There is no need to use a 2 stage procedure. All that is needed is to
weight each run either by the number of trials or by the number of
runs with the condition. The former is a better way. Weighting would
be number of trials in the run / total number of trials. If you do it
this way, then the results are approximately the same as the 2 stage
procedure.
One of the problems with the 2 stage procedure is that the predicted
HRF will use points from the subsequent run when trials are close to
the end of the run, even though those points have no relationship to
the previous trials of the earlier run.
Best Regards, Donald McLaren
=================
D.G. McLaren, Ph.D.
Postdoctoral Research Fellow, GRECC, Bedford VA
Research Fellow, Department of Neurology, Massachusetts General
Hospital and Harvard Medical School
Office: (773) 406-2464
=====================
This e-mail contains CONFIDENTIAL INFORMATION which may contain
PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED
and which is intended only for the use of the individual or entity
named above. If the reader of the e-mail is not the intended recipient
or the employee or agent responsible for delivering it to the intended
recipient, you are hereby notified that you are in possession of
confidential and privileged information. Any unauthorized use,
disclosure, copying or the taking of any action in reliance on the
contents of this information is strictly prohibited and may be
unlawful. If you have received this e-mail unintentionally, please
immediately notify the sender via telephone at (773) 406-2464 or
email.
On Sun, Jan 23, 2011 at 4:50 PM, Joonkoo Park <[log in to unmask]> wrote:
> Dear all,
> I have a question about multi-session data and how I can concatenate them.
> I am collecting data from many sessions (or runs depending on how you want
> to call them) per subject. Normally, the design matrix would include
> conditions for each session in separate columns. However, there are some
> cases where some specific experimental conditions of interest occur only few
> times (or does not occur at all) within one session. This causes problems in
> estimation of the regressors. So, I would like to concatenate data over all
> sessions and run one big model so that each conditions of interest are well
> represented in the design matrix. The problem with this approach, however,
> is that there may be spill over effects between sessions (especially when
> using a highpass filter and AR(1) model). What's the best solution to this
> problem?
> I was thinking of using the ordinary model matrix but just using the
> constant only (actually number of sessions -1) for each session, while
> including a high-pass filter and an autoregressive model. Then, I was going
> to use the residuals from this model in a second model where all the
> conditions of interests are entered in one long column (concatenated across
> sessions). Does this make sense? Is there a better approach to this problem?
> Thanks in advance.
> Joon
|