User's Guide‎ > ‎Analysis‎ > ‎

### Step 1: generating IS

- using the file calculator, multiply your ERP files with your IS so to transform your data from the sensor space to the brain space.

Alternatively,

- Using the export function, extract your period of interest (i.e. microstate where topographic/GFP modulation occurs) for each subject and average               in time.

Use the latter .ep file for the inverse

### Step 2: visualizing IS

-          create a .lm  file with:

a.       avg*.spi = matrix for solution points

b.       avg*T1.fld = T1-weigthed MR image

c.       avg*gray.fld = gray matter volume

-          read the output of the preceding step (i.e. the is-files) into Cartool

a.       You can either take just one condition or the conditions you want to contrast….

b.       Link the lm-file and the ep-file by clicking the icon “+”

-          click into the window comprising the inverse solutions and choose the (probably) optimal source distribution to explain your data

-          for this selection cycle through the solutions by clicking the “I” button (second from the right in icon line)

-          If you want to compare the responses to two conditions make sure to align the scales of the images! Go to “Options” >>> “Specify display scaling”

Making Inverse Solution Figures

-          have the laura-111.lm  from the “avebrain” -folder open and the *is.ep files you produced in the preceding step

-          choose your optimal inverse solution by cycling through the modes (click the “I” button again, the second from the right in icon line)

-          have the display scalings between conditions aligned as before…..

1. Producing Top views (on axial slices)

- cycle through the slice mode options by clicking the “////” button in the icon bar until you find the horizontal (axial) cuts

- choose slice with the maximal values by clicking the “+” button and note its z-coordinate

-export the window (CTRL+C) into a graphics program and choose your favourite slice there

2. Producing Side views (sagittal)

- remove all the visible planes (x, y, z) from your open is.ep brain view file (you only see the orientation axis then)

- enlarge *.gray.fld by double-clicking on it in the *lm-window in the left corner of Cartool

- eventually downsample the isosurface of the MNI brain in “Options” > Specify Downsampling Quality (set to “2” to make it look nice and smooth)

- then click again on one of your *is.ep brain views and activate it by this

- now link this *.ep with the *.gray.fld by clicking the respective icon (Attention: It does not work the other way around)

- specify the brain orientation you want by clicking the “timer” icon in the icon bar or just type “O” on your keyboard

-          For to change the transparency of the MR image go to “Options” >>> “Depth shifting trick for 3D inverse” and change the parameter to “Light” (this is somehow obligatory)

-          Now click “Show the Inverse Solutions” in the icon bar and choose the display you want

-          TIPP: If colors are too boring or few click on “Increase contrast” and play around

For to insert the gray.fld with the IS into the scull of the “dummie” >>>

-          open the *T1.fld by clicking on it in the *.lm window

-          click on “Y” and “Z” in the task bar to remove the upper calotte of the dummie

-          remove more of it by pressing CTRL on the keyboard and scrolling down with the right mouse button pressed (“default” is around 66)

-          click again on your *is.ep and then link each of them separately with the *T1.fld

NOTE: There is an fMRI-like approach to the inverses: First, visualize your inverse and choose the appropriate scaling (e.g. S=1.3*10^-3). Then subtract your inverse (e.g. C1-C2) and scale the subtraction to S/2 or S/3 so that each source appearing in the subtraction visualization represents a region as being twice or three times stronger or lower in one condition than in the other condition. Then, do the statistic comparison on these ROI only. Values of the solution points within ROIs can also be used as input for correlationnal analysis (see e.g. Spierer et al., 2008, Neuroimage).