If that doesn't reveal anything, the best debugging approach I know is
to create some 'fake' (i.e. simulated/synthetic) data for an individual
subject, and feed it through your analysis. If the 'fake' data is nearly
perfect, you should see nearly perfect performance. If you feed in
nonsense or noisy fake data, you should see performance degrade towards
baseline. Creating fake data can be tricky, especially if your
experiment/analysis are complex, but 100% worth the effort. See e.g.:
http://code.google.com/p/princeton-mvpa-toolbox/source/browse/trunk/core/synth/noisify_regressors.m
If your analysis works as expected on fake data, you can start to
breathe more easily that it's not a straightforward bug. At that point,
start to think about ways to visualize what's going on.
g
> Perhaps the data is somehow being overfitted?
In NNs, if the training error goes very low, you might have overfitting, you may try another classifier, e.g. Logistic Regression. Have you done detrending and z-scoring to your data?
Regards,
Rawi