Wednesday, February 24, 2010

Neuromarketing the Neurology of Facebook

Facebook brain activation - visual cortex (photo by Marc Van Rymenant)

First, we had the fictitious Neurology of Twitter study, sponsored by The Neurocritic. Now it appears there's been an actual (unpublished) fMRI study of viewing Facebook pages, conducted by Netway, a neuromarketing firm.
A global premiere: travel in the mind of Facebook users

In the digital world, where business results are ever more top of the list, user experience is one of the key factors of success.

. . .

In plain English: we can see which zones of the brain are activated when a user is performing a task.

If we can identify the activated zones of the brain, we also know the answer to the following questions:

  • Doesn’t the screen have too many elements?
  • Which parts of the screen are analysed the most by the brain?
  • Do users recognize the used visuals?
  • Do the call-to-action elements incite action?
  • Do users understand the content?

And if we do that, we can objectively measure the user experience.
I would retort that it's not possible to answer those questions by examining fMRI results alone. Eye tracking and user behavior are essential. And the danger of extrapolating user experience from the pattern of brain activity? The logical fallacy of reverse inference, flogged repeatedly on this blog. One cannot directly infer the participants' cognitive or emotional state from the observed pattern of brain activity in neuroimaging experiments (Aguirre, 2003; Poldrack, 2006).

Let's examine some of Netway's conclusions. From looking at the figure above, they say:
In the case of Facebook we see the right visual cortex has a higher level of activation. This indicates the visual elements at the left side of the interface generate more brain activity than the right-side elements.
That generally happens if the density of visual stimuli in the left visual field is greater than in the right visual field, or if there is some difference in basic perceptual features. You don't need fMRI to tell you that. Commenter Theo Vosse explains further:
Sorry, but this is nonsense. Even the low-level visual analysis is wrong. That there is more activity in one part of the visual cortex than in another only means that there is a difference in contrast, color or brightness between the left and right sides. Furthermore, since fMRI is pretty slow, this is averaged over the different places of focus, so it might just mean that people focus more on one side of the monitor than on the other. Or something completely different. You can’t tell, because you have no proper baseline.
Then we hear a bit about the dorsal ("where" pathway) and ventral ("what" pathway) visual streams. But Netway really goes awry and starts reading tea leaves when they reach the prefrontal cortex.

Facebook brain activation - semantic activation (photo by Marc Van Rymenant)

The Brodmann 44 zone is involved in recovering information in our semantic memory. This means a surfer watches the elements and this system will activate a network of knowledge about a certain word or an object.

The information that is recovered in the long-term memory during a Facebook site visit activates the semantic network. People will know what they see and that activates a set of linked information (I know this person, it is a friend of…, …).

We see the Brodmann 45 zone is not activated. If this had been the case, it would have meant the recovered information didn’t activate strong associations. That would mean the content is not very well known or not very often used by our brain.

There is absolutely no evidence for such a distinction between these two regions of the inferior frontal gyrus (not that it's entirely clear what was meant here). BA44 is activated by strong semantic associations and BA45 by weak semantic associations? That is nonsense... Neuromarketing companies do not have to subject their studies to peer review, and in fact it's detrimental to expose their proprietary methods. So buyer beware! But if corporations want to pay for such tenuous insights, I can't feel too sorry for them.


Aguirre GK (2003). Functional Imaging in Behavioral Neurology and Cognitive Neuropsychology. In: T.E. Feinberg & M.J. Farah (Eds.), Behavioral Neurology and Cognitive Neuropsychology. New York: McGraw Hill.

Poldrack RA (2006). Can cognitive processes be inferred from neuroimaging data? Trends in Cognitive Sciences 10: 59-63.

Subscribe to Post Comments [Atom]


At March 02, 2010 1:43 AM, Blogger Neuroskeptic said...

I wonder if someone could make a Facebook app that lets you upload some fMRI data and it automatically sets your status accordingly, by reverse inference:

"Neuroskeptic is... looking at something emotionally salient in the right visual field"

At March 02, 2010 6:11 AM, Blogger The Neurocritic said...

Portable brain monitoring devices that give periodic Facebook status updates will be the wave of the future. A human version of the Neurologger for pigeons can update your status based on EEG spectral analysis:

"The Neurocritic is confused by the latest bad Science paper."

