The basic function of the NEF is to map the outside world onto neurons, using the firing rates of the neurons. It also supports various neuron models. Have these firing rates been used to create EEG and fMRI signals to replicate experimental data before? If so, how?
Replicating fMRI signal was first used in the Tower of Hanoi task as seen in section 5.8 of "How to Build a Brain":
There is strong evidence that dendritic processing, driven by neurotransmitter usage, underwrites the BOLD signal (Logothetis & Wandell, 2004). It is that BOLD signal that is actually measured by MRI machines. Consequently, MRI measures neurotransmitter usage, not neural activity (i.e., spiking of neurons) in a brain area. Neurotransmitter usage is actually determined by neurons that project into the measured area, not the neurons in the area itself. Here we predict fMRI data by filtering neurotransmitter usage in the model. This filter is determined by a fixed model of how neurotransmitter fluctuations demand energy and give rise to the blood-usage-based BOLD signal.
Replicating EEG signals would be possible as well. You would just have to figure out the potential contained in each neuron as a result of it's activities and then pass the resulting signal through a scalp filter.