Basic fNIRS Data Analysis

with Homer

Neurophotonics Center, Boston University fNIRS Workshop, November 7-9, 2018

Prepared by Juliette Selb

Outline

◊ More advanced processing

◊ Basic processing :

◊ Homer file formats

◊ Other useful Homer functions◊ Exporting results

◊ Plot probe◊ Group analysis

◊ Intensity to hemoglobin concentrations◊ Filtering◊ Block-averaging

File formats

Probe source-detector geometry : .sd files

fNIRS experimental data : .nirs files

Data processing stream : .cfg files

Processed data files : groupResults.mat

.SD file format

Lambda Wavelengths used for data acquisition 1 × nWavelengths [690 830]

nSrcs Number of sources 1 × 1 1

nDets Number of detectors 1 × 1 2

SrcPos Array of source coordinates nSrcs × 3 [ 0 0 0]

DetPos Array of detector coordinates nDets × 3 [30 0 0;0 30 0]

MeasList List of measurement channels nChannels × 4 [1 1 1 1;1 2 1 1;1 1 1 2;1 2 1 2]

Note: The SD structure is also incorporated in the .nirs file format we will see in a few slides

Describes the 2D source/detector geometry

Field Definition Size Example

column 1 column 2 column 3 column 4 source idx detector idx unused wavelength idx

Matlab .sd file containing the structured variable SD with the following fields:

.SD file format: the SDgui

Sub-GUI of AtlasViewer

Enables you to create your own .sd files

6

Start AtlasViewer

Go to Tools > Make Probe → it starts the SDgui

.SD file format: the SDgui

7

.SD file format: the SDgui

We will go over this in

detail on Day 3.

.NIRS file format

Structured variable

NTime_points × 1

NTime_points × NChannels

NTime_points × NConditions

NTime_points × NAux channels

Contains an fNIRS dataset

Field Definition Size

Matlab .nirs file containing the following variables:

SD Probe geometry, same as .sd files

t Array of data time points

d Array of raw intensity time course

s Array of stimulus onsets. Different conditions are a way to separate stimuli into different groups

aux Array of auxiliary signals

Note 1: In addition to the above parameters, additional parameters can appear in the .nirs file as a result of processing the data set in Homer2.

Note 2: Conversion scripts are available for some other data formats (Hitachi, NIRx, Shimadzu, …), see http://homer-fnirs.org/faq/

.CFG file format

Note: in options, you can

hover over each function to get

help on the function and the

parameters

A .cfg file describes the steps of data analysis, a sequence of Homer2 functions

Intensity to ΔOD

ΔOD to Concentration

Block- average

ΔOD

Concentration

Hemodynamic response function

(HRF)

Intensity

Band-pass filter

Identify MAs

Prune data

Remove / Correct MAs

Regress short SD

GLM

Ignore segments

Simplest processing stream: Multiple more advanced functions we can incorporate in the stream:

Let’s create a processing stream.

• Launch HOMER2

• Option 1: Executable version: Run HOMER2.exe

• Option 2: MATLAB version: Run MATLAB, run “setpaths”

function and then call EasyNIRS (or Homer_UI) from

command window.

• Hit cancel when it asks you to load process options

file

• Go to File

• Change Directory to following folder:

\fNIRS_COURSE_2018\EXERCISES\Ex1_Basic_data_analysis\1-basic-steps

.CFG file format: the Process Stream GUI

The Process Stream GUI enables you to create your own .cfg files

Go to Tools > Process Stream GUI → it starts procStreamGUI

.CFG file format

The list of functions to choose from

InputsOutputs Function Name

The processing stream you’re creating

InputsOutputs Function Name

The help information for the

selected function

Add, remove, and move functions in the stream

.CFG file format

1 2

Intensity to ΔOD

ΔOD

Intensity

.CFG file format

Add:• hmrIntensity2OD• hmrBandpassFilt• hmrOD2Conc• hmrBlockAvg

Intensity to ΔOD

ΔOD to Concentration

Block-average

ΔOD

Concentration

Hemodynamic response

function (HRF)

Intensity

Band-pass filter

ΔOD, filtered

.CFG file format

1

2

.CFG file format

Give a name to your new processing stream

and save

.CFG file format

You can change your current processing stream in Homer2File > Load Process Options (*.cfg)

.CFG file format

1

2

.CFG file format

Click on ‘Options’

This will open the ProcessOptwindow, where you can modify the parameters of each function

Let’s analyze some fNIRS data: Homer2_UI

When opening Homer2, you need to select:• The dataset(s): a folder containing at least one .nirs file• The data processing stream: a .cfg file. If you hit ‘Cancel’, a default process stream will load.

• Start Homer2• When prompted to select a .nirs folder, navigate to and select folder

• When prompted to select a .cfg file, select my_simple_processing.cfg file we just created.

