 |
 |
Dear SPM'ers I know there is a lot of high-pass filter questions, hopefully there is room to answer one more. I am wondering if somebody could help and recommend the best high pass filter for me to use to pick up learning changes in my cognitive task. My design is based on pilot studies to optimise learning as measured behaviourally. I have 8 blocks of a cognitive task, each block consists of 22 trials (TR = 3s). Each of these blocks is preceded by a 'prompt' block of 2s. I have three such runs, two new learning and one no learning. 2 blocks of the baseline task (2 x 66s) 4 blocks of the experimental task (4 x 66s) 2 blocks of the baseline task (2 x 66s) I am interested in looking at the linear change with time over all the blocks and so I include a parametric modulation of time factor in the model. However SPM sets a default high-pass filter of 128 seconds and so I miss a lot of signal. Would it be valid if I used a high pass filter of (4 x 72s)? or even double this to look at the learning effect. Thanks a lot for any suggestions Nick Nicholas Walsh, PhD Student Department of Clinical Neuroscience, Centre for Neuroimaging Sciences, Box 089, Institute of Psychiatry, De Crespigny Park, London SE5 8AF, UK. email: [log in to unmask] phone: +44 20 7919 3053 fax: +44 20 7919 3074 http://internal.iop.kcl.ac.uk/ipublic/staff/show/?go=10220