Nature Neuroscience: How does the brain make decisions? Scientists reveal corticosteral-thalamic neural circuits for decision-making behavior

Click on the blue letter to follow us

In decision-making, frontal cortex neurons are selective
for different variables such as sensation, movement, and cognition.
One view is that the selective presence of these variables in a shared population of neurons complicates
neural activity.
This view is supported by neurophysiological records: the firing activity of a single neuron behaves similarly to a random combination
of continuous temporal variation and task selectivity.

There is an interactive projection in the motor cortex and thalamus and cortex: in the mouse tentacle motor cortex, input from the somatosensory cortex preferentially innervates superficial neurons, while input from the thalamus innervates deep neurons
.
Superficial neurons, on the other hand, preferentially project back into the somatosensory cortex, and deep neurons project the thalamus, forming different long-range projection loops
.
The frontal cortex also forms interactive neural projections with the thalamus and other cortical regions to maintain sustained activity
.

On September 28, 2022, the Nuo Li research team in the Department of Neuroscience of Baylor College of Medicine revealed the functional neuronal activity
of the frontal cortex in the decision-making behavior of the thalamus.

1

A group of neurons with different active patterns that determine behavior

Figure 1: Beard decision-making behavior paradigm

Mice use their whiskers to identify the position of the object (in front or behind, called the stimulation period) and make a choice (delay period) based on their own judgment and finally make a "left lick" or "right lick" (reaction period) to get the water reward
.
Depending on the position of the object (stimulus) mouse to make a choice of licking direction can be divided into four experimental results, correct experiments (front position left lick or back position right licking), wrong experiment (front position right lick or back position left licking, ).



The researchers recorded the electrical activity of 20,000 neurons in the anterolateral motor cortex (ALM) region of beard decision-making behavior through electrodes, and they found that these neuronal activities did not occur randomly, but encoded the behavior
according to a certain tissue structure.

According to the stimulation location, selection direction and experimental results, they divided the ALM neuronal activity into 7 different active mode neuron groups such as stimulation, selection, action, result, and response, and different activity mode neuron groups predicted different behavioral characteristics
.

For example, in the wrong experiment, the stimulation pattern characterizes the type of experiment (the direction in which the object is placed), regardless of the direction of the licking, while the selection and action patterns characterize the direction
of the mouse's licking.

On the other hand, although there are huge differences in these different activity pattern neurons, they also "share" the neuron population
.
These "shared" groups of neurons have different weights for each active pattern and contribute differently to different
behaviors.

Figure 2: 7 neuronal activity patterns in beard decision-making behavior

2

ALM accepts input from different brain regions

Retrograde tracing virus experiments and anterograde tracer virus discovery ALM receives interactive input
from the ipsilateral primary and secondary somatosensory cortex (collectively referred to here as S1/S2), ipsilateral thalamus (partial ventral medial nucleus, ventral anterolateral nucleus, medial dorsal nucleus, etc.
), and contralateral ALM.

During the stimulation of the object, the ability
to complete the task is impaired after inhibiting the activity of the contralateral S1/2 or thalamus.
During the delayed period, the light suppresses the contralateral ALM and the thalamus in the upcoming licking direction in the ipsilateral direction
.
This suggests that the input received by the ALM from three different brain regions can all influence the tentacles to determine behavior
.

3

The ALM-thalamic circuit is the primary circuit that determines behavior

Further experiments can be found that light inhibition of S1/S2 or contralateral ALM silences most neurons in the superficial layer, but the activity in the deep layers is less
affected.
Light inhibition of thalamic activity can inhibit neuronal activity in all layers of the cortex, of which the deep neuron activity is more strongly
affected.

During the experimental stimulation period, light inhibition of neuronal activity in the S1/S2 region temporarily inhibits ALM selection and stimulation activity patterns, while inhibition of thalamic activity inhibits ALM selection and stimulation activity patterns
for a long time.
In all active patterns, the effect of light inhibition of thalamic activity was somewhat stronger than that of the S1/S2 region and the contralateral ALM region, suggesting that the tentacles determine behavioral dependence on the ALM-thalamic circuit
.

Figure 3: Activity patterns of different neural circuits regulating decision-making behavior

summary

This paper reveals how the neuronal population of the frontal cortex encodes perceptual decision-making behavior, and reveals the neural loop mechanism behind it: the cortic-thalamic circuit plays a key role
in decision-making behavior.

【References】

1.
Yang, W.
, Tipparaju, S.
L.
, Chen, G.
 et al.
 Thalamus-driven functional populations in frontal cortex support decision-making.
 Nat Neurosci (2022).
 

https://doi.
org/10.
1038/s41593-022-01171-w

The images in the article are from references