Terry Morris is in charge of distributing the 2000 SPM course video. Does
anyone know his email address? I want to order the video but the price in
american dollars is not listed on the website.
> ----------
> From: Klaus Hönig
> Reply To: Klaus Hönig
> Sent: Friday, January 12, 2001 1:22 PM
> To: [log in to unmask]
> Subject: mixed model: events and epochs
>
> Dear SPMers,
>
> the experimental design I adopted is a mixture of a block design and
> different (2) events within each of the two blocks.
> In the first case, I specified a pure event-related model for four types
> of
> events:
>
> [type 1 events,baseline],[type 2 events, baseline], [type 1 events,
> activation], and [type 2 events, activation]
>
> In order to detect greater activation in type2 events compared to type1
> events, irrespective of baseline or activation epoch, is it correct to
> specify the following two contrasts and combine them in a conjunction
> analysis?
>
> -1 1 0 0 and 0 0 -1 1
>
> What would be the difference to one single contrast ?
>
> -1 1 -1 1
>
> In a second model I added the two box-car regressors to the four event
> trains:
> Two 'box-car' regressors (1 for the 'baseline' epoch, and 1 for the
> 'activation' epoch) and 4 event trains for [type 1 events,baseline], [type
> 2 events, baseline], [type 1 events, activation] and [type 2 events,
> activation]
>
> Would the following contrast
>
> -1 1 0 0 -1 1
>
> detect those voxels that are activated more during the activation epoch
> than during the baseline epoch and additionally more activated during
> type2
> events than during type1 events in the activation condition?
>
> Would a masking procedure instead be more appropriate?
>
>
> Thanks in advance for helpful comments.
>
>
> Klaus
>
>
>
> ------------------------------------------------------------------
> Klaus Hönig - Universität Bonn - Psychiatrische Klinik
> Sigmund-Freud-Str. 25 - 53105 Bonn - Tel. 0228-2876946
> Fax 0228-2876097 - EMail [log in to unmask]
> ------------------------------------------------------------------
>
>
|