Note: If you’re already in Homer2 and need to change the .nirs directory or the .cfg file:• Go to File > Change Directory, and select your nirs folder• Go to File > Load Process Options (*.cfg) and select your .cfg file

\fNIRS_COURSE_2018\EXERCISES\Ex1_Basic_data_analysis\1-basic-steps

\fNIRS_COURSE_2018\EXERCISES\Ex1_Basic_data_analysis\1-basic-steps\my_simple_processing.cfg

Homer2_UI

Source / detector geometry

Data time series

.nirs files in your folder

Modify options and run processing stream

Display options

What data type to display

Displaying the raw fNIRS data

Click next to the source

Displaying the raw fNIRS data

Right click on a channel will remove it from the display.

Displaying the raw fNIRS data

Left click on a channel will remove it from the analysis.

Click again to return back.

Displaying the raw fNIRS data

Select wavelength to display

Zooming in

Zoom in

Zooming in

Modifying the processing stream options

1 - Open the processing options

2 - Set hpf (high-pass filter) and lpf (low-pass filter) to 0, i.e. no filtering

Running the processing stream

hpf and lpfare set to 0

Select Run

Zoom reset

Intensity to ΔOD

Select OD

ΔOD to Hb concentrations

Select Conc

ΔOD to Hb concentrations

Choose HbO/HbR/HbT to display.Use Control or Shift to show more than one.

Band-pass filtering

Showing HbO

x-range fixed

Band-pass filtering

hpf = 0.00lpf = 0.5

Select Run

Band-pass filtering

hpf = 0.00lpf = 0.5

unchecked

Band-pass filtering

hpf = 0.01lpf = 0.5

Select Run

Band-pass filtering

hpf = 0.1lpf = 0.5

Select Run

Go back to hpf = 0.01 and re-run

hpf = 0.01lpf = 0.5

Select Run

This dataset doesn’t have stimulus onsets, we will enter them manually.

Go to Tools > StimGUI

StimGUI

You can add one or multiple stimulus marks.

StimGUI

StimGUI

StimGUI: delete the stim mark

(1) Select stim mark by mouse.

(2) Delete

StimGUI

You can create stim marks using the auxiliary signals

StimGUI

(1) 22:25:347(stim marks every 25 seconds starting at t = 22s)

(3) Save

(2) Select 1

Stimulus onsets

Back to the Homer2_UI main window, the stim marks are visible.

Stimulus onsets: rejecting stim marks

(2) Select ‘Stim Reject’

(3) Drag a box around the stim mark to edit

(1) Go back to raw data

Stimulus onsets: rejecting stim marks

Stimulus onsets: rejecting stim marks

Stim mark now appears as dotted line (toggled off)

Stimulus onsets: rejecting stim marks

Let’s say we want to remove these two stimulus onsets from the block-average

Displaying the HRF

(2) Show Run HRF

(1) Run

Displaying the HRF (shift + left click)

Display both HbO and HbR

Exporting results

(2) Right-click on a time trace.

(1) Should be unchecked.

Display options: waterfall

Waterfall display of channels

Visualize all HRFs: the Plot Probe feature

(2) Check ‘Plot Probe’

(1) Back to zero

Visualize all HRFs: the Plot Probe feature

Show time markers

Other useful Homer processing functions

See ‘Motion artifacts’ lecture

See ‘GLM and short separation regression’ lecture

Group average: data structure

SubjectName1_run01.nirs SubjectName1_run02.nirs . . . . . . . . . . . . . SubjectName1_runn1.nirs

SubjectName2_run01.nirs SubjectName2_run02.nirs . . . . . . . . . . . . . SubjectName2_runn2.nirs

. . . . . . . . . . . . .

SubjectNamem_run01.nirs SubjectNamem_run02.nirs . . . . . . . . . . . . . SubjectNamem_runnm.nirs

Any directory that contains one or more files with the .nirs extension is considered by Homer2_UI to be a group, organized into sub-groups called sessions (or subjects).

You can compute the HRF:

• at the run level

• at the session (or subject) level

• at the group level

Data processed in Homer2_UI is saved automatically to disk: • Run data is saved in the corresponding .nirs file• Subject and group processing results are saved to

groupResults.mat

To go back to the state before any processing was done to the dataset, go to File > Reset File

Group average: exampleFile > Change directory to:

\fNIRS_COURSE_2018\EXERCISES\Ex1_Basic_data_analysis\2-group-averaging

Group average: exampleFile > Load Process Options (*.cfg)

Choose your processing stream *.cfg

Group average: example

Folder structure

Group average: example

(3) You can run the analysis at the run level, session level, or group level

(1) Choose a specific run

(2) Click on the display to visualize signal

Group average: example

Compute Run HRF

Group average: example

Compute Session (Subject) HRF

Group average: example

Compute Group HRF

Group average: example

Note: All subjects can also be in the same folder, provided they have the proper name nomenclature

Group average: example

We will get back to this dataset tomorrow and perform more advanced data analysis.