BRAIN$?
Brian Knutson
23.01.01
(Note: The following is excerpted from several commentaries on a review of progress in consumer neuroscience, contained in this reference: Hutchinson, J. W., Reimann, M.,
Knutson, B., & Huber, J. Commentaries on “Reconsidering the path for
neural and psychological methods in consumer psychology.” Journal of Consumer Psychology.)
Old goals
Once upon a time in Hollywood, young director James Cameron sought to convince a group of
studio producers to fund a new movie. Dispensing with the usual slides, he turned over his
script.
He first wrote one word: ALIEN
After a pause, he added a letter: ALIENS
Finally, he added a vertical line: ALIEN$
As we now know, the project was funded (Chilton, 2022).
The director’s successful strategy implies basic principles for seeking support for a new venture. First, confirm prior success. Second, demonstrate potential generalization. Third, highlight added value.
These principles for inspiring investment might extend beyond entertainment to scientific ventures. For instance, over a decade ago, reviewers suggested that the emerging field of “Neuromarketing” (or Consumer Neuroscience) should strive to replicate (e.g., by supporting robust inference), to generalize (e.g., by revealing hidden information), and to add value (e.g., by offering benefits over existing methods; Ariely & Berns, 2010). The passage of time raises an opportunity to reflect on whether consumer neuroscience has met these desiderata. Though the current review seems to imply that it has not (e.g., “…’consumer neuroscience’ has thus far not lived up to its promises...” Clithero et al., 2023), what does the evidence suggest?
BRAIN
Can brain activity replicably predict individual choice? Inspired by animal neuroscience and early human neuroimaging studies, researchers used Functional Magnetic Resonance Imaging (or FMRI) to explore whether brain activity can predict choice. Researchers leveraged FMRI for this purpose based on its temporal (on the order of seconds), spatial (on the order of millimeters), and depth (subcortical) resolution for visualizing anticipatory activity in evolutionarily-conserved motivational circuits prior to choice (Knutson & Greer, 2008; Bartra et al., 2013; Clithero et al., 2013). Over a decade of research currently indicates that activity in a few brain regions can predict subsequent choices to purchase products, at a level approximating or exceeding self-report measures (e.g., ~75% versus 50% chance; reviewed in Levy et al., 2012). Consistent with comparative research, relevant regions include the Nucleus Accumbens (NAcc; associated with anticipating gains), the Anterior Insula (AIns; associated with anticipating losses as well as gains) and the Medial PreFrontal Cortex (MPFC; associated with balancing anticipated gains versus losses, as well as other considerations including uncertainty and time; Samanez-Larkin et al., 2015). Although questions remain and methods continue to improve, this evidence indicates that localized brain activity can robustly predict consumer choice in individuals. After the first replication, reverse inference transforms into forward inference. In predicting individual purchases with brain activity, although researchers may not have identified a “buy button,” they certainly can target activity in relevant “hedonic hotspots.”
BRAINS
Can neural predictions of consumer choice generalize? After using an individual’s brain activity to predict subsequent choice, researchers began to explore whether they could use group brain activity to forecast the choices of other groups of people (sometimes called “neuroforecasting”). Following early examples of neuroforecasting demand for popular music and health advertisements, subsequent studies verified and extended neuroforecasts of consumer demand to other markets (partially reviewed in Knutson & Genevsky, 2018). Remarkably, not only could group brain activity from predictive circuits forecast demand out-of-sample, but it could also do so above and beyond more conventional measures collected from those samples (e.g., subjective ratings or choice). Together, these findings illustrate that neuroforecasting can generalize beyond experimental samples, sometimes even better than commonly-measured behavioral variables, but have yet to clarify when or why.
BRAIN$?
