Progress in cognitive neuroscience has been undermined by an over-reliance on brain imaging techniques

You’re sitting in a packed courtroom. It is the last day of an intense and sensational trial. Opening statements and cross examinations have been fired back and forth like a game of blitz table tennis; witnesses have come and gone. Now the moment you have all been waiting for is here – the accused takes the stand.

The defence counsel starts arguing that the accused is not guilty of his crime. Why?, he asks rhetorically. His answer is to unveil a large photographic exhibit to the jury. A photo of a brain, the accused’s brain, complete with coloured ‘hot spots’. What does it all mean, you wonder. The defence lawyer starts explaining. You hear descriptives such as ‘receptor abnormalities’, ‘impaired cognitive functioning’, ‘neural correlates’ and ‘the prefrontal cortex – the home of violence’ (Reeves et al, 2003).

In the next breath, he points to the schizophrenic tendencies in his client’s brain – this is supposedly where the voices ‘told him to do it’. His conclusion is that his client is ‘abnormal’ and as a result he did not ‘know’ what he was doing (Compton, 2010). As the dust settles in the courtroom, the jury, consisting of ordinary men and women, sit silent. Some look impressed, others struggle to connect the dots. But there is no doubt that a picture has painted which a thousand words could not. The visual ‘evidence’ seems clear.

Or does it? What has been presented and explained? How has this brain evidence been put together? Does the exhibit show the brain as it really is? Has the defence skillfully skipped over some important facts? Is there such a thing as ‘schizophrenic centre’ in the brain? What can brain imaging reveal that other techniques cannot?

‘Brain imaging’ and ‘neuroscience’ are two very ‘in’ buzzwords at the moment. There is no doubt that cognitive neuroscience – the study of how thoughts occur in the brain – and brain imaging – the use of technology to directly or indirectly study the brain as it processes or responds to information – are both fascinating and potentially revolutionary in what they can both tell us about how we think and how the brain is involved in this process (Mathias, 1996). But caution does need to be exercised, as with anything still in the beginning stages. Even though now widely used, brain imaging is far from being an infallible or precise science (Compton, 2010). Its songs of praise need to be balanced with its reports of error and subjectivity (Compton, 2010). Moreover, researchers have to get the basics right, such as developing a cognitive ontology – where mental processes are explicitly described, as in the case of the Cognitive Atlas Project (Poldrack, 2010; Gonsalves & Cohen, 2010).

The media often reports of scientific breakthroughs and discoveries using brain imaging, as well as how studies have found links between behaviours and brain structures, e.g. “region X makes us do Y” or “we have found the ‘XYZ’ centre” (Beck, 2010). Unfortunately, the media is sometimes guilty of misrepresenting brain imaging data and of oversimplifying complex psychological theories and layers to mere ‘neurons and brain tissue’ (Decety & Cacioppo, 2010). As with the defence lawyer in the opening paragraph, many are using brain imaging technology to make sense of complex psychological processes (Beck, 2010). But just because we can now see the brain in action, doesn’t mean the instrument we are using automatically has all the answers. Brain imaging is a tool, a very powerful one, but a tool nonetheless (Decety & Cacioppo, 2010). It is not a conscious being with insight, experience, judgement and acquired knowledge. The tool still requires a good operator. Giving the impression that brain imaging is a source of wisdom or a holy grail, is something to be avoided (Decety & Cacioppo, 2010; Diener, 2010). It is the same as a telescope, which enables us to observe the universe. The telescope does not hold the answers to the secrets of how stars are born, why they are born, and so on. But it brings us closer to these answers, by looking at old questions in new ways (Decety & Cacioppo, 2010).

What is clear from the opening paragraph is that brain imaging has cultivated a kind of cult status (Compton, 2010). In the same way as we are impressed by photos of galaxies and nebulae, the layperson is impressed by scientific language, and even more so, when said language is combined with a brain image (Beck, 2010). However, while a photo of the brain may ‘show’ a lot, it doesn’t explain the cause or concept behind the activity recorded, e.g. an image of someone experiencing grief, doesn’t tell us what grief is or why that person may be experiencing grief (Beck, 2010).

