2011/3/10 飞鸟 <[log in to unmask]>:
> Dear Vladimir,
> 　　Thanks for your detailed reply. As for what you said just now, I still
> have two questions:
> 　　1. You said that 'There are
> some other cases when they might be used, when it is known that two
> areas are at the same level of cortical hierarchy. '
> 　　　Do you mean that two different sources may have lateral connections as
> long as they are in the same level ? Or at least, I can assume there exist
> lateral connections between them when I construct a specific model ?

Yes

 In
> addition, could you recommend me any literatures about cortical hierachy ?

The classical reference is http://www.ncbi.nlm.nih.gov/pubmed/1822724
There might be also more recent reviews on the topic.

>       2. I'm not very clear about 'Note that if
> there is bi-lateral symmetry in your model it wouldn't make sense to
> assign different types to symmetric connections in the two hemispheres. '
> 　　Do you mean that whether it is forward connections or backward connection
> has no influences on the model in that case?


I meant that if you have in your model left V1, right V1, left V2 and
right V2 it wouldn't make sense if the connection between V1 and V2 is
forward on the left and backward on the right. They should either both
be forward or both backward and in this particular example you can
find from the literature that forward is more likely.

Best,

Vladimir



> 　　Thanks a lot!
>        Haoran.
>
> --
> Haoran LI (MS)
> Brain Imaging Lab,
> Research Center for Learning Science,
> Southeast University
> 2 Si Pai Lou , Nanjing, 210096, P.R.China
>
> At 2011-03-10 18:16:32，"Vladimir Litvak" <[log in to unmask]> wrote:
>
>>Dear Haoran,
>>
>>2011/3/10 飞鸟 <[log in to unmask]>:
>>> Dear SPM's users,
>>> 　　When we construct a specific DCM model, we should specify different kinds
>>> of connections to those sources, but:
>>> 　　1. I wonder that whether the lateral connections only exist between the
>>> sources which are in the same level ? And, what does ' the same level ' mean
>>
>>Typically, lateral connections are for connecting bi-lateral symmetric
>>sources i.e. for modelling inter-hemispheric connections. There are
>>some other cases when they might be used, when it is known that two
>>areas are at the same level of cortical hierarchy. Usually such
>>information is available for the well-studied visual system.
>>
>>> ? Does it mean the two regions own the same function, such as the left
>>> visual region and the right visual region ?
>>
>>That's one example.
>>
>>> 　　2. Does forward connections only exist between the sources which are in
>>> the different level ? Such as low-level and the high-level ?
>>> 　　3. Do backward connections are always generate from high-level regions and
>>> terminate in those low-level regions ?
>>
>>Yes, that's the idea.
>>
>>> 　　All in a word, how could we specify the different connections among those
>>> sources ? Any help will be appreciated.
>>>
>>
>>A principled way of doing this is model comparison. Of course if you
>>want to try the 3 possible options for each connection, you'll
>>probably have very large number of possibilities. But typically you
>>would have information about at least some of the connections from the
>>literature. So if you are not sure about a particular connection, you
>>can try all the relevant options and do model comparison. Note that if
>>there is bi-lateral symmetry in your model it wouldn't make sense to
>>assign different types to symmetric connections in the two
>>hemispheres. This further constraints your options.
>>
>>Best,
>>
>>Vladimir
>>
>>> --
>>> Haoran LI (MS)
>>> Brain Imaging Lab,
>>> Research Center for Learning Science,
>>> Southeast University
>>> 2 Si Pai Lou , Nanjing, 210096, P.R.China
>>>
>>>
>
>
>