Artificial Intelligence May Help Override Fearful Memories

SCIENTISTS have created a procedure that can remove specific fearful memories from the brain, which could be used as a new therapy for treating fear-related disorders. A team of neuroscientists used a combination of real-time brain imaging, artificial intelligence, and small monetary rewards to override fear responses to specific memories in healthy volunteers. This has the potential to provide an effective alternative to common aversion therapies for conditions such as phobias which require the patient to be exposed to the feared objects, and which many choose to avoid.

The team conducted the small study by creating a fear memory in 17 healthy volunteers. Each participant was placed inside a magnetic resonance imaging (MRI) scanner and given an electric shock every time a specific image was shown. The induced fear response could then be identified and recorded from enhanced activity in the brain’s amygdala, which controls fear.

The team then employed an artificial intelligence algorithm to spot patterns of activity that corresponded to the brain accessing this fearful memory, even while the patient was not actively thinking about it. Each time the pattern was detected, the patients were told they had received a small monetary reward. This allowed the researchers to eventually overwrite the fearful response to the memory with a positive one. “In effect, the features of the memory that were previously tuned to predict the painful shock, were now being reprogrammed to predict something positive instead,” explained the lead researcher of the study, Dr Ai Koizumi, from the ATR Computational Neuroscience Laboratories, Kyoto, Japan.

The participants were not aware of the induced fear response and were not told why their brain activity was being recorded in the study. The fearful memory was removed without the individual having to consciously confront the painful experience. This means that the new technique developed by the scientists, termed Decoded Neurofeedback, could provide a less confrontational approach in clinical treatment for patients.

“To apply this to patients, we need to build a library of the brain information codes for the various things that people might have a pathological fear of, say, spiders,” explained Dr Ben Seymour, Behavioural and Clinical Neuroscience Institute, University of Cambridge, Cambridge, UK, who was also involved in the study. “Then, in principle, patients could have regular sessions of Decoded Neurofeedback to gradually remove the fear responses these memories trigger.”

Jack Redden, Reporter

Keywords: Artificial Intelligence, Fear, Memory, MRI, Neurology, Neuroscience, Phobias, ‘Decoded Neurofeedback’