To start with, if you come from fMRI background, there are few important differences between the haemodynamic imaging (fMRI, PET) and neurophysiological imaging (MEG, EEG) approaches, all closely related to each other:
1. Timing aspect. In MEG/EEG, you can follow brain activation with millisecond precision. Using the CBU machine, you will typically sample your MEG signal with 1kHz sampling rate, which will enable you to have a snapshot of the brain activation every millisecond at all 306 sensors.
MEG activations, as a direct reflection or neuronal activity, last considerably less time than BOLD, typically only a few hundred milliseconds.
During this short time, the activation is usually not uniform and has complex dynamics.
2. As the result of the above, time-locking of your stimuli to the MEG data becomes very important. Here, the major methodological difference with fMRI is that you have to send triggers from your stimulation equipment to the MEG acquisition machine and record them with the data. These can later be used for time-locking brain responses to the stimuli. Unlike the fMRI experiments, the MEG machine will not send any pulses or triggers to the stimulation computer.
The same refers to any responses you would like to collect and refer your data to. Triggers from button presses, voice reactions etc should be fed to the acquisition machine to be recorded with the data.
3. The time course of activation is not known beforehand, and you will not be fitting your data to any function (cf. HRF convolution in fMRI). You will, instead, record the activation patterns as they appear in the recording and simply analyse their parameters (amplitude, latency, topography) numerically (after a number of preprocessing steps, of course.)