Or a portable near infrared spectroscopic device can be adapted for automatic updates from your baby, using the Baby hat method for optode-scalp coupling:

"Baby Neurocritic is crying and needs a diaper change."

At March 02, 2010 12:59 PM, Blogger Marc Van Rymenant said...

I am delighted to see that the discussion has begun. The subject is indeed a passionate one, a very large one, and one that still is in its starting blocks. I am delighted because this will allow me to gain more knowledge.

Just a reminder: the reason why I write articles on this blog is to bridge the gap between the enormous amount of scientific knowledge and subjective creativity when it comes to creating digital screens. I do not publish scientific articles on my blog.

I am not a scientist. That’s why, over the years, I have surrounded myself with several PhDs and professors. They are our Scientific Council.

The combination of business and scientific efforts allows us to book progress in the field of practical techniques. It’s not about revolutionizing whatsoever; it’s about obtaining concrete business results for our clients (e.g. the 140% increase of online sales of Nespresso, in no more than 8 months, or the 29% increase of calls to eBay).

Within the frame of our methodology we perform user tests in order to combine a maximum of different objective parameters providing us precise data on the non-conscious behaviour of users (ocular fixation, pupillary diameter, attention focus, memorisation, ...).

It is within this context that Laurent (Doctor Laurent Hermoye), Arnaud (Professor Arnaud Pêtre) and myself have decided to complete our data with IRMf data allowing us mainly to measure the brain activity generated by two different screens, based on the contrasts they generate. The objective: to comprehensively investigate the activations of visual and semantic zones.

After having received the approval of the ethical committee we have initiated our work with an IRM 3 Tesla Philips, equipped with a visual projection system. We have performed an anatomical and functional IRM on 5 healthy patients between 20-40 years old/CSP+ for a period of 40 minutes.

The experimental paradigm consisted in the random presentation of 3 conditions:
- the site, 4 pages, approx. 4 sec. per page
- the site, 4 (personalised) pages, approx. 4 sec per page
- blank page, 6 sec in between each block.

Each tester had received a genuine and precise scenario for each of the two sites that needed to be tested.

The analysis of the IRMf data has been done using SPM5 (Statistical Parametric Mapping) with a minimal TMAX threshold of 2.3.

All the conclusions are derived from scientific work (Science,...). Some examples:

• Bro 45: Wagner, A. D. (2002). Cognitive control and episodic memory: Contributions from prefrontal cortex. L. R. Squire & D. L. Schacter (Eds.). Neuropsychology of Memory (3rd ed.), pp. 174-192. New York: Guilford Press

• Bro 44: Dr. Darren R. et al.(1995), Functional imaging of human right hemispheric activation for exploratory movements ,Annals of Neurology, Volume 39, Issue 2 (p 174-179)

• Bro 45/46: Buckner, R. (1996). "Contributions of specific prefrontal brain areas to long-term memory retrieval". Psychonomic Bulletin and Review 3: 149–158.

• Associative visual zone v4: *Moran & Desimone. “Selective Attention Gates Visual Processing in the Extrastriate Cortex” Science, Vol. 229, No. 4715. (Aug. 23, 1985), pp. 782-784.

• …

You will easily understand that this type of information is not interesting for the majority of UX professionals, and most certainly not on a blog :-)

Would it be possible for the different people who want to leave a comment to also provide some concrete examples and the accompanying business results? That will allow me to supply my professional knowledge with new concrete elements.

I do want to thank you again for the shared passion on UX.

A splendid day to you all...

At March 03, 2010 10:11 AM, Blogger The Neurocritic said...

Hi Marc,

Thanks for commenting and providing more information about the study. As we know, academic scientists have different interests and motivations for conducting these types of experiments than those in the business community. Neuromarketers want to see results for their clients (however this is measured), and neuroscientists want to see careful methodology and experimental design, appropriate data analysis methods, and proper interpretations of results based on the current literature.

There a large number of elements to control when comparing a participant's brain and behavioral reactions to two web sites. For instance, are your participants already familiar with the site? It seems to me quite a lot less engaging than a personalized Facebook page...


Post a Comment

<< Home

eXTReMe Tracker