The study, published in the PLOS Computational Biology journal, postulates that primates’ ability to come up with a wide variety of behaviors not anticipated by evolution is in fact a product of connections between neurons that form recurrent loops where inputs can rebound and mix in the network. Such reservoir networks can generate a unique combination of inputs that can then be used to develop the correct behavior for any new situation.
The researchers demonstrated their findings by training a network to perform a novel problem-solving task, and then by comparing the neuron activity in that model with that of a pre-frontal cortex of a primate trained to perform the same task. Needless to say, the comparison revealed striking similarities in the activity of neurons in both instances.
This discovery is an important step in understanding the local recurrent connectivity in the brain which essentially allows primates to tackle unfamiliar tasks.