The brain speeds up perception, guessing what will happen next

Imagine that you pick up a glass of Apple juice, as you suspect, the juice to take a SIP and discover that it’s actually a light beer. Even if you are not against beer, this time a breath of horrible taste. Because the context and internal state, including expectations, affect the way all animals perceive and process sensory information, explains Alfredo Fontanini, a neuroscientist from the University of stony brook in new York. In this case, expecting the wrong stimulus leads to surprises and negative reactions.

But it is not limited by perceptual quality. Among other effects, sensory system, which expects the incoming data, good or bad, can also increase the speed at which an animal detects and responds to identificeret them.

A few years ago Fontanini and his team found a direct neural evidence of this effect of acceleration in the taste cortex, the part of the brain that is responsible for perception of taste. Since then, they tried to determine the structure of cortical circuits, which led them to such results. And now, she found. Last month they published their findings in Nature Neuroscience: network model with a special type of architecture which not only provides a new understanding of how the expectation, but also delves into broader issues about how scientists should consider the perception more broadly. More than that, it comes up with decision theory, which assumes that the brain really is jumping to conclusions, not building them.

Quick feelings, and active state

Taste is the least understood sense — has become the perfect to search. Once the taste hits the tongue, it takes several hundreds of milliseconds and activity in the taste cortex begins to reflect the incoming data. “By the standards of the brain is forever,” says don Katz, a neuroscientist from the University of Brandeis in Massachusetts. “In the visual cortex because it takes less time”, which complicates the determination of the effect of expectations that scientists wanted to study.

In 2012 Fontanini and his colleagues conducted an experiment in which rats heard a sound (a”forward signal”), and then received a small portion of the taste through a tube in his mouth. By itself, the taste could be sweet, salty, sour or bitter, and a preemptive signal did not contain information about which of the four may follow.

However, scientists have discovered that these General expectations can make the neurons of the gustatory cortex to recognize the stimulus is almost two times faster than when the rats received flavor, not hearing the first sound. The latency period decreased from approximately 200 milliseconds to 120 milliseconds.

Fontanini wanted to know what type of neural network could theoretically provide faster encoding. And in this he was helped by a colleague at stony brook, a neuroscientist Giancarlo La Camera, who has previously worked on the modeling of spontaneous brain activity that occurs even in the absence of stimuli.

In the last few decades has been increasingly emphasized that most of the activities in sensor networks generated from the inside rather than under the influence of external stimuli. If you compare the activity in the visual cortex of the animal in complete darkness with the activity of the animal, which looks around, the difference will be small. Even in the absence of light sets of neurons in the cortex begin to fire together, either simultaneously or predictable waves. This joint actuation is stored in a so-called metastable States from a few hundred milliseconds to several seconds, and then the picture activity switches. Metastability, or the tendency to jump between transition States continues after the introduction of the stimulus (the stimulus), but some States, as a rule, manifest themselves more often to certain stimuli and is therefore regarded as the “encoding status”.

La Camera and others (including Katz) previously simulated metastability, creating a so-called cluster network. In her group of excitatory neurons were highly correlated, but with stimulating individual neurons were randomly connected inhibitory neurons. “This cluster architecture required for the creation of metastability,” says Fontanini.

Scientists have discovered that the same network structure was needed to recreate the effects of expectations. In metastable models with a clustered architecture, the scientists simulated total pre-emptive signal with the subsequent emergence of a particular taste stimulus. By doing this, they successfully reproduced the pattern of accelerated encoding, which Fontanini observed in rats in 2012: transitions from one metastable state to another have become faster than allowed the system to achieve faster encoding of States.

These results showed that simply by creating a network that shows these metastable activity diagrams, “you can catch a variety of neurological reactions that mimic the taste sensations,” says Fontanini.

The work of scientists was notable for the fact that provided information about what the architecture should look for in the taste cortex, and perhaps also in other sensory cortical layers. Now neuroscientists decide how to treat taste: some argue that the number of neurons can encode a “sweet” and the other “salty”, creating a very specific neural signatures for certain tastes. Others associate it with the wider models, in which most neurons respond to most tastes. The work done Fontanini and his colleagues supports the latter theory, at the same time providing projections about how should look this link.