Can neural forecasts of consumer demand add value? One concrete way to add value is to offer an application with a better benefit-to-cost ratio (Ariely & Berns, 2010). Although neuroimaging methods exact costs in terms of expenses and expertise, they might also confer benefits involving fewer subjects and higher signal-to-noise measurement of mechanisms that drive choice. For instance, a typical FMRI study (e.g., n~40 subjects for one hour each billed at $500.00) currently costs about $20,000.00 (not including staff and subject compensation; Clithero et al., 2023). This estimate does not drastically diverge from current pricing of focus groups or randomized telephone surveys on larger samples (e.g., n~2000 subjects). These estimates will likely change, however, with improvements in design and analysis, and the number of subjects required will depend on the robustness and generalizability of each method.
Another more abstract way to add value is to improve theory about consumer choice. While the mechanics of neuroforecasting have yet to gracefully meld with existing consumer theory, cumulative findings seem to broadly support accounts that stress the primacy of early implicit affective responses, followed by integration with more deliberative considerations (e.g., Samanez-Larkin et al., 2015). A tantalizing but as yet unrealized possibility is that neural signals might eventually inform researchers not only about which decision components guide specific individual choices, but also move different kinds of markets.
New goals
In summary, consumer neuroscientists have harnessed brain activity to replicably predict individual choice, to generalize forecasts to choices of other groups, and to add value to existing measures. Research therefore appears to have delivered on the desiderata proposed over a decade ago (Ariely & Berns, 2010). So why the pessimism about the contributions of consumer neuroscience (Clithero et al., 2023)? While researchers have met old goals, new goals may have risen to take their place. After demonstrating the possibility of neuroforecasting, researchers can do much more to delineate both the advantages and limits of applications. Any novel measure (central or peripheral, electrical or chemical, and so forth) must still run the gauntlet of satisfying measurement criteria (e.g., reliability, validity, generalizability). An interesting implication of current findings is that measures which are closer to motivational circuits might more rapidly achieve measurement quality. Researchers also now have an opportunity to develop standardized protocols as well as performance benchmarks for evaluating new methods and tracking advances. Approaching the goal of linking levels of analysis, but from a different angle, building from a core set of replicable and generalizable findings may offer the most direct route (Knutson & Srirangarajan, 2019). Theoretical integration has also lagged behind exploration of applications (possibly reflected by local undercitation; Clithero et al., 2023), and could benefit from greater synergy. Still, the demonstrated replicability, generalizability, and added value of consumer neuroscience would seem to justify investment. Who will reap the returns of such an investment, however, remains to be seen.
References:
Ariely, D., & Berns, G. S. (2010). Neuromarketing: the hope and hype of neuroimaging in business. Nature Reviews Neuroscience, 11(4), 284-292.
Bartra, O., McGuire, J. T., & Kable, J. W. (2013). The valuation system: a coordinate-based meta-analysis of BOLD fMRI experiments examining neural correlates of subjective value. NeuroImage, 76, 412-427.
Chilton,
L. (2022). James Cameron confirms ‘urban legend’ about his Aliens film pitch
was in fact true. The Independent, December 27.
Clithero, J.A., Karmarkar, U.R., Nave, G., Plassmann, H. (2023). Reconsidering the path for neural and physiological methods in consumer psychology. Journal of Consumer Psychology, In Press.
Clithero, J. A., & Rangel, A. (2014). Informatic parcellation of the network involved in the computation of subjective value. Social Cognitive and Affective Neuroscience, 9(9), 1289-1302.
Knutson, B., & Genevsky, A. (2018). Neuroforecasting aggregate choice. Current Directions in Psychological Science, 27(2), 110-115.
Knutson, B., & Greer, S. M. (2008). Anticipatory affect: neural correlates and consequences for choice. Philosophical Transactions of the Royal Society B: Biological Sciences, 363(1511), 3771-3786.
Knutson, B., & Srirangarajan, T. (2019). Toward a deep science of affect and motivation. Emotion in the Mind and Body, 193-220.
Levy, D. J., & Glimcher, P. W. (2012). The root of all value: a neural common currency for choice. Current Opinion in Neurobiology, 22(6), 1027-1038.
Samanez-Larkin, G. R., & Knutson, B. (2015). Decision making in the ageing brain: changes in affective and motivational circuits. Nature Reviews Neuroscience, 16(5), 278-289.
