There is an appalling lack of quality in the vast majority of brain imaging studies trying to find correlations between brain characteristics and either mental activity or mental characteristics. The great majority of such studies use way-too-small study group sizes such as only 15 or 20. A press release from the University of Minnesota Twin Cities announced results which indicate that such small-sample correlation-seeking brain imaging experiments are utterly unreliable. The headline of the press release is "Brain studies show thousands of participants are needed for accurate results." The abstract of the paper in the science journal Nature can be read here. The paper is entitled, "Reproducible brain-wide association studies require thousands of individuals."
Recently an attempt was made to find correlations between brain imaging and psychological characteristics. Rather than following the usual nonsense of getting about 15 or 20 new subjects to have their brain scanned, the study took the much wiser approach of using brain scan data from 23,810 subjects already collected by the UK Biobank. I can't over-emphasize how important it is for correlation-seeking experimental studies to use existing brain scan data whenever it exists and is sufficient. When an experimental study does unnecessary fresh new brain scans of a few dozen subjects (particularly using the more powerful 7T scanners) rather than using existing brain scan data already gathered in much greater quantities, this amounts to both poor science practice and also subjecting humans to needless risk without any medical justification. See my post here for the type of risks involved.
The authors of a new study really "shook the trees" and "went the extra mile" looking for some correlation between brain states and psychological states. The study is entitled "The legibility of the imaged human brain." We read this:
"Across 23810 unique participants from UK Biobank, we systematically evaluate the predictability of 25 individual biological characteristics, from all available combinations of structural and functional neuroimaging data. Over 4526 GPU hours of computation, we train, optimize, and evaluate out-of-sample 700 individual predictive models, including multilayer perceptrons of demographic, psychological, serological, chronic morbidity, and functional connectivity characteristics, and both uni- and multi-modal 3D convolutional neural network models of macro- and micro-structural brain imaging."
When the brain scan data was collected by the UK Biobank, which occurred before this study began, the subjects also had tests done or interviews to determine psychological traits which have these names in that database: mood swings, miserableness, irritability, sensitivity, fed-up, nervous, anxious, tense, worry, lonely, guilty. The authors of the new study made the most elaborate efforts to "slice and dice" the data, trying to find some correlation between brain states of their 23,000 subjects and psychological states of those subjects.
Since they tried 700 different models, we are reminded of the old adage "keep torturing the data sufficiently, and it will confess to anything." But in this case there was no confession. The authors of the new study were unable to find any real correlation between brain states and mental states. A section of their paper referring to brain imaging (such as fMRI scans) is entitled "Psychological characteristics are poorly predicted by imaging."
The authors found that they could predict age and sex well from brain imaging data, and also handedness (whether one is right-handed or left-handed). But it was a totally different story for psychological characteristics. We read this: "The addition of any neuroimaging, whether structural or functional, generally offered no material benefit" in being able to predict a psychological characteristic. The authors state this: "Our analysis shows that whereas constitutional characteristics—age, sex, and weight—are highly predictable from neuroimaging, psychology, chronic illness, and serological characteristics are not."
You could fairly summarize the results by saying this: brains don't look any different when someone is moody, miserable, irritable, sensitive, fed-up, nervous, anxious, tense, worrying, lonely or guilty. The result of the authors is consistent with the claim that the brain is not the source of the human mind.
No comments:
Post a Comment