By Nicole Gelfand
Children imitate our every action- from their very first words to even the most miniscule of habits they acquire from their parents. Children are a firsthand example of how human learning often takes place by observing other individuals, a term referred to as observational learning. From a young age human brains associate observed actions with the rewards and consequences that follow, to subsequently “learn by watching” and change behavior.
A recent study from UCLA and Caltech explores which specific neurons in the brain are responsible for encoding observed behaviors. The researchers, including Erie Boorman, now an assistant professor at the UC Davis Center for Mind and Brain and Department of Psychology, were able to identify an area of the brain called the rostral anterior cingulate cortex (ACC), where neurons showed evidence of learning from observation. The two other brain areas studied did not show this pattern of response.
“These signals had not yet been identified at the single neuron level in any species,” Boorman said. “It is particularly exciting to find these learning computations in single neurons in the human brain because neuroscientists typically only have access to indirect measures of the activity of large ensembles of neurons in humans.”
Prior to the study, the researchers implanted the subjects with electrodes in order to record the responsiveness of individual neurons in three brain areas, the rostral anterior cingulate cortex (rACC), amygdala (AMY), and the rostromedial prefrontal cortex (rmPFC). (The subjects were patients with a form of epilepsy, and were being fitted with electrodes to monitor their condition.)
The research team then conducted a card game task for the subjects to perform. The players had to choose cards from either the left or the right deck. One deck provided only a 30 percent chance of winning, while the other deck had a 70 percent chance of winning. Each round, a subject was first responsible to play individually and then was instructed to observe as the other subjects played out the card game.
Neurons in the rostral ACC showed prediction errors — differences between what the observed player received and what the subject expected them to receive. This signal is considered an engine for learning.
Their findings suggest that neurons in all three brain areas are triggered through self-experience, but only neurons in the rACC signaled learning.
Boorman said that continuing research at the single neuron level could relate microscopic signals to the macroscopic signals from neuroimaging. This would help bridge the gap in understanding between insights that can be gained at the single cell level in animal models, and the indirect measures of activity that can be readily recorded in humans during complex behaviors in health and disease.
Nicole Gelfand is a Strategic Communications intern.