Feed on
Posts
Comments

Odysseus had himself tied to his ship’s mast and ordered his men to use bee wax in their ears and ignore his orders to free him. And so they did. Odysseus was curious about what the Sirens sounded like, and how the three bird-women seductresses were able to lure sailors with their enchanting music and voices to shipwreck on the rocky coast of their island.

In very much the same way, we should strap our minds to solid ground and avoid the lures of popular neuromarketing. Just as much as I am relieved to see basic cognitive neuroscience books being recommended as essential readings in neuromarketing, I am equally puzzled by recommendations of highly dubious books such as Lindstrøm’s “Buyology” and Pradeep’s “The Buying Brain” (book review upcoming soon to this blog).

Lindstrøm’s Buyology (just like his Brand Sense) is presented as a major book on neuromarketing. Building on proclaimed numerous brain scanning experiments, Lindstrøm presents the results of these studies, and how they provide new ways to understand marketing effects and consumer behaviour.

The problem is that the book is highly speculative and misinterprets and misrepresents neuroscience. Although it seems to be based on science, it is only indirectly so, and in a luring, invalid way. While Buyology has several problems of many kinds, I will here focus on one of the most telling ones.

Lindstrøm claims to have done over 1.000 brain scans. The vast majority of these were done in Richard Silberstein’s lab (about 90%) using his special take on EEG. The remaining 100 or so fMRI scans done were with Gemma Calvert at Neurosense. One of the problems with the book is that no information is given about the methods, the analyses or the detailed results from either of the studies. In other words, there are no ways we can check whether the results Lindstrøm claims to have had, are valid. To my knowledge, neither Calvert or Silberstein are happy with the way Lindstrøm has used the results and their names in Buyology. Calvert has stated that she will not use the results for publication.

Publication and documentation is key, especially when it comes to strong claims such as those made by Lindstrøm. Lindstrøm has responded to my criticism here in Denmark, by claiming that the results should be out soon. That was more than two years ago.

Most importantly, Lindstrøm makes several logical errors. For example, in the study of smokers, he claims that the warnings put on cigarette packages had the opposite effect of what was intended (to scare smokers). His reasoning is like this: the warnings produced stronger activation of the ventral striatum (nucleus accumbens). Since other studies have demonstrated that this structure is involved in the expectancy of reward, it should be straightforward to conclude that the activation seen in smokers means that the warning signs made them expect a reward. In other words: warning signs make smokers think (positively) of smoking…right?

This is a prime example of a logical fallacy called “reverse inference“. Basically it means that explaining some brain activation with findings from other studies is wrong. In simplified terms, it goes like this:

Rule: When it rains, I open my umbrella
Valid use of rule:
- Observation: it rains — we can conclude that I will open my umbrella

Invalid use of rule:
- Observation: I open my umbrella — it rains

The invalid conclusion above is due to the fact that I can open my umbrella for other reasons. I could, for example, want to open it to let it dry.
If the rule was that “when it rains, and only when it rains, I will open my umbrella”, we would accept the argument. My opening the umbrella could only come to be if and only if it rained. We’ll get back to this.

In the same way, Lindstrøm makes the logical fallacy of reverse inference, in this way:
Rule: when the organism expects a reward, nucleus accumbens is activated
- Observation: nucleus accumbens is activated, hence we can conclude that the organism expects a reward

The problem is, the nucleus accumbens could be activated for other reasons. Indeed, this structure is involved in many other functions, and not particularly to reward expectancy. In fact, in a study by Levita et al (2008) it was reported that the nucleus accumbens was activated when subjects expected a punishment. This makes the argument made by Lindstrøm fall apart. We don’t really know what his smokers did and why their brain responded in the way they did.

Assigning strict brain structure to functional neuroimaging results is hard. Suffice to say that most researchers are much more careful with assigning strict labels to such small structures as the nucleus accumbens. Due to the data preprocessing steps in fMRI and the large individual variability in brain morphology, one should rather prefer to use the vague term “ventral striatum”.

Indeed, if top researchers and journals make categorical mistakes in pointing at key brain structures, why should be trust an undocumented claim by a non-neuroscientist such as Lindstrøm?

As I will report soon at this spot, Pradeep does not fare much better. His book “The Buying Brain” is one long presentation of  how immensely better we can be at measuring communication effects, preferences and purchase intentions, and how good his own company is in this respect. But, again, there is no real documentation for this claim. There are developed tools and tests from the NeuroFocus group, but if there is no external validation of such tests, why spend thousands of dollars on this?

Is it, then, not time for us to take an odyssean approach to neuromarketing? Sirens abound, leaving no other options but to tie ourselves to the mast and aim for the horizon.

-Thomas

One Response to “An odyssean take on the sirens of neuromarketing”

  1. [...] let me be right up front and remind you that the explanation offered by Martin Lindström is really not tenable. Rather, it is a prime [...]

Leave a Reply

Sharing Buttons by Linksku