Written by Chen Zimei, Yu Yuguo

Responsible editor - Wang Sizhen

Editor—Binwei Yang

The human brain consumes a lot of energy, and how to effectively allocate the energy that occupies 20% of the body's energy consumption by each brain with different perceptions and cognitive functions is crucial to understanding how the brain works and building a brain intelligent computing model[1].


The density of neurons and glial cells contained in various brain regions of the human brain, the density of synapses, the length and density of neural connections, and the rate of neural activity are all important factors in determining energy consumption [2].


But so far, systematic research and mathematical modeling of these problems are almost blank
.

On September 15, 2022, Fudan University's Laboratory of Intelligent Complex Systems and the Fahmeed Hyder team of Yale University's Center for Quantitative Nuclear Magnetic Resonance jointly published a research article entitled "A 3D atlas of functional human brain energetic connectome based on neuropil distribution" at Cerebral Cortex[3].

。
The researchers used the data obtained from tissue staining and in vivo PET carbon isotope imaging for analysis and modeling, and took the lead in constructing the spatial distribution density map of human brain neurons and synapses, as well as the digital model of human cortical energy distribution map, and then calculated the three-dimensional electrical activity map of the human cerebral cortex, which is of great significance for understanding the structure and function of ultra-low energy consumption of
the brain.

One.

The authors first obtained a density map
of neurons and glial cell distribution in the whole brain by clustering data [4] of a 7404 slice of the complete human brain[4].

In addition, according to the in vivo PET imaging data of SV2A radioligand [11cJUCB-J] in the human brain[5], the synaptic density distribution map of the whole cerebral cortex was calculated and analyzed (Figure 1), and the average synaptic density order of each brain region was obtained, of which the nucleus, cingulate gyrus and frontal lobe were significantly higher than the synaptic density
of other brain regions.

Figure 1 Spatial distribution of human brain cell and synaptic density map (Source: Yu Y, et al.
, Cereb Cortex, 2022 ) II.
Construct a digital energy distribution map of the human brain

The researchers have established two computational models: Model 1 (M1) is the use of the same discharge activity rate (f) in each brain region of the brain; Model 2 (M2) The rate of discharge activity of each brain region of the brain is a function of the spatial position of all voxels in the brain region (M2
).

Both models exhaustively calculate the basal metabolic energy consumption (HK), resting potential maintenance energy consumption (RP), action potential energy consumption (AP), synaptic transport energy consumption (ST), neurotransmitter recovery (NR) and nerve calcium activity energy consumption (Ca)
for all neurons and glial acid.

The energy profiles of both numerical calculation models are accurately compared with the voxel levels of the glycogen metabolic rate density CMRglc (ox) measured by PET in the lucid and closed eye state, and for M2, the f-spatial distribution of the human brain neuronal activity map is accurately calculated by matching the numerical values of the energy metabolic rate of glucose oxide measured experimentally at various voxel scales
.

For Model 1 (M1
).

The average frequency of electrical activity f, which best matches the statistical distribution of real brain energy, is 1.
16 Hz, and M1 also predicts a large amount of metabolic heterogeneity in living brains
.

Most of the brain energy consumption density calculated by M1 differed from the real experimental value (Figure 2), and only a few 6-8 brain regions showed significant or high or low differences, indicating that the human brain is at rest, and the activity rate and glucose metabolism rate of most brain regions are in a similar fluctuation range
.

Figure 2 Density map of energy distribution in various brain regions of the human brain based on fixed discharge frequency model 1 (M1) (Source: Yu Y, et al.
, Cereb Cortex, 2022) Next, the frequency f in each voxel is a free variable, and the measured CMRglc (ox)
is optimally fitted, which predicts the electrokinetic activity of neurons at the voxel level in all cortical brain regions.
It has great spatial distribution heterogeneity (the average rate of neuroelectrical activity of the whole brain in the resting state is about 1.
2 Hz
).

Interestingly, the resulting glucose metabolic rate distribution of M1 and M2 is statistically similar to the experimental measurement CMRglc (ox) by 90% and 99.
9%, respectively
.


