[pymvpa] Problem with eventrelated_dataset

Roberto Guidotti robbenson18 at gmail.com
Fri Nov 23 11:44:26 UTC 2012


Well, surfing the code I notice that building the BoxcarMapper it uses the
max of the duration of the events and this is reasonable to have the same
feature length for every event.

Solved!
RG

On 23 November 2012 12:25, Roberto Guidotti <robbenson18 at gmail.com> wrote:

> Dear all,
>
> I have had a strange problem building a simple eventrelated_dataset. I've
> loaded my dataset and then I've tried to build up an event related one.
>
> >> events =  find_events(chunks = ds.sa.chunks, targets = ds.sa.targets)
> >> evds = eventrelated_dataset(ds, events=events)
>
> ValueError: Illegal box (start: 656, offset: 0, length: 7) with total
> input sample being 660.
>
> but the last events are:
>
>  {'chunks': 5.0, 'duration': 7, 'onset': 638, 'targets': 'condition1'},
>  {'chunks': 5.0, 'duration': 4, 'onset': 645, 'targets': 'fixation'},
>  {'chunks': 5.0, 'duration': 7, 'onset': 649, 'targets': 'condition2'},
>  {'chunks': 5.0, 'duration': 4, 'onset': 656, 'targets': 'fixation'}
>
> Trying to build the dataset with events[:-1] it works, so it is
> indipendent from the durations, I mean that the events duration could be
> not equal.
>
> Thank you
> Roberto
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20121123/648ba2ce/attachment-0001.html>


More information about the Pkg-ExpPsy-PyMVPA mailing list