We have developed an on line real time brain mapping which also is used in carotide research for the decision "shunt or not to shunt" by calculating a so-called "mirror mode" , where differences in electric power (after FFT) are displayed to show the blood supply from the other hemisphere. This Softare runs under Windows. You can download it from Mewicon GmbH. Please see link.

Best regards

The details will depend a very a bit depending on things like what analysis software you're using, what frequency range you're measuring and what population you're interested in (children, patients, unselected adults, etc.); but, in general, the steps go something like this:

1) Do a little bit of  filtering - although less than you might do for a time domain study. A notch filter will probably be enough (use 50 or 60 Hz depending on what the line frequency is where you live).

2) Reject artifacts. In frequency domain studies, it's often better to remove ocular artifacts rather than correct for them. So choose some criteria (the criteria will likely depend on your population - e.g. children typically have noisier data than adults) for detecting artifacts and remove those data.

3) Segment your data into segments lasting for something like 2 seconds and overlap them by about 1.5 seconds to correct for issues that arise when you impose the Hamming window.

4) Rereference your data. Mastoids and average are typical - however, there are some data that suggest that the choice of reference can really make a difference for asymmetry. I would recommend using the reference scheme that is most appropriate for your field/research questions.

5) Conduct a fast fourier transformation with a Hamming window. Taper the most distal data of each epoch (the most distal 10% should do). Convert to power spectra and average across epochs.

6) Compute power in the spectrum of interest and then asymmetry scores (typically right minus left). The details will depend a bit on which spectrum you're interested in. For example, it's common to compute alpha asymmetry as LN(right) minus LN(left).

Hope that helps

Wilfred..what link?

Hi Valia, perhaps this paper helps:

Issues and assumptions on the road from raw signals to metrics of frontal EEG asymmetry in emotion. Biological Psychology, 67,183-218.