the screenshot shown above is:ERP plot for gradiometers(left) and ERP plot for magnetometer(right) .And the pattern shown with two types of sensors are different,when the ERP plot for gradiometers shown P1 component,the the ERP plot for magnetometer shown N1 component(2D layout);just like in the time series pattern(show below,top red line is gradiometers,and bottom line is magnetometer which I already check)
I think this is just a scalling difference between gradiometers and magnetometer ,but what should I do if I need to use like N1 and P1 component to explain my data
You cannot expect to process MEG signals as EEG ERP.
The value read by a MEG magnetometer depends not only on the amplitude of the field at this point in space (like an EEG electrode) but also on its orientation. Change the orientation of the coil, you flip the sign. It gets even more complicated with gradiometers, they produce derivatives of the magnetic field in a given direction. You cannot relate what you observe here directly to the EEG ERP litterature.
In general signs of MEG signals are difficult to understand, especially with Elekta systems, which mix different types of sensors in the same machine. This is one of the reasona for which when processing MEG signals we commonly prefer to work in source space (spatial estimation of the sources) or frequency space (power in specific frequency bands).
Thank you for your quick answer,it’s very helpful;Another question I want to figure out is :how does brainstorm merge magnetometer and gradiometers into one single ERP plot?(these two types of sensors’ signal scale is far away from each other)
Another element. The gradiometer are made as the difference between two (or mor) magnetometers. The amplitudes recorded will taper fast with the distance to the source. This is good when you have strong noise sources in the area. The magnetometer do not have this tapering. So if you don’t have a big noise problem, it is my oppinion that you should work with magnetometers.
