Untangling the roots of human intelligence is something that is sure to keep neuroscience in a job for many years to come, although the findings of a new study may provide a good starting point. According to the paper, which appears in the journal PLOS, the primate brain is inherently capable of adapting to a limitless number of unique situations, thanks to a particular quirk in its wiring.
For many years, the brain was thought of as an impossibly complex machine, containing an array of different parts that perform specific and highly predictable functions, yet which, when combined, can achieve amazingly convoluted feats of cognition. However, according to this model, the brain’s computational power is defined and limited by the number of different parts it contains, since each component can only carry out one function, rather than improvising.
However, studies later began to show that while some of the brain’s neurons do indeed perform just one task by responding to specific stimuli, others are actually able to respond to a wide range of different things. These neurons – called mixed selectivity neurons – massively increase the brain’s capacity, raising the number of tasks it can perform by a factor of several million.
Crucially, mixed selectivity neurons are particularly prevalent in the prefrontal cortex – the part of the brain responsible for conscious thought and learning.
The brain was once thought of as a mechine-like contraption. Christian Lagerek/Shutterstock
This discovery inspired scientists to start developing computer models that operate along the same principles as brains containing mixed selectivity neurons, in order to try and unscramble the workings of our thought-makers. In the latest computational study, researchers decided to try and emulate the brain’s functioning using a so-called reservoir network.
Reservoir networks are computer models with high mixed selectivity, yet are defined by the fact that their functioning is not determined only by the stimuli that they are presented with at a given moment, but by past experiences as well. This is because previous neuronal inputs and outputs determine the state of each mixed selectivity neuron, and therefore continue to exert an influence over their likely response to future stimuli.
The researchers trained the reservoir network to perform a problem-solving task, whereby it had to learn which of four targets presented on a screen was the “correct” one. At the same time, the team trained macaque monkeys to perform the same task, by giving them a fruity treat when they picked the right target.
By comparing the patterns of activity of the neurons in the reservoir network to those of the prefrontal cortex in the macaques’ brains, the study authors found a number of striking similarities. As such, they propose that primate brains – which includes the human brain – are “pre-adapted” to be able to learn the correct skills in a limitless number of life situations, thanks to this reservoir-like capacity, whereby previous neuronal inputs continue to loop and rebound through time.
Primates like macaques and humans have amazingly versatile brains. pathdoc/Shutterstock
