Scientists have discovered the mechanism by which attention regulates the value encoding of the brain's prefrontal lobe

On October 22, Nature Communications published a paper
titled "Reward salience but not spatial attention dominates the value representation in the orbitofrontal cortex.
" 。 The research was completed
by the Yang Tianming Research Group of the Center for Excellence in Brain Science and Intelligent Technology (Institute of Neuroscience), Chinese Academy of Sciences.
The study used the electrophysiological recording of awake macaques to reveal that the value coding of the orbitofrontal brain region is mainly affected by the salience of the reward from the bottom up, and is not regulated by the top-down attention, which provides help
for scientists to understand the neural mechanism of value decision-making.

In daily life, the process of making choices is often accompanied by a shift
in attention.
When choosing a book in a bookstore, the decision process may be guided by top-down attention, such as consciously looking for books with certain keywords; Attention may also be unconsciously drawn
to bestsellers with prominent locations and bright colors.
This attention influences the decision-making process
in a bottom-up way.
Previous research has shown that orbitofrontal cortex plays a key role in value-based decision-making, and its value coding may be regulated
by attention.
It is generally believed that just as the visual cortex encodes visual information, the orbitofrontal lobe's encoding of value is regulated by attention, and attention can filter irrelevant information so that neurons mainly encode the value
of the object of interest.
However, previous studies lacked rigorous behavioral or electrophysiological measures to confirm attention size and location to rigorously validate
this hypothesis.

The research group of Yang Tianming, a researcher at the Center for Excellence in Brain Science and Intelligent Technology (Institute of Neuroscience), Chinese Academy of Sciences, has designed a visual inspection task
related to attention and reward.
In this task, two graphics appear on the screen, and the macaque is prompted to pay attention to one of them and report possible brightness changes
in this graphic.
At the same time, different graphics represent juice rewards
of different sizes.
If the macaques correctly report changes in brightness, the computer randomly selects one of the two graphics on the screen as a reward for the macaques
.
This reward method decouples the pattern that the macaque pays attention to from the reward it actually receives, making it easier for researchers to distinguish
between attention and reward.

In this behavioral paradigm, the macaque is prompted to actively divide attention from the top down to complete the task
.
However, because macaques have been learning for a long time, familiarity with these figures can bring rewards
.
Larger figures are highly rewarding for macaques and are able to attract their attention
in a bottom-up way.
When macaques perform tasks, top-down attention sometimes contradicts bottom-up attention, so the figure that rewards a large amount may not be the graph
that the macaque should focus on.
So, how does the macaque brain divide attention in this case? The team combined macaque behavior with the activity of neurons in the dorsal prefrontal lobe of the brain and found that the attention of macaques is mainly guided by top-down attention, because attention is truly task-related
.
Macaques focus on the right graphics with cues: they are more correct and react faster
.
However, studies of recording OFC neurons have shown that despite ample evidence that macaques' attention is focused on the prompted picture stimulus, OFC neurons mainly reflect the more valuable of the two patterns (significant pattern), while the less valuable one (non-significant pattern) is almost completely ignored by OFC
, even when it is concerned by macaques.
Top-down attention does not dominate the value coding of the OFC, but rather the bottom-up reward saliency plays a dominant role
in the value coding of the OFC.

This achievement breaks with the previous popular view of OFC value coding, pointing out that OFC has a stable value coding
that is not affected by tasks and top-down attention.
This coding may be used for the processing of value information at the level of consciousness and may involve value-related eye movements and attention conversion calculations
.
The research work was supported
by the Ministry of Science and Technology, the Chinese Academy of Sciences and the Shanghai Municipality.

Figure 1.
Visual inspection tasks
.
The macaque watches two pictures, and different pictures correspond to different juice rewards (0~8 drops of juice).

One of the images will change in brightness, and the macaque needs to report by looking at the yellow target point at the top of the screen as soon
as it detects the change in brightness.
The researchers gave the macaques a hint (rectangular frame) in advance, and the graphics inside the box did not change
in brightness.
Therefore, the macaque needs to focus on the figure without a rectangular box
.

Figure 2.
The encoding
of OFC neurons for value variables.
Using linear regression analysis, it was found that OFC neurons encoded the value of the more significant pattern of the two graphs (yellow line) and weakly encoded the value of the insignificant pattern (green line
).

Source: Center for Excellence in Brain Science and Intelligent Technology, Chinese Academy of Sciences

The manuscripts of this account have been opened for WeChat "quick reprinting", please indicate the source of reprinting

For reproduction through other channels, please contact weibo@cashq.
ac.
cn