BrainFlowZZZ

Dataset from the 2023 manuscript titled **_Sizes and Shapes of Perivascular Spaces Surrounding Murine Pial Arteries_** by Raicevic et al. DOI: 10.21203/rs.3.rs-2587250/v1.

## Overview
The **14 datasets** from **9 subjects** include the original 3D two photon microscopy data from three channels which show tracer in the vessel, PVSs, and microspheres. Additionally, each dataset also includes the final binary segmentation of the PVS and vessel used to generate the model and statistics in the manuscript. Additional details regarding the subjects, tracer injection, image acquisition, and segmentation can be found in the manuscript at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9949243/.

## Content
For easier navigation, below is a mapping between the NWB file names and the datasets referenced in the manuscript.
1. sub-21-07-19-b-act (Mouse 6, dataset K)
2. sub-21-09-01-b-act (Mouse 8, dataset M)
3. sub-21-09-20-b-act (Mouse 9, dataset N)
4. sub-21-10-08-b-act (Mouse 7, dataset L)
5. sub-BPN-M4 (Mouse 3, dataset E and F)
6. sub-BPN-M6 (Mouse 4, dataset G and H)
7. sub-BPN-M7 (Mouse 5, dataset I and J)
8. sub-BPN-OLD-M2 (Mouse 1, dataset A and B)
9. sub-BPN-OLD-M3 (Mouse 2, dataset C and D)

## Data Reading Instructions
The .nwb files can be viewed using PyNWB or MatNWB. To install and set up, please visit <https://www.nwb.org/how-to-use/>. Below, we show how to open and view an .nwb file using MatNWB.
### Loading the image data
```matlab
nwb = nwbRead(PATH_TO_NWB_FILE);

% first check what color channels are present
>> nwb.acquisition
ans =
3×1 Set array with properties:
TwoPhotonSeriesChanA: [types.core.TwoPhotonSeries]
TwoPhotonSeriesChanB: [types.core.TwoPhotonSeries]
TwoPhotonSeriesChanC: [types.core.TwoPhotonSeries]
```
The above code load the .nwb file and the output tells us that there are three channels present in the nwb file, which are ChanA, ChanB, and ChanC. Then, to load the actual data from a channel,
```matlab
% load the image data from ChanA
>> chanAdata = nwb.acquisition.get('TwoPhotonSeriesChanA').data.load();

% check its shape
>> size(chanAdata)
ans =
1 512 512 181
```
### Loading the segmentation masks
For an overview of the mask for ChanA, for example,
```matlab
>> nwb.processing.get('ophys').nwbdatainterface.get('ImageSegmentation').planesegmentation.get('PlaneSegmentationChanA').image_mask.data
ans =
DataStub with properties:
filename: '.\sub-BPN-M4_ses-20210524-m1_obj-1c8nyxo_ophys.nwb'
path: '/processing/ophys/ImageSegmentation/PlaneSegmentationChanA/image_mask'
dims: [512 512 181]
ndims: 3
dataType: 'logical'
```
To load the actual mask data into array (may take several seconds to load),
```matlab
% load mask from ChanA
>> mask = nwb.processing.get('ophys').nwbdatainterface.get('ImageSegmentation').planesegmentation.get('PlaneSegmentationChanA').image_mask.data.load();

% check its shape
>> size(mask)
ans =
512 512 181
```