I've got a problem with flexible factorial design, both in "all scans", both in "subjects by subjects" mode (SPM, last patch installed, working on matlab 7.1).
I've a design with almost 20 subjects, three acquisition in tree different times, tre tasks each acquisition, performed in two different ways.
Because we were primarly interested in "acquistion time" effect, i thought the best second level model was a flexible factorial design.
I tried to enter the scan in one go, and i tried to enter the scans subject by subjects, in both cases SPM design the matrix, start estimating the model, but then give me an error message about a non-existing field Vbeta (that in fact doesn't exist in the SPM structure, i've checked).
I don't really can figure out where is the problem: i've checked the first level and the contrast i've entered in the second level (a lot of contrats :) , and everything seems good to me.
Maybe I'm doing something wrong in the "conditions" field of each subject (or in the "factor matrix), for the "one go")?
I've designed a matrix with one factor only ("time of acq") and for each subject I entered the contrasts, entering in "conditions" and array of number of scan x 1, writing 1 for the first acquistion, 2 for the second etc.