Discussion in 'Other Health News and Research' started by MeSci, May 5, 2015.
This paper looks interesting.
Full text at link.
Does this indicate a genetic correlation -- mothers genetically susceptible to depression pass on something that shows up as different amygdala functional connectivity in their infants? Or is something biochemical active in the mother's body chemistry altering the infants' brains? Were the mothers all medicated for depression, or all not medicated? That would make a big difference in the interpretation of the results. Would medicating mothers so that the symptomology (which is what the study measured) disappeared eliminate the alteration in infant amygdala functional connectivity? Is it about the symptoms, the inherited structural differences in the brain?
The paper seems to be claiming that the issue is intrauterine environment, but I don't see that they controlled for genetic transmission of structural aspects. If the alteration of amygdala functional connectivity is, in fact, only correlated to depression symptoms, this would make an excellent argument for high attention to proper medication for depression in pregnant women. I don't see that this paper proves that the alteration is strictly the result of intrauterine environment. There are too many confounding factors that they haven't controlled for, as far as I saw. It looks like all they checked was 1) did mom have depression symptoms based on a questionnaire (not diagnosis by a qualified psychiatrist), and 2) did the MRI on the infants show changes. They show correlation between those two things, but that's not all that helpful.
Correlation of a physical result to the answers on a questionnaire is only the most general of information. What is the physical result actually correlated to -- the mother's emotional state that day, her overall mental health, the emotional impact of pregnancy hormones, her genetic predisposition to mental health problems, the effectiveness of her medication for depression (if any)? The correlation is very unclear.
The authors' interpretation seems based on their belief about the nature of the correlation, not any evidence. And they seem to be completely missing the major scientific fact that correlation is not causation.
This is an interesting paper, but it has way too many holes, imo. Like most psych papers, it extrapolates far beyond it's data.
No, I don't think so. You didn't provide evidence that mom's symptoms alter amygdala connectivity, you provided evidence that how mom answers a questionnaire on a given day is correlated to infant amygdala functional connectivity. We still don't know exactly what the alteration is correlated to -- there are too many confounding factors. And we certainly can't draw conclusions about causation from the available data.
The rest of their conclusion, however, seems more sound:
I would say they did show something associated with maternal mental health during pregnancy is correlated to alteration in infant amygdata connectivity.
Yes - I only had a very quick look before posting, and when I saw reference to MRI I assumed that it was a rigorous study. It seems strange that they used MRI for the children but no objective measurements for the mothers. They could have used biochemical correlates of depression such as interferon-gamma or cortisol.
There are no accepted biomarkers for depression, so no, they couldn't have used objective measures.
Thinking about this, how many illnesses do have unequivocal, consistent biomarkers? A number of studies have found associations between depression and cortisol and interferon-gamma, so I think it would be useful to know whether these or other biomarkers in mothers were associated with the brain abnormalities in the offspring.
If you use those markers, all you will have shown is a correlation between them and amygdala connectivity in offspring, not the latter and depression. That's no better than the questionnaires.
If you are going to critique a study, you need the background research to do so. You need to know the sensitivity and specificity of the rating scales, versus that of the physiological markers you mention. My guess is that the physiological markers are no better than the rating scales, or they would have been widely used by now.
Maybe we'd do better to stop defining depression as a specific condition to use in studies like this, and more as a set of symptoms with more-or-less consistent biochemical correlates, and just use objective correlates/markers when analysing findings such as brain abnormalities in offspring. Depression is so heterogeneous and with a range of causes.
I am tending more and more towards @alex3619's view that supposedly-psychiatric illnesses would be better studied and treated as physiological conditions, which can be studied objectively. Then we would have more homogeneous study groups, with more reliable and meaningful results.
Symptoms such as depressed mood could just be treated as clues to what to test for, along with other symptoms.
Well, everyone would like that, but depression cannot be treated as a physiological condition until we know its physiological correlates. It might possibly be so that depression, as defined in the DSM, is too heterogeneous for this to become a reality. I personally would like to see the DSM ditched, and the RDoC used in research instead.
I agree that the divisions at the above link can be of value. It is crucial to understand the reasons for depression, anxiety, fear, etc. If they are due to a real or imminent threat they should be treated differently from persistent negative responses to a threat that no longer exists, or perhaps never did.
But what I mean by 'treated as clues' is 'treated' as in the above sentences - things like depression, anxiety/fear or fatigue should be treated (regarded/classified, etc.) as symptoms/clues to an illness, as are/should breathlessness, polyuria, etc., not as specific illnesses in their own right. You can't 'treat' them - in the therapeutic sense - until you know the cause(s) and the physiological correlates, which will be different in different cases.
You can also try a Google Site Search
Separate names with a comma.