Monday, January 9, 2023

Study Finds No Robust Link Between Brain Structure and Personality

 "You are your brain" is the nonsense that neuroscientists keep pushing. But the data tells us otherwise. Let's consider the important question of whether there is any robust link between brain structure and personality. This was the topic of a 2019 paper entitled "Empirical examination of the replicability of associations between brain structure and psychological variables." Reviewing previous studies, the paper says, "Our results revealed that among healthy individuals 1) finding an association between performance at standard psychological tests and brain morphology is relatively unlikely 2) significant associations, found using an exploratory approach, have overestimated effect sizes and 3) can hardly be replicated in an independent sample."

We read the following about an analysis on a data repository that included hundreds of brain scans and personality tests on the people being scanned:

"Kharabian Masouleh et al. have now used brain scans from hundreds of healthy volunteers from an already available dataset to try to resolve the issue. The volunteers had previously completed several psychological tests. These measured cognitive and behavioral aspects such as attention, memory, anxiety and personality traits. Kharabian Masouleh et al. performed more than 10,000 analyzes on their dataset to look for relationships between brain structure and psychological traits. But the results revealed very few statistically significant relationships. Moreover, the relationships that were identified proved difficult to replicate in independent samples.'

Oops, that's pretty much just we would expect if your brain has nothing to do with your personality. So why do we see all these studies claiming to link brain structure with personality? The paper suggests two answers: (1) a use of way-too-small study group sizes, and (2) publication bias, under which negative results go unreported. 

It is well-known that these are two of the worst problems in experimental neuroscience. The paper suggests that "studies with 200 to 300 participants are still too small." But the typical neuroscience study does not even use 100 participants. Typically brain scan studies use fewer than twenty subjects, very often fewer than 15. With the type of way-too-small study group sizes used in most experimental neuroscience studies, what you are getting are probably merely false alarms and noise. 

Publication bias is the well-known fact that scientific journals prefer to publish positive results: studies that report some real effect rather than just a null result. This means that studies reporting null results tend to go unpublished. In many cases experimenters getting a null result will not even bother to write up the results in the form of a scientific paper. That is called "the file drawer effect." So, to give a hypothetical example, suppose that 9 out of 10 researchers trying to find a link between merry cheerfulness and brain structure find no significant link. It could be that only the 1 in 10 researchers reporting such a link get published. The result in the literature is then misleading. Maybe you'll do a Google search for "brain structure and cheerfulness" and find only papers reporting a link, even though 9 out of 10 researchers failed to find such a link. 

Referring to "structural brain behavior" or SBB associations, the authors report this:

"In particular we found a considerable number of SBB-associations that were counterintuitive in their directions (i.e., higher performance related to lower gray matter volume). Furthermore, subsampling revealed that for a given psychological score, negative correlations with GMV [gray matter volume] were as likely as positive correlations." 

Such results are not surprising if you make the correct assumption that the brain is not the source of the human mind. The study says this:

"Our empirical investigation of the replicability of SBB [structural brain behavior] in healthy adults showed that significant associations between psychological phenotype [personality] and GMV [gray matter volume] are not frequent when probing a range of psychometric variables with an exploratory approach. Where significant associations were found, these associations showed a poor replicability...When looking at a range of psychological variables, significant associations with GMV [gray matter volume] were very rare."

The authors suggest that there is a lot of misrepresentation going on in neuroscience papers, under which authors misstate the effect size they found. Speaking foolishly and ungrammatically, the authors state "brain structure can certainly not be questioned as the primary substrates of behavior," which contradicts the data they have reported in the paper. That silly statement notwithstanding, they have produced quite a good paper showing the lack of evidence showing a link between brain structure and personality.  

Vastly understating the gigantic dysfunction in experimental neuroscience, the authors state this:

"These findings suggest that samples consisting of ~200–300 participants have in reality still low power to identify reliable SBB-associations [structural brain behavior associations]among healthy participants. However, the sample size of SBB studies is usually substantially smaller. "

Substantially smaller? Figure 5 of the paper shows study group sizes used in neuroscience studies between 2001 and 2017, and shows an average of only about 15 participants. In rodent-based studies, we typically get "junk science" experiments involving fewer than 15 animals per study group.  We read, "Our study pointed out the need for big data samples to identify robust associations between psychological variables and brain structure, with sample size of at least several hundreds of participants."  

So what has been going on for decades is that our neuroscientists have not been using even one tenth of the study group sizes they needed for reliable results. 

No comments:

Post a Comment