Seed-based fMRI resting state analysis

The Resting state fMRI analysis module can be used to generate and visualize correlation maps from acquired resting state fMRI time series. The analysis is a seed-based analysis, used to investigate how the time series in one region in the brain correlates to other regions.

The correlation coefficient is calculated as Pearson correlation coefficient:

where the x_i are the signal intensities of the time curve for the pixel under investigation, and the y_i are the signal intensities of the time curve of the seed region. This latter time curve is calculated as the mean of the time curves of the pixels in the seed region.  

The module is started by selecting Resting State fMRI analysis from the Modules toolbar, with the resting state DICOM data loaded and selected.

Starting the module will create a copy of the original data, and automatically start any preprocessing that is selected (motion correction, Gaussian smoothing and bandpass filtering). Preprocessing steps can be changed and redone (see Preprocessing in resting state fMRI). After preprocessing is done, the seed-curve and corresponding correlation map from a seed ROI position in the middle of the middle slice (lower left image) will be displayed in the main window (right).

You can move the ROI by dragging it around the image, and the time curve and correlation map will update interactively. You can change the displayed slice for the seed ROI using the scroll bar on the upper left image. You can also change the slice of the correlation map using the slider on the left-hand side of the map.

You can change the displayed map in the main window to display positive and/or negative correlations, with a given threshold using the slider below. Note that the negative correlation is given as a positive value.

In the example above, the positive correlation map is thresholded at a relatively high value of 0.829. In this case, the same slice is selected for the seed ROI and the positive correlation map, and displays where in the brain a correlation above the selected threshold can be seen. There is a high degree of correlation with the seed itself, as expected, but also with the area in the opposite hemisphere. Note that this is included as an example only.

Selecting Process will generate the output maps:

If positive, negative or both correlations are selected, the following maps will be generated:

  • The correlation map, showing all correlations (positive and negative) from the selected seed ROI
  • The positive (negative/both) correlation map including the values above the selected threshold only
  • The original input data (first time point) with the positive (negative/both) as overlay, and with the seed ROI included.

If positive and/or negative correlation are not selected, the output map will be the correlation map only.

 

Further reading:

Preprocessing in resting state fMRI

 

Tips

Link images together for simultaneous scrolling through slices using shift + left click. See Options -> General Options -> Image Linking.

 

 Logo

Description automatically generated