Two features of the study seem fishy. The first is the manual
designations of individual brain areas. Although they studied the
reliability of the designations and
appeared very careful, there must be some persistent error in defining
the brain areas. The second is that people making these designations
may have a
systematic bias in the brain areas they think are associated with
various disorders. This is different than being blind to the subject
diagnosis. We did
something similar in a study that required designating the hippocampus.
I noticed that each person on the team had a different idea of where the
hippocampus ended and the parahippocampus began. There were other
uncertain areas. Each rater just had to make a decision. However, if
the raters in this study
systematically sampled more of the hippocampus for the depression group
and less for the schizophrenia group then the classifications might
represent
such a systematic difference. Since they did not use a normalized
method common across all the subjects, this could have happened and they
should have
examined it.
The second is the high rate of classification compared to the
reliabilities of the diagnostic methods. The SCID reliability varies
from approx .6 to .9, depending
on the diagnosis. Presumably their classification rates should be
lower, given the error in making anatomical designations and
measurements and
the error in making diagnoses. The extremely high rates of
classification suggest that some systematic bias is linking the brain
measurements to the
diagnostic clusters.
Finally, these overly empirical and descriptive studies do not enlighten
us concerning how the brain mediates these disorders.
Mike Williams
Assaf, B., Mohamed, F. B., Abou-Khaled, K., Williams, J. M., Yazeji, M.,
Haselgrove, J. & Faro, S. (2003). Diffusion Tensor Imaging of the
Hippocampal formation in temporal lobe epilepsy. American Journal of
Neuroradiology, 24, 1857-1862.
On 12/8/12 11:00 PM, Teaching in the Psychological Sciences (TIPS)
digest wrote:
Subject: MRI Diagnosis
From: don allen<dap...@shaw.ca>
Date: Sat, 8 Dec 2012 10:15:55 -0700 (MST)
X-Message-Number: 3
I just read the following study in PLOS one titled: Anatomical Brain Images
Alone Can Accurately Diagnose Chronic Neuropsychiatric Illnesses.
It can be found here:
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0050698
The claims of the study seem impressive:
"In MRI datasets from persons with Attention-Deficit/Hyperactivity Disorder,
Schizophrenia, Tourette Syndrome, Bipolar Disorder, or persons at high or low familial
risk for Major Depressive Disorder, our method discriminated with high specificity and
nearly perfect sensitivity the brains of persons who had one specific neuropsychiatric
disorder from the brains of healthy participants and the brains of persons who had a
different neuropsychiatric disorder."
The research design seemed to be adequate (at least to me) but I don't have
enough detailed information about MRI to know whether this is a really
important breakthrough or just another soon-to-be-forgotten study. The fact
that it was published in PLOS one rather that Science or Nature makes me
suspect the latter. Would anyone with more expertise in interpreting MRI data
like to provide some comments on the study?
Thanks,
-Don.
---
You are currently subscribed to tips as: arch...@jab.org.
To unsubscribe click here:
http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=22259
or send a blank email to
leave-22259-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
