Hello,
I simulated fmri data via R’s neurosim. For 20 subjects, there was a 200-second experiment in which the onset (only a single stimulus) occurred at every 30 seconds, starting at 1s. TR=1, and I used a 64x64 grid, so there are 4096 voxels… I then convolved these onsets with the canonical HRF to get the evoked stimulus, and ran a mixed effects model: Bold=EV+subject, where subject is random. Unfortunately, some voxels, not all, return a singularity error. Does anyone know why I would get singularity issues in a design such as this? Thank you.