Author: Allegra Rustici
Editor: Altay Shaw
In 1956, Phillip K. Dick published “The Minority Report,” later adapted into a film in 2002 starring Tom Cruise as the chief of the specialised police force, “Precrime”. Set in a crime-free society in the year 2054, the plot centres around the premonitions of three mutant humans called “precogs”, enabling Precrime to intercept criminal activity before it occurs. This seemingly futuristic concept, however, draws intriguing parallels to “neuroforecasting” in contemporary society. This term describes the use of small-scale brain activity to predict large-scale collective behaviour.
Neuroforecasting Crime
In the context of crime prediction, one USA-based research team found that EEG examination during a P300-based Concealed Information Test could identify individuals guilty of terrorist activity. The Concealed Information Test, also known as the Guilty Knowledge Test, involves multiple choice questioning in which critical items consistently associated with a diverse response to control items allow knowledge about an event to be inferred. The P300 wave, examined via EEG in this study, is a measurable electrical brain response most commonly elicited in response to the detection of a target (critical) stimulus amongst a succession of standard stimuli. Via these means, in a mock terrorism attack planning task, the team identified 12 out of 12 guilty individuals (who had prior knowledge of the mock attack), with no false positives in the 12 control innocent subjects.
Another USA-based study employed fMRI to predict which of a sample of 96 ex-convicts were more likely to be rearrested within 4 years of being released from prison. During a simple go/no-go task, it was shown that ex-convicts with relatively low fMRI activity in the anterior cingulate region during this impulsivity task had doubled odds of being rearrested 4 years later than ex-convicts with high activity in this region.
Though via different means to P.K.Dick’s envisioned precogs, predicting criminality is certainly feasible. However, this inevitably raises ethical considerations clearly bypassed by Precrime: should all individuals with neural markers for criminality be arrested and imprisoned? This invites the classical debate of free will versus determinism. The hard determinist would argue that since free will is an illusion, individuals with these markers have no control over their actions and therefore will inevitably commit a crime at some point. Nonetheless, given the prevailing (59% of the population) belief in free will, arresting or monitoring individuals based solely on neural markers may be deemed unethical.
Neuroforecasting the Market
Perhaps more ethical circumstances to which neuroforecasting can be extended is to predict consumer behaviour. This has been carried out in contexts such as gambling, investments, purchases, and advertisements. Amazingly, neural markers consistently outperform conscious reflection and opinion in predicting aggregate choices or responses.
Predicting Population Response to Advertisements
In a study on anti-smoking campaigns, fMRI-measured medial prefrontal cortex activity of a sample of smokers predicted population-level advert success. Participants viewed three real-life advertisements promoting the National Cancer Institute’s quit smoking hotline (ads A, B and C) while under fMRI, after which they were asked to evaluate (by preference and perceived effectiveness) the three ads. Comparing these results to population-level call volumes in the month preceding and following the campaign launch, it was found that fMRI predictions matched population-level campaign success (in the order C>B>A) whilst self-reported predictions did not match (B>A>C).
Predicting Musical Hits
In a 2012 study, a sample of teenagers was asked to listen to music clips from the internet while under fMRI and after to report their subjective rating of each song. Three years later, this data was then compared to population-level internet download statistics. The researchers discovered that averaged brain activity from the teenage sample, particularly in the nucleus accumbens, significantly correlated with aggregate song download volume , whilst self-reported preference did not.
In a study published just last year, on the other hand, a slightly different approach was used. Via combining the neural data from a sampled 33 individuals to a machine learning (a prediction-based subset of artificial intelligence) statistical model, “hit” songs versus “flop” songs were identified with 97% accuracy (i.e., which songs were the most/least successful on the market in terms of numbers of streams).
For one last example, in a study of news virality from last year, researchers found neural activity to be predictive of population-level sharing of news articles in the USA. Sampled brain activity predicted article virality more robustly than rated intentions to share each article. This study replicated the findings of a 2017 study which highlighted the predictive power of nucleus accumbens and medial prefrontal cortex activity in particular.
The future is now!
These examples represent only a fraction of the diverse contexts to which neuroforecasting has been applied. Beyond optimising marketing strategy and enhancing product allure to consumers, neuroforecasting holds the key to understanding human behaviour. Despite its currently limited use, which is primarily down to financial and ethical drawbacks, neuroforecasting (particularly in combination with artificial intelligence) promises to revolutionise our understanding of the aggregate: the human herd as a whole.
kinesismagazine.com 