17: Measuring The Brain
The author takes a dim view of the companies (well over 100 of them) that offer brain scanning to marketers. He regards them as "snake oil salesmen" who are not explicit of the methodology they use or the validity, accuracy, or even the meaning of the metrics they sell - and that the information they offer is of little if any value. As a result, companies are paying large amounts of money to be led woefully astray, and would do well to be more knowledgeable and discerning about brain measurements.
There are a collection of different techniques used for obtaining measurements pertaining to the brain - even the valid ones have strengths and weaknesses that make them appropriate for some things but not for others, both in terms of what the instruments are capable of measuring and the way in which the measurements are interpreted.
In general, it must be conceded that the science of brain measurement remains very primitive. The tools we have today are not capable of gathering sufficiently granular data from trillions of variables, nor at the speed at which neurons fire.
While the tools at our disposal today are undeniably better than those used in the past, it also cannot be denied that there is still much progress to be made before they can claim the level of accuracy of other scientific measurements.
Single-Cell Measurement
It is entirely possible to detect whether a single neuron is firing, and detect with some precision the degree to which it fires. The example is cited of an experiment at Cal Tech where a single neuron was noted to fire when shown a specific image (oddly, a photo of Jennifer Aniston) but not when the subject was viewing other images.
This does not mean that this is the only neuron involved in the process, as it is a small part of a larger network, but it does establish the fact that observation is possible at the cellular level. Also, we can observe only that the cell is active, which is still fairly general given that it may be communicating to ten thousand other neurons, and each synapse can involve a number of chemical signals.
Electroencephalography (EEG)
The EEG was discovered in 1929, when it was found that electrodes placed on the scalp can detect electrical activity in the brain. The measurements were improved by computer technology, which is more precise than former methods in interpreting the data numerically and storing data in real time.
While it is fairly sensitive to the level of activity, it is not particularly specific to where that activity occurs: even the 64-electrode unit can only trace a signal to a fairly large area of the brain. These 64 data points give us only a very superficial sense of the several trillion possible electrical events that may be occurring at any moment. Worth noting: the equipment used in most marketing research is a 16-electrode unit, which takes even fewer data points and less frequent readings.
EEG is also limited to electrons that are relatively near to the surface of the brain, and the activity beneath it is undetectable due to its depth within the organ or because it may be concealed by stronger activity in the same general vicinity. Since sub cortical regions such as the hypothalamus are considered to be extremely important, their invisibility to EEG is concerning.
As such, when marketers leverage EEG studies, even using the most sophisticated hardware and software available, they should be well aware of the weakness of this methodology, and especially so when EEG results claim to be measuring emotion in the limbic system (which is a deep-brain structure).
Magnetoencephalography (MEG)
The MEG is used in a similar fashion to the EEG, but instead of measuring electrical activity in the brain, it measures the magnetic fields that result from that electrical activity. It is less intrusive, but also more indirect in the way in which brain activity is interpreted.
The MEG can measure at millisecond-intervals (same as EEG) but can pinpoint the source of electrical activity to within a few millimeters of the cortical surface - and like EEG, it is limited to surface activity.
The MEG does not produce an image of the brain, but can be used in conjunction with an MRI scan to superimpose an indication of activity over an image of the subject's brain - and it is critical to use the two in conjunction because the "hills and valleys" of the brain's surface have a strong effect on MEG readings.
The medical use of MEG is to determine areas of the brain that should be avoided during brain surgery. The use of MEG for behavioral research remains highly questionable.
Transcranial Magnetic Stimulation (TMS)
TMS is an experimental technique that consists of applying magnetic impulses on the scalp of a test subject to excite or inhibit activity in small areas of the brain. For example, applying an impulse to some areas of the brain can cause the subject's hands to twitch, and other areas make it difficult for the subject to control the voluntary movement of their hands.
While all of this is great fun at parties, it still remains experimental and does not seem likely to yield any practical application for marketers.
Indirect Measures Of Neural Activity
The techniques above attempt to measure electronic activity of neurons directly. A few other methods exist for measuring activity indirectly, such as examining the blood flow and oxygen levels in the brain.
Positron Emission Tomography (PET) is one technique for doing so, but because it is expensive and requires the patient to be injected with a radioactive tracer to measure blood flow, it is not often used.
Magnetic Resonance Imaging (MRI) is commonly used for brain imaging, and is the technology most laymen consider when "brain scans" are mentioned. The device into which the subject is inserted generates a strong magnetic field (600,000 the magnetic field of the Earth), which causes the blood oxygen atoms to line up, giving a very precise map of blood flow through the bodily tissues.
These machines cost $4 million, weigh more than six tones, and have significant requirements for power and radiation shielding. Scanning a patient requires a significant amount of time during which the subject must remain immobile. Combined, these factors make MRI a highly impractical tool for market research.
Very Indirect Measurements Of Brain Activity
There are "several" measurement technologies that consider other biometrics, such as the skin's conductivity or the movement of the subject's eyes, to infer the function of the brain without a reliable causal connection.
The author indicates that he will not expend much effort into considering them, and neither should anyone else - they are simply invalid.
'Reading' The Measures
Ultimately, what all measurements of the brain deliver is a picture of the "noise" in the brain, which can be compared to similar pictures of the same activity at different times - and the difference between the measurements must be interpreted to determine its significance.
Neuroscientists collect data about the brain in a resting state, then provide a given stimulus, then assume that the difference in the picture of the brain between the two states was caused by the stimulus. The procedure is repeated for a number of subjects to rule out idiosyncrasies, and general observations can be made. All of this conforms to standard practices of the scientific method.
However, there is much conjecture over what the "resting" state of the human brain happens to be, and whether we can isolate a given stimulus, along with they typical array of questions about contamination of experiment and results and the accuracy of a given approach to interpretation.
Creativity In Designing Brain Scanning Experiments
None of these methods, nor any method that might be conceived, will produce meaningful results if the experiments in which they are used are ill-conceived. To merely "push someone into an MRI machine asking them to think of something and hoping you will learn something" is unlikely to produce any usable results. And when a researcher attempts to make sense of data that is bad in the first place, the outcome has less to do with the actual phenomenon and more to do with the creativity of the researcher in designing the experiment and forcing some meaning upon the outcome.
Where marketing attempts to conduct laboratory-like research, the problems are amplified by unsound design: they hypothesis is unclear, the methodology is inappropriate, the sample is not representative, there is no control group, etc.
The author uses an experiment that investigated spatial navigation skills in taxi drivers, as a means to show the methodology that a truly scientific inquiry uses. (EN: I'll skip the details, but it is very detail-oriented and focused.)