There are many pivotal brain regions (HUBs) with high-density network connections in the brain[6,7], and the relationship between energy consumption and the corresponding human cortex neurons[8] and synaptic density and activity rate is still an unsolved problem
.

For HUB brain regions, energy consumption density decreases exponentially with brain region volume, while for non-HUB brain regions, this relationship is reversed
.

The calculated average activity rate of all brain regions, f is linearly related to the glucose energy consumption of the respective brain regions, indicating that signal activity is the main energy consumption factor
of human brain energy consumption.
Figure 3 Three-dimensional map and activity map of the functional energy connection group of the human brain in the waking resting state (closed eyes
).

(Source: Yu Y, et al.
, Cereb Cortex, 2022 ) Figure 4 Comparison of synaptic/nonsynaptic component related energy maps (Source: Yu Y, et al.
, Cereb Cortex, 2022 )

III.
Nonlinear coupling relationship between cell density, synaptic density, energy metabolic rate, and activity rate

The pivotal brain region spends more on ST and AP (up to 72% of the total cost, compared to only 68% of the non-central brain area), indicating that the pivot brain area is more active in the communication process and that the pivot brain area generally has higher energy
consumption than the non-pivot brain area.

In addition, the energy-consuming density of the pivotal brain region decreases with the increase in neuronal density, while the non-pivot brain area increases
.

This reflects the fact that each pivotal brain region prioritizes energy to functional activity rate and synaptic connection density at limited energy consumption levels rather than higher neuronal densities (Figure 5
).

For all brain regions, especially the pivotal brain area, the higher the density of neurons or synapses, the lower
the rate of neuronal activity.

The cost of non-signaling processes in both the pivot and non-pivotal brain regions is linearly related to neuronal density, however, the signaling costs of the pivotal brain regions decrease sharply with the increase in neuronal density, while the signaling costs of the non-pivot brain regions increase
dramatically.

This inverse relationship between f and energy consumption rates reflects the evolutionary adaptability of the human brain to the distribution of energy.
[9]

Figure 5 The relationship between the anatomical and metabolic components of the hub and non-hub regions (Source: Yu Y, et al.
, Cereb Cortex, 2022)
The article conclusions and discussions, inspiration and prospects are summarized in summary, the study is the first in the world to construct the first cortical neuron density and synaptic density spatial distribution map of the human brain based on experimental data calculation and analysis, and then constructs a three-dimensional energy digital map calculation model of the cortex, predicting the heterogeneous activity rate of all cortical brain regions; An inverse function relationship
between energy consumption and brain volume was observed in the pivotal and non-pivotal brain regions.

Smaller pivotal brain regions have higher synaptic density, energy consumption, and activity rates than non-pivotal
brain regions.

This study provides a new perspective for understanding how the human brain works, how to allocate energy to support neurons, glial cells, rich synaptic connections and neural functional activities, and also provides a reference design basis
for the development of brain-like intelligence models and neuromimetic chips.

It is worth pointing out that the current research also has some shortcomings, for example, there is a lack of more data on healthy human brain resources recorded by high-quality ultra-high-resolution imaging technology; Moreover, the glucose metabolism rate of the human brain currently measured is not dynamic and cannot reflect the energy map characteristics of time-changing under behavioral tasks, which requires more investment in experimental and theoretical research in
the future.
https://doi.
org/10.
1093/cercor/bhac322
This is a 5-year international cooperation research project, and this article is one of a series of
research results.

Professor Yu Yuguo of Fudan University is the first author and co-corresponding author of the paper, and Professor Fahmeed Hyder, director of the Center for Quantitative Nuclear Magnetic Resonance at Yale University, is the co-corresponding author
.

This research was obtained from the Chinese Brain and Brain-like 2030 Major Science Program Project (2021ZD0201301
).

