Dear Vamshi,
Vamshi Krishna wrote:
> Hi Penny,
> I read your paper. And i have basic doubt which is not clarifying my self.
> How Causal models==> Regression models
> and converse is not true.
> So what are the examples that i can consider.
> what is the best way to desctrize these causals and regression.
> Thanks
The causal models you refer to, eg Dynamic Causal Models (DCM),
are defined by a set of differential equations
dz_i/dt=f(z)
where the neuronal activity in area, z_i, changes as a function
of activity in other regions. Simultaneously, activity in area j will
also be changed. So this is like standard differential equation models
describing neural networks (eg. integrate and fire/Hodgkin-Huxley nets)
with instantaneous change and mutual causality.
In standard regression models an independent variables x are deemed to
cause a change in a dependent variable y. Additionally, this process is
susceptible to additive noise. These relations are not usually described
as causal but one can make inferences about partial correlation etc.
The differential equations in DCMs are discretized using standard
numerical integration methods eg. Runge Kutta
(see eg. http://en.wikipedia.org/wiki/Numerical_integration).
Best wishes,
Will.
>
> reagrds
> Vamshi
>
> ________________________________
>
> From: Will Penny [mailto:[log in to unmask]]
> Sent: Tue 3/25/2008 4:07 AM
> To: Vamshi Krishna
> Subject: Re: requesting for clarification of concepts
>
>
>
> Hello again,
>
> Perhaps you could read this paper:
>
> http://www.fil.ion.ucl.ac.uk/~wpenny/publications/integration.pdf
>
> Best,
>
> Will.
>
> Vamshi Krishna wrote:
>> Dear Sir, Thanks. I understood. Bute i like know in general differnce
>> between Casual and regression. Thanks regards Vamshi
>>
>> ________________________________
>>
>> From: Will Penny [mailto:[log in to unmask]] Sent: Tue
>> 3/25/2008 3:49 AM To: Vamshi Krishna Subject: Re: requesting for
>> clarification of concepts
>>
>>
>>
>>
>> Dear Vamshi,
>>
>> Regression models look for instantaneous correlations or partial
>> correlations between variables.
>>
>> The 'Dynamic Causal Models' I refer to use differential equations to
>> model the dynamic relationships between variables (neuronal
>> activity).
>>
>> Very best wishes,
>>
>> Will.
>>
>> Vamshi Krishna wrote:
>>> Dear Sir, I read your paper 'Comparing causal models' which is
>>> interesting. Could you explain the major difference between Causal
>>> models and regression models, since i am not able to get the
>>> information from Text book/Internet. Thanks Regards
>>>
>>>
>>> T.Vamshi Krishna
>>>
>>> Research Analyst
>>>
>>> VisualBeam Technology
>>>
>>> <http://owa.steelwedge.com/exchweb/bin/redir.asp?URL=http://www.visualbeam.com/>
>>>
>>>
>>>
>>> 1st Floor Uma Hyderabad House,
>>>
>>> Raj Bhavan Road, Somajiguda
>>>
>>> Hyderabad,AP, INDIA
>>>
>>>
>>>
>>>
>>>
>>>
>> -- William D. Penny Wellcome Trust Centre for Neuroimaging University
>> College London 12 Queen Square London WC1N 3BG
>>
>> Tel: 020 7833 7475 FAX: 020 7813 1420 Email: [log in to unmask]
>> URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
>>
>>
>>
>>
>>
>
> --
> William D. Penny
> Wellcome Trust Centre for Neuroimaging
> University College London
> 12 Queen Square
> London WC1N 3BG
>
> Tel: 020 7833 7475
> FAX: 020 7813 1420
> Email: [log in to unmask]
> URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
>
>
>
>
>
--
William D. Penny
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
Tel: 020 7833 7475
FAX: 020 7813 1420
Email: [log in to unmask]
URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
|