Dear Chu Liu, You should use the source extraction functionality (spm_eeg_inv_extract) accessible via batch or GUI. Best, Vladimir On Wed, Mar 5, 2014 at 2:14 AM, Chu Liu <[log in to unmask]> wrote: > dear all > > I need your advance about lead field matrix and how to get the projector > of row signals(sensor space) on source space. > > > description: My row signals are obtained from 275-channel MEG at a > sampling rate of 300HZ, 2 seconds. resting-state. > after preprocessing of SPM8, I get a structure D contains a data matrix of > <275*600> > > question: I implemented 3D source reconstruction to the structure D, in > the step of forward model, I can get a matrix <275*8196> which > has SPMgainmatrix in its name. It is regard as "so called" lead field in > the SPM manual. 8196 is the number of dipoles. > If I want to get the projector of signal from sensor space to source > space, can I multiply the "so called lead field' matrix <275*8196> with > data matrix of structure D(<275*600>) and get a matrix of <8196*600>? > > or as the manual says, the "so called lead field" is just former step of > "inverse". after inverse, I can get many new projector ,such as U,L,J,T. > Among them, J is regarded as "condition projector", T is "time projector". > can I get the projector of signal from sensor space to source space by > multiply J with T? U is spatial pojector and L is lead field, can I get the > projector of signal from sensor space to source space by multiply U with L? > > > I will appreciate your assistance very much. >