Note

Go to the end to download the full example code

Reading and Exploring an NWB File¶

This tutorial will demonstrate how to read, explore, and do basic visualizations with an NWB File using different tools.

An NWBFile represents a single session of an experiment. It contains all the data of that session and the metadata required to understand the data.

We will demonstrate how to use the DANDI neurophysiology data archive to access the data in two different ways: (1) by downloading it to your computer and (2) streaming it.

We will briefly show tools for exploring NWB Files interactively and refer the reader to the NWB Overview documentation for more details about the available tools.

See also

You can learn more about the NWBFile format in the NWB File Basics tutorial.

The following examples will reference variables that may not be defined within the block they are used in. For clarity, we define them here:

import matplotlib.pyplot as plt
import numpy as np

from pynwb import NWBHDF5IO

We will access NWB data on the DANDI Archive, and demonstrate reading one session of an experiment by Chandravadia et al. (2020). In this study, the authors recorded single neuron activity from the medial temporal lobes of human subjects while they performed a recognition memory task.

Download the data¶

First, we will demonstrate how to download an NWB data file from DANDI to your machine.

Download using the DANDI Web UI¶

You can download files directly from the DANDI website.

Go to the DANDI archive and open this dataset
List the files in this dataset by clicking the “Files” button in Dandiset Actions (top right column of the page).

Choose the folder “sub-P11MHM” by clicking on its name.

4. Download the NWB data file “sub-P11HMH_ses-20061101_ecephys+image.nwb” to your computer by clicking on the download symbol.

Downloading data programmatically¶

Alternatively, you can download data using the dandi Python module.

from dandi.download import download

download("https://api.dandiarchive.org/api/assets/0f57f0b0-f021-42bb-8eaa-56cd482e2a29/download/", ".")

PATH                                      SIZE    DONE           DONE% CHECKSUM STATUS          MESSAGE
sub-P11HMH_ses-20061101_ecephys+image.nwb 72.6 MB 72.6 MB         100%    ok    done
Summary:                                  72.6 MB 72.6 MB                       1 done
                                                  100.00%

See also

Learn about all the different ways you can download data from the DANDI Archive here

See also

Streaming data

Instead of downloading data, another approach is to stream data directly from an archive. Streaming data allows you to download only the data you want from a file, so it can be a much better approach when the desired data files contain a lot of data you don’t care about. There are several approaches to streaming NWB files, outlined in Streaming NWB files.

Opening an NWB file with NWBHDF5IO¶

Reading and writing NWB data is carried out using the NWBHDF5IO class. NWBHDF5IO reads NWB data that is in the HDF5 storage format, a popular, hierarchical format for storing large-scale scientific data.

The first argument to the constructor of NWBHDF5IO is the file_path. Use the read method to read the data into a NWBFile object.

filepath = "sub-P11HMH_ses-20061101_ecephys+image.nwb"
# Open the file in read mode "r",
io = NWBHDF5IO(filepath, mode="r", load_namespaces=True)
nwbfile = io.read()
nwbfile

/home/docs/checkouts/readthedocs.org/user_builds/pynwb/envs/stable/lib/python3.8/site-packages/hdmf/backends/hdf5/h5tools.py:213: UserWarning: No cached namespaces found in sub-P11HMH_ses-20061101_ecephys+image.nwb
  warnings.warn(msg)

session_description: New/Old recognition task for ID: 9.

identifier: H11_9

session_start_time: 2006-11-01 00:00:00-07:00

timestamps_reference_time: 2006-11-01 00:00:00-07:00

file_create_date

2020-05-26 19:08:31.224415-07:00

related_publications: ('Faraut et al. 2018, Scientific Data; Rutishauser et al. 2015, Nat Neurosci;',)

acquisition (2)

events

resolution: -1.0

comments: no comments

description: The events coorespond to the TTL markers for each trial. For the learning trials, the TTL markers are the following: 55 = start of the experiment, 1 = stimulus ON, 2 = stimulus OFF, 3 = Question Screen Onset [“Is this an animal?”], 20 = Yes (21 = NO) during learning, 6 = End of Delay after Response, 66 = End of Experiment. For the recognition trials, the TTL markers are the following: 55 = start of experiment, 1 = stimulus ON, 2 = stimulus OFF, 3 = Question Screen Onset [“Have you seen this image before?”], 31:36 = Confidence (Yes vs. No) response [31 (new, confident), 32 (new, probably), 33 (new, guess), 34 (old, guess), 35 (old, probably), 36 (old, confident)], 66 = End of Experiment

conversion: 1.0

offset: 0.0

unit: n/a

data

timestamps

timestamps_unit: seconds

interval: 1

experiment_ids

resolution: -1.0

comments: no comments

description: The experiment_ids coorespond to the encoding (i.e., learning) or recogniton trials. The learning trials are demarcated by: 80. The recognition trials are demarcated by: 81.

conversion: 1.0

offset: 0.0

unit: NA

data

timestamps

timestamps_unit: seconds

interval: 1

stimulus (1)

StimulusPresentation

resolution: -1.0

comments: no comments

description: no description

conversion: 1.0

offset: 0.0

unit: meters

data

timestamps

timestamps_unit: seconds

interval: 1

dimension

format: raw

distance: 0.7

field_of_view

orientation: lower left

keywords

epoch_tags: set()

electrodes

description: metadata about extracellular electrodes

id

colnames: ('x', 'y', 'z', 'imp', 'location', 'filtering', 'group', 'group_name', 'origChannel')

columns: (, , , , , , , , )

electrode_groups (9)

Neuralynx-cheetah-microwires-19