The power of a well-chosen image; EEG measures of brain activity and exercise

This picture:

Bd36lfSCUAAWG3h.png-large
…occasionally does the rounds on Twitter, often spurred by tweets from the kind of evidence-phobic accounts that publish whole lists of mind-blowing ‘facts’, at least 50% of which are made up. This picture has also spurred about a billion blog posts (like here, here and here), somewhat unsurprisingly, written by the kind of people who like to get their scientific evidence from a single image on Twitter.

So what’s the problem here? What the image appears to suggest at face value is that brain activity is increased after a short bout of exercise (a 20-minute walk). Sounds reasonable, right? We know that exercise has various effects on brain function, and exercise in general is definitely a good thing, now that the Western world is suffering from massive rates of obesity, diabetes, etc. I really don’t have a problem with the message here, more in the way that it’s presented.

The brain images are clearly from an EEG, but beyond that, there’s very little information in the images about what it actually represents. There are lots and lots of different things you can measure with EEG technology, such as the P300, Error-Related Negativity, C1 and P1, or much slower neural oscillations across a wide range of frequency bands. We have no information about what particular measure this image is describing. Secondly, we have no information about what the colours mean. Heat-map colour scales on brain images like this often represent statistical values (usually or scores), which is a convenient way of representing a large amount of numerical data in a visual-friendly format. Here though, we have no colour-scale information, so we have no idea what the colours represent.

Here’s some brain images I just created from some MRI data I had laying around. Took about three minutes.

brain_threshold

Big difference, right? Somewhat counter-intuitively, the left and right images above are actually the exact same functional brain data, all I did to create the right one was to lower the statistical threshold on the colour-overlay, to essentially say “Show me more results, I don’t care if they’re statistically reliable or not.” People who do this kind of work are very clued-in to these kinds of issues, and would always look for a colour-scale on these kinds of images in research papers. Clearly though the general public aren’t that conversant with statistical issues in brain imaging, because why would they be?

What we do have in the original image is an attribution to a guy called Chuck Hillman at the University of Illnois. Dr Hillman appears to be a perfectly respectable scientist, performing some perfectly respectable research focussing on the interaction between exercise and the brain. I have absolutely no problem with Dr Hillman or his obviously very worthwhile research. Looking through his articles, I can’t find an image which matches the one at the top of the post, although this paper  (PDF, Figure 2, page 548) does contain one that’s somewhat similar. That image shows the amplitude of the p300 wave during a particular task, after a period of reading and a period of exercise. Unfortunately the colour-scale here is in raw units of EEG signal (micro-volts) so it’s not totally clear if that represents a statistically-significant difference or not. If anyone can work out where the original image at the top comes from, please let me know in the comments!

As something of an aside, is an increase in brain activity necessarily a positive thing? Oxidative stress can potentially occur as the result of an increase in brain metabolism, and oxidative stress has been implicated as a potential causal factor in a huge variety of problems, from cancer to Alzheimer’s. One could even argue that lower brain activity is better because it indicates a more efficient use of cognitive resources; performing the same task, with less activity, equals greater efficiency. Although using the concept of ‘efficiency’ in this way is currently fairly controversial.

The essential point here is that when images like this are presented in academic papers or presentations, they come packaged with a whole host of caveats, qualifications, and additional information. Of course, scientists often try to make visually arresting images in order to present their results with maximum impact and clarity, and (as long as they don’t cheat in some way) that’s entirely appropriate, and indeed useful. The problem comes when someone else takes those images, strips them of this essential contextual information and presents them uncritically, often in order to further their own agenda or aims. Without the context, these images become pretty much meaningless. If this kind of thing happened to some result from my own research, I’d be pretty embarrassed about it. As ever, a critical approach to this kind of un-critically presented ‘evidence’ is crucial.

 

Advertisements

14 responses to “The power of a well-chosen image; EEG measures of brain activity and exercise

  1. Reblogged this on People Performance Potential and commented:
    There’s an incredibly important principle in here about the “use” of neuroscience that has been worrying me for some time.
    If we are not discerning… if we don’t ask the right questions… if we don’t talk to the right people… if we don’t understand the research… then we are in danger of misinformed and unethical practice.
    It’s already happening… let’s just make sure we don’t fall into that trap and that we highlight weak representations of apparent neuroscience. Neuroscientists deserve that from us don’t they?

  2. A Brainscan Picture says more than a thousand Words – depending who manipulated it 😉

    Thanks for the Insights on the Workings and Option of creating Scan Images.

    It would be great to get more technical Insights how these Scan Images are created (like creating different Settings for colouring Activity etc.) to learn what these Images actualy can show us …

    Let’s say you scan a Person doing a simply Task and show the Scan Image with different Settings. Then have an Scan Image of the same Person doing a more demanding Task – and also Variations of Scan Images.

    Just an Idea … 😉

  3. I believe the paper in question is Hillman / Pontifex, “The effect of acute treadmill walking on cognitive control and academic achievement in preadolescent children”, Neuroscience 159. http://kch.illinois.edu/Research/Labs/neurocognitive-kinesiology/files/Articles/Hillman_2009_TheEffectOfAcute.pdf Here is a PDF (With the images on page 1050, the incongruent trials, in black and white).

  4. Tried to find the source of the original image by playing with time ranges on Google’s search tools . Failed to find the source unfortunately.

    The oldest copy I could find is in this PDF file authored by “aschlorff” from October 2009 which is hosted by an elementary school alongside one of Hillman’s papers. Same brain images and text but different appearance though. http://www.fmschools.org/webpages/twiniecki/files/Compare%20brain%20scans%20of%20a%20sedentary%20child%20and%20an%20active%20child%20taking%20the%20same%20test.pdf

    Next appearance chronologically and first I can find with the metallic silver background (although different font, text style to the above image) is in an April 2011 PDF by Marianne Szeto called “Aerobic exercise primes the brain for optimal learning” hosted on the City of San Francisco website.
    http://www.sfgov3.org/Modules/ShowDocument.aspx?documentid=744

    Next appearance is in a May 2011 PDF called “Learning is a Moving Experience” by Paul Zientarski which is hosted on the “President’s Council on Fitness, Sports & Nutrition”. Same scan images and almost identical text as the 2009 copy, but with the addition of shadows on the text and “Research/” prepended to the credit. http://www.fitness.gov/pdfs/learning-is-a-moving-experience.pdf

    All the same brain scan images and similar text but different appearance. I get the feeling these images were provided by Dr. Hillman as PPT files that allowed people to go in and modify the style to suit themselves. Note that all the above PDF files started as PPT files according to the Title or Content Creator metadata on the files.

  5. Pingback: Good read: The power of a well-chosen image; EEG measures of brain activity and exercise | From experience to meaning...

  6. Good work Alistair. I dug a bit more and looks like this could possibly be the original source … here’s a link to a power point presentation prepared some time ago by Dr. Hillman. See slide 20 in particular http://www.slideshare.net/youth_nex/hillman-revised1007

  7. Pingback: The vices and virtues of brain imaging | What is behavioral?

  8. Pingback: La sélection cérébrale de la semaine (numéro 1) | Cerveau en argot

  9. Pingback: The Importance of Activity for Learning | UKEdChat.com - Supporting the #UKEdChat Education Community

  10. Pingback: The power of a well-chosen image; EEG measures of brain activity and exercise

  11. Pingback: The Amen Clinics advertising is dangerous, disingenuous bullshit | NeuroBollocks

  12. Pingback: Boomers – Got Fit (bit)? | hcldr

Leave a Reply to halcy@DIGITAL SPORTS (@halcy) Cancel reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s