The human brain has cognition of the number '4', but not the number '5'! Scientific evidence supports that the four heavenly kings F4 are all '4' (doge)｜Nature

Please answer within half a second. How many apples are there in the picture below:

Do you know at a glance that there are four on the left, and... many on the right? Why can't we instantly determine which one is on the right?

In addition to this phenomenon, there are some similar examples, such as the "Four" Great Heavenly Kings, F "4", "Four" Xiao Hua Dan, etc. Why does this all have to do with the number "4"?

Nature's latest report provides a direct basis to explain this phenomenon - The brain recognizes "4" but not "5".

Researchers from the University Medical Center of Bonn, Germany, observed the activity of single neurons and found that the brain uses a different system to recognize the number "1-4" and to recognize the number "5-9",The boundary is "4".

Specifically, when neurons process the number 1-4, specific neurons are used; while when processing 5-9, the response is not specific and will be interfered by adjacent numbers.

For example: a neuron that prefers the number 3 will only respond to 3, while a neuron that prefers the number 8 will respond to 8, but also to the numbers 7 and 9

This discovery is of great significance to understanding the nature of human thinking. Lisa Feigenson, a psychologist at Johns Hopkins University, said:

Fundamentally, the question involves the structure of the mind: What forms the basis of the human mind?

The human brain recognizes "4" but not "5"

In an article published in "Nature" magazine in 1871, economist and logician William ·Stanley Jevons investigated the human counting ability and came to the following conclusion:

Some people are unable to judge the size of the number "5"

Some people are unable to judge the size of the number "5"

Researchers subsequently analyzed that this was due to the brain using only one counting system, which is less accurate for larger numbers. Others hypothesized that the difference stems from the fact that there are two independent neural systems. The system was used to count and then further studied using techniques such as electroencephalography and functional magnetic resonance imaging, but the results were inconsistent and it was impossible to determine which model was correct.

Until recently, a research team from the University of Bonn in Germany used new technology to record the activity of single neurons in the human brain, and finally discovered the difference in neural coding between small numbers and large numbers.

Specifically, the data were obtained from 17 epilepsy patients. In order to help doctors find the source of epilepsy, these patients underwent invasive medical monitoring, that is, microelectrodes were implanted in the brain

While the patients underwent monitoring surgery, the microelectrodes recorded801 The activity of a single neuron. These neurons come from four areas of the brain: the amygdala, hippocampus, ventromedial temporal lobe, and ventrolateral temporal lobe.

During the recording process, the patient needs to do a quantity judgment task:

The screen will display non-symbolized "dot matrix" images in the range of 0-9, and each image will last half a second. . Participants need to press the left and right keys to indicate whether the number on the image is even or odd.

The dot matrix is ​​presented in standard format and control format. The following are three dot matrix arrangements. ：

△ (Left) Standard layout, the size and position of points are variable; (Middle) Control layout, the total area and density of points are balanced; (Right) Points are linear Arrangement

In the behavioral test, the patient's judgment of small quantities was significantly more accurate, and the reaction time was shorter, showing "subization" characteristics.

(Saturation, a term in psychology and cognitive science, refers to the human ability to quickly estimate small quantities of items.)

When After the quantity is greater than 4, the judgment slows down and the error rate increases, reflecting the "quantity estimation" process.

The researchers calculated the error rate and reaction time respectively, and finally determined that the upper limit of the subization range was on average 3.7 (error rate) and 3.6 (reaction time).

The analysis of neuronal responses further confirmed the accuracy of the behavioral results

Decimal digitization system & large number estimation system

In a single neuron At the meta-level, the researchers also observed similar coding patterns

They found neurons that selectively responded to non-symbolized quantities in the four recording brain areas, and these selective neurons covered 0 -9 complete quantity range.

The proportion of selective neurons in different brain regions is around 15.1%, which is significantly higher than random levels.

After analyzing the neuron tuning curve characteristics and correlations, the researchers found:

For small numbers 1-4, the neurons are highly selective for the preferred number and inhibit the non-preferred number Quantitative response, thus improving the recognition effect.

For a neuron that is sensitive to changes in quantity, it will produce the maximum response to a specific quantity. The quantity that causes the maximum response is the "preferred quantity" of the neuron

When the number is greater than 4, the tuning curve of the neuron becomes wider and the selectivity decreases, and returns to the baseline level as the number increases, which is consistent with the characteristics of number estimation

The researchers then further used SVM to decode selective neurons, conduct state space analysis and clustering analysis, and conduct statistics on neuron groups The test also found that there is a classification boundary for quantity judgment, and the most significant difference is between quantities 4 and 5.

This boundary is very consistent with human intuitive judgment

This study provides direct evidence that a system of "subitization" of small numbers and a system of "estimation" of large numbers can coexist

The "subized" system may be related to attention and working memory, which explains our sharp judgment on small amounts of information, while the brain needs to rely on slower and systematic information processing for input that exceeds capacity.

However, the researchers also mentioned that the more complex network mechanisms need to be further explored, such as whether there are similar coding differences in other brain areas and whether this phenomenon can also be observed in more complex cognitive tasks. Temporal effects of inhibitory loops, etc.

Paper link: https://www.nature.com/articles/d41586-023-03136-w

The above is the detailed content of The human brain has cognition of the number '4', but not the number '5'! Scientific evidence supports that the four heavenly kings F4 are all '4' (doge)｜Nature. For more information, please follow other related articles on the PHP Chinese website!