National Natural Science Foundation of China (U20A20221, 81761128011), Shanghai Major Special Project of Brain-like Brain (2018SHZDZX01 and 2021SHZDZX0103), Shanghai 2021 Outstanding Academic Leader and many other projects were funded
.
Selected articles from
previous issues [1] Inflamm Regen-Wu Anguo/Qin Dalian/Wu Jianming team revealed the pathologic of Alzheimer's disease related to the improvement of active monomers of Hmong medicine to catch yellow grass [2] Transl Psychiatry - Reticular meta-analysis: Drug treatment strategies for hyperprolactinemia caused by antipsychotics [3] Nat Metab - Xiong Wei's research group elucidated a new target for alcohol and cannabis to cause motor toxicity [4] PloS Genet-Lingyan Xing/Liucheng Wu/Junjie Sun collaborate to reveal a new mechanism of non-cellular autonomic degeneration of motor neurons in spinal muscular atrophy [5] Neuron - a new molecular mechanism of intrasynaptic initiation protein regulation of ultrafast endocytosis [6] JNNP-Qiu Wei/Yu Qingfen's team discovered the susceptibility gene of familial optic nerve myelitis spectrum disease [7] Nat Neurosci - breakthrough! Brain electrical stimulation can sustainably improve work and long-term memory in the elderly[8] Nat Methods—Zhang Yang's team released a common structural comparison algorithm for proteins/nucleic acids and their complexes: US-align[9] J Neuroinflammation Review - COVID-19 and cognitive disorders: nerve invasion and blood-brain barrier dysfunction [10] CRPS review - Shen Guozhen team reviews the research progress of optoelectronic artificial synaptic devices Quality research training course recommendation [1] R Language Clinical Prediction Biomedical Statistics Special Training (October 15-16, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing) [2] Course Preview | How can neural stem cells be cultured efficiently? This time to give you a thorough lecture (September 22, 2022 (Thursday) 19:00-20:30) Conference/Forum Preview [1] Course Preview | How can neural stem cells be cultured efficiently? This time to give you a thorough lecture (September 22, 2022 (Thursday) 19:00-20:30) [2] Preview | Neuromodulation and Brain-Computer Interface Conference (October 13-14, Beijing Time) Welcome to "Logical Neuroscience" [1] Talent Recruitment - "Logical Neuroscience" Recruitment (Network Part-time, Online Office) References (Slide Up and Down)

[1] Shulman RG,Hyder F,Rothman DL.
Base line brain energy supports the state of consciousness.
Proc Natl Acad Sci.
2009:106(27):11096–11101.

[2] Yu Y,Herman P,Rothman DL,Agarwal D,Hyder F.
Evaluating the gray and white matter energy budgets of human brain function.
J Cereb Blood Flow Metab.
2018:38(8):1339–1353.

[3] Yu Y, Herman P, Rothman DL, Agarwal D, & Hyder F.
(2022), A 3D atlas of functional human brain energetic connectome based on neuropil distribution, Cerebral Cortex, 1-17，https://doi.
org/ 10.
1093/cercor/bhac322.

[4] Amunts K,Lepage C,Borgeat L,Mohlberg H,Dickscheid T,Rousseau ME,Bludau S,Bazin PL,Lewis LB,Oros-Peusquens AM,et al.
BigBrain: an ultra high-resolution 3D human brain model.
Science.
2013:340(6139):1472–1475.

[5] Finnema SJ,Nabulsi NB,Eid T,Detyniecki K,Lin SF,Chen MK,Dhaher R,Matuskey D,Baum E,Holden D,et al.
Imaging synaptic density in the living human brain.
Sci Transl Med.
2016:8:348ra396.

[6] Liang X,Zou QH,He Y,Yang YH.
Coupling of functional connectivity and regional cerebral blood flow reveals a physiological basis for network hubs of the human brain.
Proc Natl Acad Sci.
2013:110(5):1929–1934.

[7] Tomasi D,Wang GJ,Volkow ND.
Energetic cost of brain functional connectivity.
Proc Natl Acad Sci.
2013:110(33):13642–13647.

[8] Herculano Houzel S.
Coordinated scaling of cortical and cerebellar numbers of neurons.
Front Neuroanat.
2010:4:12.

[9] Bullmore ET,Sporns O.
The economy of brain network organization.
Nat Rev Neurosci.
2012:13(5):336–349.

End of article