Studying brain structure and function. A way to gain better understanding of social interaction?

Are we gaining a better understanding of social interaction through studying brain structure and function?

In the following text the above point will be carefully enlightened from different perspectives. After a definition of the research area of social neuroscience, its most prominent method, the functional magnetic resonance imaging – method (fMRI) will be discussed. Subsequently, problems and stumbling blocks in neuroscientific argumentation will be addressed by analyzing two studies using the fMRI method. It will be outlined, why the overoptimistic evaluation of the explanatory potential of neuroscientific results and the relating arrogant attitude of some researchers, feeds the critics. In the next step, a selection of important findings of the last 16 years will be used to prove the importance of social neuroscience for a better understanding of social interaction. This will include examining the structures of the ‘social brain’ (Kennedy & Adolphs, 2012). Eventually, there will be enough evidences collected to come to a profound evaluation of the claim in the title.

The emergence of the scientific fields of social behaviour and neuroscience is labeled social neuroscience. Its aspiration is to explain the neural processes which underlie social interaction. The notion of social neuroscience was first mentioned in 1992 by Caccioppo and Berntson, stressing the need of interdisciplinary, multilevel analysis for a deeper understanding of social behaviour (Adolphs, 2010). The upcoming of functional magnetic resonance imaging (fMRI) in the early 2000s acted as catalyst for the, by then, fast growing discipline.

What is social neuroscience exactly? Liebermann (2007) defines social neuroscience as the ‘study of the processes in the human brain that allow people to understand others, understand themselves, and navigate the social world effectively’. Thus, social neuroscience shares the same goals as the discipline of social psychology. The remarkable difference is the tools used. E.g. social neuroscience works with fMRI, positron emission tomography (PET) or transcranial magnetic stimulation (TMS); tools which allow tempting perspectives on brain function and structure and brought up insights of big importance.

On the other hand, the set of tools of social neuroscience carry huge risks regarding the feasible interpretation of the collected data. These risks, and the disadvantages of the methods used to scan the brain, serve as a fundamental issue in the scientific debate about the range in which social neuroscience can explain social interaction. That is to say, in the end the results are just as good as the methods used and there are certain boundaries to the tools used for brain studies. This is particularly important because the approach of studying behaviour by the use of fMRI to investigate cognitive processes, is threatening the established methods of research areas as social psychology, behavioural economics and even facets of political science (Adolphs, 2010). The aspiration of the fMRI method to replace traditional methods bears the risk of overrating its validity and sometimes results in vastly questionable research designs.

Therefore, in the following we will take a deeper look on the advantages and disadvantages of one of the most frequently used brain study methods. Thereafter, the risk of false inferences of collected data about brain structure and function will be illustrated on the basis of two studies. Studies using MRI-technology make it possible to nearly give a real-time view on the brain. MRI technology measures changes in blood flow, guided by the underlying assumption that an activated brain region requires more oxygen, which is transported through blood. Consequently, a higher blood flow in a specific region is assumed to be correlated with neural activity (Höhl, 2014). That the fMRI method is not measuring brain activity directly, is a common argument for sceptics of the method (Schmundt & Felix, 2012). Nevertheless, the advantages compared to the preceding CT Scan method are substantial: MRI offers a much higher spatial resolution without exposure to radiation as it is non-invasive. Compared to positron emission tomography, the fMRI method offers the opportunity to create event-related designs, which suit much more the demands of psychological experiments. On the negative side, the BOLD signal is not measuring neural activity in a direct way. Furthermore it lacks temporal resolution and does not provide information on causality of findings; a fact, that is often the reason for false conclusions from the data given.

According to that, there is good reason to take a deeper look on a study’s methodology, especially in this example: The study, conducted by renowned neuroscientists Iacoboni, Freedman and Kaplan from the University of California, was published in The New York Times with the title ‘This Is Your Brain on Politics’ (2007). The research team’s matter of interest was to examine the way US-citizens feel about the candidates aiming for the presidential office in 2008. Therefore, they asked 20 test persons into the fMRI scanner, where they were shown pictures of the candidates. Thereafter the conductor presented film clips to the test persons, showing the candidates speaking in public. The last trial contained the pictures from the first trial. The test persons were asked to do a before and after comparison (on a scale from 0-10; very unfavourable – very favourable). This more than questionable design resulted in following ‘findings’: Having seen pictures of Hillary Clinton, participants who evaluated her as unfavourable showed activity in the anterior cingulate gyrus, which is in the judgement of the researchers ‘an emotional center of the brain that is aroused when a person feels compelled to act in two different ways but must choose one’ (Iacobini et al., 2007). Pictures of Mitt Romney correlated with activity in the amygdale. The interpretation of the research team was that Mitt Romney causes fear in the voters. These are very strong and very specific inferences based on a simple brain scan.

