Compressiongzip
Compression opts4
Uncompressed size (bytes)813424
Compressed size (bytes)488421
Compression ratio1.6654156967042777
timestamps (link to acquisition/PupilTracking/pupil_minor_radius/timestamps)
HDF5 dataset
Data typefloat64
Shape(101678,)
Array size794.36 KiB
Chunk shape(1589,)
Compressiongzip
Compression opts4
Uncompressed size (bytes)813424
Compressed size (bytes)604896
Compression ratio1.344733640162937
timestamps_unit: seconds
interval: 1

Time references¶

The timestamps in an NWB file should be in seconds with respect to the global start time of the session. The original time references were relative to scan start which would have produced negative timestamps for the behavior data. Therefore, all times were shifted to the earliest behavioral timestamp.

In this file for instance, time 0.0 corresponds to the first timestamp for the treadmill velocity data while the two photon imaging started 10.899 seconds after.

In [29]:
treadmill_timestamps[0]
Out[29]:
np.float64(0.0)
In [30]:
two_photon_series1_timestamps[0]
Out[30]:
np.float64(10.899238834999977)

To browse the file interactively you can install the optional nwbwidgets package and call nwb2widget(nwbfile). It is not installed here because its current release is incompatible with the pynwb version this notebook uses.

Downloading NWB files ¶

This section demonstrates how to download one of the sessions from this dataset using dandi.

In [31]:
from dandi.dandiapi import DandiAPIClient
from dandi.download import download as dandi_download

dandiset_id = "000402"
file_path = "sub-17797/sub-17797_ses-9-scan-4_behavior+image+ophys.nwb" # file size ~67GB

# The folder where the file will be downloaded
dandiset_folder_path = "microns/nwbfiles/"

# The file path on DANDI
with DandiAPIClient() as client:
    asset = client.get_dandiset(dandiset_id, 'draft').get_asset_by_path(file_path)
    dandiset_url = asset.api_url

# Download the file
dandi_download(urls=dandiset_url, output_dir=dandiset_folder_path, get_metadata=True, get_assets=True)

PATH                                            SIZE    DONE            DONE% CHECKSUM STATUS          MESSAGE
sub-17797_ses-9-scan-4_behavior+image+ophys.nwb 66.8 GB 66.8 GB          100%    ok    done                   
Summary:                                        66.8 GB 66.8 GB                        1 done                 
                                                        100.00%
In [32]:
# Downloading all NWB files for this dandiset

# dandiset_url = "https://dandiarchive.org/dandiset/000402/draft"
# dandi_download(urls=dandiset_url, output_dir=dandiset_folder_path, get_metadata=True, get_assets=True)