# How are isopotential maps computed from EEG channel data

I have some EEG recordings of $1$ minutes, with $500$ samples per seconds and $63$ channels. For each recordings, this effectively yields a $30,000\times 63$ matrix $M$.

How do I extract isopotential maps from the data?

My first try was to simply take each row $m_i\in M_{i,j}$ such that $m$ has length $63$ and plot them with respect to the $x,y$ coordinates given by the location of each of the $63$ channels... However, I seem to get bad plots...

I also tried to bandpass the data first. For alpha waves, my Mathematica code looks like this:

BANDPASSALPHA[x_] :=
BandpassFilter[x, {8 Pi, 13 Pi}, SampleRate -> 500]
bpalpha = Map[BANDPASSALPHA, dataz];


I bandpassed the columns of the matrix $M$. That is the signal $x$ that is used in BANDPASSALPHA[x] is one of the columns of $c_j\in M_{i,j}$ and are of length $30,000$

According to Wikiepdia: http://en.wikipedia.org/wiki/Isopotential_map, the spatial derivative needs to be computed first...

Does anyone know how isopotential maps from EEG recordings?