The problem gets even clearer in another newspaper article. A psychologist from Princeton University claimed ‘Bikinis Make Men See Women as Objects’ (2009), as scans allegedly confirm. The lead researcher Susan Fiske claimed that when men are shown pictures of women in bikinis, the brain region associated with tool use is activated (Dell’Amore, 2009). Both of the studies, which were made publicly accessible to many people, fell into the trap of invalid reverse inference. In case of the latter study, the logic in the argumentation is as follows: if tools activate region XY and women in bikinis activate the same region XY, then men perceive women in bikinis as objects. This sort of argumentation is only acceptable, when the area affected is responsible for only one task. This is the case in the primary visual cortex. For every other brain region it is not acceptable. Thus, it is very important to set experiments in (social) neuroscience in a way that the concerned cognitive process is induced reliably and can be differentiated sufficiently from other cognitive processes (Höhl, 2014). Studies like above are a huge problem for the credibility of neuroscientific results. Laying aside poor designed studies, still, the problem of the mapping between brain structures and psychological processes is one of the big current debates in neuroscience nowadays (Adolphs, 2010). This makes it important to keep in mind that ‘no single brain structure maps any central psychological process’ (Adolphs, 2010).

Neuroscientific argumentation (especially using brain scans) is very appealing to the public. A study shows that images of brain scans have a vastly convincing influence on the public awareness of research results (McCabe & Castel, 2008). That means that people generally like to believe in what they can actually see on a scan. Indeed, brain images are much more accessible to a broader audience than dry evaluation tables. That makes the issue of poor neuroscientific designs even more significant. When it comes to reproducibility and validity of neuroscientific studies, findings paint a rather negative picture. For instance, a meta analysis of studies using the fMRI method estimates that around 10-20% of the brain activations reported in the examined studies are false positives (Wager, Lindquist & Kaplan, 2007). In another study, repeated fMRI measurements lead showed an overlap of only 30% in areas activated (Schmundt & Felix, 2012). Compensatively it must be said that this is not solely a problem related to the use of fMRI. Psychological studies in general have a massive problem with reproducibility due to the infamous ‘p-hacking’-practice, in order to raise the chances of the paper to be published. In a big reproducibility analysis, more than half of the examined studies (all from the year 2008) were not reproducible (Arts et al., 2016), regardless of the methods used. But related to the hype that the upcoming of brain scans caused, it is justified to question if the hustle was appropriate.

It can be clearly said that inadequate neuroscientific argumentation, the boundaries of the methodology and knowledge about the power of brain images on public awareness raise skepticism towards new findings and it is true that there are studies that do not improve our understanding of social interaction. But besides a careful eye on a study’s methodology and argumentation, can it really be said that studying brain structure and function is mostly irrelevant for a better understanding of social interaction?

A central point of critics on social neuroscience concerns neural plasticity. Critics argue that neuroscientists search for clearly definable regions for complex cognitive functions, although the brain’s plasticity is flexible enough to compensate for the dysfunction of even large regions (Schmundt & Felix, 2012). But most of the critics would not go so far to say that the neuroscientific perspective would be useless. What issues them more, is the arrogance of some brain researchers, who carry the explanatory potential of fMRI outcomes and their impact way too far (Schmundt & Felix, 2012). That is to say, the pessimistic position on neuroscientific findings is not least based on the grandiose promises some neuroscientists made in the early 2000s.

Some neuroscientists’ overestimation of the explanatory power of their methods raised a sort of sobering when evaluating what goals have really been realized over the years. This gets clearer, when taking a look at an article called Das Manifest, which was written by eleven neuroscientists and published in a German journal for Psychology and Neuroscience in 2004 (Elger et al., 2004). In this article the renowned experts analyzed the current state of neuroscience and gave a very optimistic future outlook on the field of neuroscience in ten years. The experts argued that there is progressing understanding on the level of molecular neuroscience and cognitive neuroscience, but the intermediate level, the level of the interaction of neural networks is largely unexplored (Elger et al., 2004). They predicted that in ten years there will be new knowledge about how neural circuits code, evaluate, store and read neural information. Following the authors, through this development there will be huge progressions made in the field of psychotropic drugs and it will even be possible to predict psychic anomalies and problematic development in the human brain. Revisiting the article ten years later the results are disappointing: There have been methodical developments but still there is a huge lack of knowledge about the connection of neural levels. Prediction of individual behaviour is unrealistic at the moment and there are no sensational innovations on the field of psychotropic drugs and still the root for many neurodegenerative and psychiatric illnesses are not completely enlightened. All in all, the major progressions in the field of neuroscience have been made in the technical area and the authors had to admit that they overestimated the explanatory potential of neuroscience (Höhl, 2014).

[...]