What a brain image portrays is not a direct mirror reflection of the brain but rather a composite image. It shows the ratio between deoxygenated and oxygenated hemoglobin in the blood and the measurement of contrasting activities performed under observation. The image is not the result of a single task or action (Beck, 2010). Therefore what we know is not black and white, it involves a lot of fuzziness. It is actually important to know what the contrasting activities are in order to correctly interpret neural reactions (Beck, 2010). If a man is shown a series of photos, including one of a good-looking woman, and his brain lights up, we need to know what the other photos are of, e.g. dogs, his wife, old grannies or other good-looking women (Beck, 2010).

Furthermore, a brain image is a reconstructed product (Reeves et al, 2003; Sarter, Berntson & Cacioppo, 1996): when two tasks are contrasted during a study, two brain images are acquired; the measured differences between the two images are standardised and this ‘tweeking’ can produce inaccuracy; signals emitted by the brain can also mean many things, and; sometimes the signals are not strong enough to determine if they are ‘noise’ or really ‘something’ (Sarter, Berntson & Cacioppo, 1996). Interpreting activity and signals is a very fine process, and researchers need to take many things into account. It is not a simple case of saying region X produces process Y (Sarter, Berntson & Cacioppo, 1996). Even if region X is active, it doesn’t necessarily mean that it is involved in process Y (Sarter, Berntson & Cacioppo, 1996; Gonsalves & Cohen, 2010). Region X can also be involved in process Z but can light up at the same time as process Y (Gonsalves & Cohen, 2010). Likewise, if activity is absent during an task, it cannot be concluded that there is no activity in region X (Bec, 2010; Sarter, Berntson & Cacioppo, 1996). Instead, it may mean that there was no significant difference between regions X, Y and Z (Beck, 2010).

Activity in general, whether increased or decreased, can be caused by a variety of internal or external circumstances (Beck, 2010; Compton, 2010). Not only can psychological processes activate brain regions, but also body movement, alcohol, drugs, psychotherapeutic treatments and even coffee (Waddle, 1996; Reeves et al, 2003).

It is too simplistic to start mapping cognitive processes with brain regions and say we now know how the brain works (Shimamura, 2010; Sarter, Berntson & Cacioppo, 1996). Researchers may search for ‘where’ in the brain (also known as localisation) cognitive processes are ‘housed’ (Shimamura, 2010), but this ‘where’ approach is seen as ‘deductively’ invalid (Gonsalves & Cohen, 2010). Psychological processes are complex and often involve many subprocesses, so it is unlikely that only one brain region is involved (Gonsalves & Cohen, 2010). For example, the amygdala, anterior cingulate, posterior parietal cortex or hippocampus are not the only structures involved in fear, conflict, episodic retrieval or memory respectively – even if brain imaging has helped us to localise their activity (Shimamura, 2010). They are part of a network and communication of interactions (Poldrack, 2010; FMRIB, 2011). By sticking with the ‘where’ approach, cognitive neuroscience risks becoming the ‘new phrenology’, where a person’s typology is characterised by the structure of their skull (Shimamura, 2010; Poldrack, 2010). Furthermore, attaching a biological basis to everything, risks overlooking how psychology plays an important role in all of this (Shimamura, 2010). People may have a genetic or hereditary tendency to put on weight, but they also develop patterns of over-eating and over-drinking out of their own free will (Shimamura, 2010).

The good news is that many researchers are moving away from the ‘where’ approach (Sarter, Berntson & Cacioppo, 1996; Gonsalves & Cohen, 2010; Shimamura, 2010). The bigger picture they say, is about understanding the ‘what’ and ‘how’ things work (Shimamura, 2010), e.g. how do neural structures play a role in the interaction of whole brain functions (Shimamura, 2010). After all, how much more do we know about what or how a moral decision or emotional process is/works by simply seeing a brain region lit up (Beck)? Not a lot. Imaging does not explain behaviour, it only shows it (Beck, 2010). Finding the mechanisms and networks involved is more valuable than finding the location itself (Shimamura, 2010). For some (Poldrack, 2010) the goal is about predicting involvement in a cognitive process by the pattern of activity in a brain region. This is called ‘selective association’, and shows that researchers cannot make just ‘any’ kind of association or correlation between mind and brain activity (Poldrack, 2010).

To illustrate this point, the following are some examples: Shimamura (2010) uses the heart, an instrumental organ in the circulatory system, but not the only one involved. Poldrack (2010) uses the example of a knife, technically-speaking a cutting tool, but it can also be used for screwing or removing flesh. Decety and Cacioppo (2010) use Lego to show that pieces are like the outputs of brain structures, and the various configurations of the Lego pieces are the different psychological processes. Lastly, Sarter, Berntson & Cacioppo (1996) use the heater-temperature analogy to show top-down and bottom-up processes also influence activity.

