If you are talking about DCM for ERP, you should always include the onset of the ERP in your model. You cannot start modelling from the middle of the waveform because the model should 'know' how to get to that point. If you include the onset you can choose when to end the ERP. This was done systematically in Marta Garrido's paper about feedback loops.
Sent using BlackBerry® from Orange
From: 飞鸟 <[log in to unmask]>Sender: "SPM (Statistical Parametric Mapping)" <[log in to unmask]>Date: Thu, 26 May 2011 16:08:41 +0800To: <[log in to unmask]>ReplyTo: 飞鸟 <[log in to unmask]>Subject: [SPM] Can DCM deal with short-term signal?Dear SPM's users,Simply put, if I have a long-term signal, such as 0~1000 ms, could I divide it into 20 short-term signals(0~50,51~100,...956~1000) to do DCM analysis? In other words, does DCM has any preferences for the time of signal, long or short? Or neither of them?Thank you very much for any advice!--
Haoran LI (MS)
Brain Imaging Lab,
Research Center for Learning Science,
2 Si Pai Lou , Nanjing, 210096, P.R.China