Dear all,
Just a quick question for the neuroimaging community.... Having read
Russ Poldrack's recent paper in TICS on the problems inherent in
using neuroimaging data to support reverse inference (I agree with
Russ), I was wondering what the views about lesion and TMS data were?
Lesion and TMS data are often described as being capable of
demonstrating causality in cognitive neuroscience studies. So, given
evidence that a lesion (or disruption) to region X impairs, say,
phonological processing permanently (or temporarily), can this
evidence be used to support a reverse inference that region X was
probably engaged in similar processing during performance of a
different task with a putative phonological component?
For the record, the question is motivated by reviewer comments that
quite literally (and oddly) focussed on a discussion of mechanisms
potentially involved in a task and not directly under investigation
in one of our studies. Mechanisms we stated clearly that our data
could not adjudicate between. We cited complementary neuroimaging,
lesion and TMS data demonstrating the problem of attributing
attention or memory selectively to the IPL. The reviewers chose to
focus solely upon the neuroimaging data we cited and its inability to
support reverse inference. Not one mention of the lesion or TMS data.
Thoughts?
regards,
Greig
__
Dr Greig de Zubicaray
Senior Research Fellow
Centre for Magnetic Resonance
University of Queensland, QLD 4072, Australia
Tel: (+617) 3365 4100 (Office)
(+617) 3365 4250 (B106, Ritchie Building)
Fax: (+617) 3365 3833
Staff Page: http://www.cmr.uq.edu.au/CMR_Staff_Zubicaray.htm
fMRI Lab Page: http://www.cmr.uq.edu.au/Research_Imaging_fMRI.htm
|