However, it may be that discovering the ‘how’ of the brain is something that cannot be ‘sufficiently’ answered by brain imaging (Decety & Cacioppo, 2010). Some say this is because psychological functions cannot be localised in the brain because they are products of neuronal activity and not the neuron itself (Sarter, Bernston & Cacioppo, 1996). Determining what can be localised was started back in 1812, with Legallois’ first brain localisation experiments (Sarter, Bernston & Cacioppo, 1996). The conclusion was that sensory and motor functions can be localised, but not the more complex psychological ones (Sarter, Bernston & Cacioppo, 1996).

There are advantages of saying that region X is correlated with function/process Y. Drawing such a conclusion often acts as a stepping stone to further research and discovery (Beck, 2010). For example, when people experience grief, activity in the nucleus accumbens is recorded (Beck, 2010). As this region is associated with pleasure and reward, it may be concluded that people derive some kind of pay off from their grief (Beck, 2010).

Based on what has so far been discussed, one could conclude that the approaches and perspectives held on brain imaging are without a doubt divided. On the one hand, brain imaging is seen as a concrete science, because it produces something that we can visually see and understand (Diener, 2010). In contrast, psychology is seen as an abstract and subjective science, with a lot of divisive theories and challenges (Diener, 2010). In other words, psychology is not as ‘easy’ as brain imaging. Essentially, we have been able to open up the brain, and bring something previously hidden out into the open. This certainly gives the brain a special aura. In contrast, thoughts, emotions and psychological problems, still remain hidden in the realm of the mind. But what if a machine could show us what a thought looks like, could track and trace a thought as it moved through the brain, perhaps then psychology and the mind would also acquire a new status too (Compton, 2010)?

This argument highlights the ancient debate of mind vs brain, and is certainly relevant here in the case of brain imaging. Are mental processes created by the brain or do mental processes control the brain? Second to this is whether activity in the brain is based on centristic or holistic patterns (Sarter, Berntson & Cacioppo, 1996). Equipotentiality means that cognitive functions activate the whole brain, and not just region X (Sarter, Berntson & Cacioppo, 1996). Whichever you believe, they all need to satisfactorily answer the critical questions of validity, credibility, reliability and adequacy (Diener, 2010).

There is no question that brain imaging definitely has a place and value. In the past, before it was used, concepts, theories and notions about the brain were pieced together by studying brain damaged patients (Gonsalves & Cohen, 2010). The downside of this was that the brain was only studied in its injured form. Today, with modern brain imaging technology, healthy brains can be seen in action (Gonsalves & Cohen, 2010).

Brain imaging has also played a major role in understanding brain functioning, determining what something is in the brain, analysing cognitive processes and organising the brain and mind (Decety & Cacioppo, 2010; Gonsalves & Cohen, 2010). In addition, it has done this in a non-invasive way (Decety & Cacioppo, 2010). Brain imaging has also been involved in myth-busting, such as determining that the prefrontal cortex and posterior parietal cortex – and not just the hippocampus – are as equally important as each other in memory processing (Gonsalves & Cohen, 2010). Or that blood flow to the brain during goal-related processing activity in the prefrontal cortex may mean a combination of things: such as motor, memory, and choice responses (Waddle, 1996).

However, brain imaging has also had its share of problems. One of its weaknesses is that it is only as good as its technology. To overcome its technical limitations, intimate knowledge of the capacities and specialisations of available technologies, and how best to combine them, is critical. For instance, the use of EEG and fMRI has been suggested in bridging temporal and spatial issues in brain imaging (FMRIB, 2011). This combination helps to record the extremely fast activity of neurons (FMRIB, 2011). Using PET would be a bad choice as it cannot record temporal activity (FMRIB, 2011). Low resolution is also a problem, and prevents determining what the real cause of activity is in the brain (Waddle, 1996). Nerve cells need to be studied on a fine-grain level (Waddle, 1996). Another point is that brain imaging compares the brain to a norm, and so the norms have to be reliable (Compton, 2010).

Despite its hits and misses, brain imaging should not be used as a way of pushing psychology aside. Even though we have not reached the stage where we can record a thought, hold it or show it to the world, there is no reason why we should discard psychology in favour of brain imaging (Miller, 2010). Miller (2010) argues that we lack the knowledge about causal relationships between biological and psychological events, and without these we have no way of being able to use brain imaging properly. For example, contrary to our opening paragraph example, Miller (2010) states that there is no ‘schizophrenic centre’ in the brain, its symptoms are psychologically-oriented. He also believes that we must be careful with our language: the brain does not ’cause’ psychological events; rather they are ‘associated’ with one another (Miller, 2010).

There are a few recommendations that can be taken from this review into brain imaging. Success in the future depends on a cooperation between psychology and biology; brain imaging should be seen and used as a bridge between both sciences (Shimamura, 2010), rather than a be-all and end-all. A lot of work needs to be done first in psychology (developing a cognitive ontology), before mapping can be applied to the brain. A solid, comprehensive and sophisticated understanding of psychology and other sources and tools also has to be attained (Decety & Cacioppo, 2010).

Success in using brain imaging is also dependent on the design of studies and the frameworks that guide their interpretation and inference (Gonsalves & Cohen, 2010). This is because in real life, more than two functions or conditions are contrasted at the same time (Compton, 2010). Life is a little more spontaneous and unpredictable than that.

Integrating and using other approaches (Diener, 2010), such as behavioural, experimental, physiological and psychological ones – otherwise known as the Golden Triangle – is also critical (Decety & Cacioppo, 2010). Lastly, researchers will have to use a very clear and explicit language stating what brain imaging can and cannot do (Compton, 2010).

In sum, a map is not the territory that it represents (Miller, 2010).


  • Beck, D. M. (2010). The Appeal of the Brain in the Popular Press. Perspectives on Psychological Science 2010 5(6): 762. doi: 10.1177/1745691610388779
  • Compton, E. S. (2010). Not Guilty By Reason of Neuroimaging: the Need for Cautionary Jury Instructions for Neuroscience Evidence in Criminal Trials. Vanderbilt Journal of Entertainment and Technology Law 12(2): 333. Available online at law/archive/download.aspx?id=5095. Retrieved April 14, 2011.
  • Decety, J., Cacioppo, J. (2010). Frontiers in Human Neuroscience: The Golden Triangle and Beyond. Perspectives on Psychological Science 2010 5(6): 767. doi: 10.1177/1745691610388780
  • Devlin, H. (2011). Common criticisms of FMRI & counter-arguments. FMRIB Centre, Department of Clinical Neurology, University of Oxford. Available online at Retrieved April 14, 2011.
  • Diener, E. (2010). Neuroimaging: Voodoo, New Phrenology, or Scientific Breakthrough? Introduction to Special Section on fMRI. Perspectives on Psychological Science 2010 5(6): 714. doi: 10.1177/1745691610388773
  • Gonsalves, B. D., Cohen, N. J. (2010). Brain Imaging, Cognitive Processes and Brain Networks. Perspectives on Psychological Science 2010 5(6): 744. doi: 10.1177/1745691610388776
  • Mathias, R. (1996). The Basics of Brain Imaging. Nida Notes 11(5). Available online at Retrieved April 14, 2011.
  • Miller, G. (2010). Mistreating Psychology in the Decades of the Brain. Perspectives on Psychological Science 2010 5(6): 716. doi: 10.1177/1745691610388774
  • Poldrack, R. A. (2010). Mapping Mental Function to Brain Structure: How Can Cognitive Neuroimaging Succeed? Perspectives on Psychological Science 2010 5(6): 753-716. doi: 10.1177/1745691610388777
  • Reeves, D., Mills, M. J., Billick, S. B., & Brodie, J. D. (2003). Limitations of Brain Imaging in Forensic Psychiatry. Journal of the American Academy of Psychiatry and the Law 31: 89-96.
  • Sarter, M., Berntson, G. G., & Cacioppo, J. T. (1996). Brain Imaging and Cognitive Neuroscience: Toward Strong Inference in Attributing Function to Structure. American Psychologist 51(1): 13-21.
  • Shimamura, A. P. (2010). Bridging Psychological and Biological Science: The Good, Bad and Ugly. Perspectives on Psychological Science 2010 5(6): 772. doi: 10.1177/1745691610388781
  • Waddle, D. (1999). Brain Imaging: Is It the New Phrenology?. National Undergraduate Research Clearinghouse, 2. Available online at Retrieved April 14, 2011.

Posted by Jasmin on Jul 17, 2011

← Professional graphology Blog Index