焦点功能:
Network Communication in the Brain
Editorial: Network Communication in the Brain
1
Daniel Graham
2
, Andrea Avena-Koenigsberger
3
, and Bratislav Miši´c
1心理学系, Hobart and William Smith Colleges, 日内瓦, 纽约, 美国
2University Information Technology Services, 印第安纳大学, 布卢明顿, 在, 美国
3McConnell Brain Imaging Centre, Montréal Neurological Institute, 麦吉尔大学, 蒙特利尔, Quebec, 加拿大
关键词: Brain connectivity; Communication models; Connectome; Network dynamics; 控制-
lability
开放访问
杂志
抽象的
Communication models describe the flow of signals among nodes of a network. In neural
系统, communication models are increasingly applied to investigate network dynamics
across the whole brain, with the ultimate aim to understand how signal flow gives rise to
brain function. Communication models range from diffusion-like processes to those related
to infectious disease transmission and those inspired by engineered communication systems
like the internet. This Focus Feature brings together novel investigations of a diverse range of
mechanisms and strategies that could shape communication in mammal whole-brain
网络.
How does a massive network of neurons give rise to intercommunication across the entire
脑? As advances are made in understanding mammalian brain network structure (看, 例如,
Assaf et al., 2020; Bassett & 斯波恩斯, 2017), the question of how such networked neural ele-
ments intercommunicate, and ultimately give rise to brain function, is undoubtedly one of the
most intriguing scientific inquiries today (Avena-Koenigsberger et al., 2018).
我
D
哦
w
n
哦
A
d
e
d
F
r
哦
米
H
t
t
p
:
/
/
d
我
r
e
C
t
.
米
我
t
.
t
/
/
e
d
你
n
e
n
A
r
t
我
C
e
–
p
d
我
F
/
/
/
/
/
4
4
9
7
6
1
8
6
7
0
6
8
n
e
n
_
e
_
0
0
1
6
7
p
d
.
t
Being the pinnacle of complex systems, brain networks can be studied across a spectrum of
spatial and temporal scales that span various orders of magnitude. On one end of the scale,
single-neuron biophysical models only partially constrain the range of possible solutions as
to how communication takes place (and single-neuron models are themselves undergoing re-
想象, 例如, Sardi et al., 2017; Gidon et al., 2019). 另一端, at the whole-brain level,
emergent network dynamics could resemble physical processes such as diffusion or driven
动力系统, but could also resemble dynamics of infectious diseases, engineered com-
munication networks like the internet, or other systems. This Focus Feature investigates a di-
verse range of mechanisms and strategies that could influence communication across mammal
whole-brain networks.
F
乙
y
G
你
e
s
t
t
哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3
Leading off is compelling evidence from Seguin et al. (2020) as to the importance of including
communication goals and constraints in modeling brain network dynamics. Using human struc-
tural and functional imaging data, Seguin et al. show that existing approaches based on network
structure alone predict little variance in node activity. 相比之下, approaches that include an
explicit model of network communication perform substantially better. Seguin et al. find that the
best predictors assume signals are communicated on random walks or on short paths deter-
mined from a node’s knowledge of local network structure. Simulated activity under these ap-
proaches also performs almost as well as empirical functional activity in predicting behavioral
dimensions of individuals, such as tobacco use.
引文: Graham, D ., Avena-
Koenigsberger, A。, & Mišic, 乙. (2020).
Editorial: Network Communication in
the Brain. 网络神经科学, 4(4),
976–979. https://doi.org/10.1162/netn
_e_00167
DOI:
https://doi.org/10.1162/netn_e_00167
版权: © 2020
麻省理工学院
在知识共享下发布
归因 4.0 国际的
(抄送 4.0) 执照
麻省理工学院出版社
Editorial: Network Communication in the Brain
The realization that many complex networks share common architectural traits and statistical
properties has afforded network neuroscience a broader perspective of communication models
that can be implemented to study brain networks. Three papers in this Focus Feature examine
mechanisms that are novel or that are applied in a new context.
Lella et al. (2020) use a model of network communication inspired by infectious disease sprea-
ding to illuminate Alzheimer’s disease (广告). They construct an analytical model of network-
wide communication from structural imaging of healthy humans and those with AD. The model
assumes that nodes can utilize redundant paths, a property that is quantified via a measure
termed communicability (see Estrada & Hatano, 2008; Crofts & Higham, 2009). This measure
shows much greater differences between patient and control participants compared to a short-
est path measure. Strikingly, and counterintuitively, the model shows that nodes in the AD
brain are actually closer to one another in terms of the communicability measure, presum-
ably due to the pattern of network damage engendered by the disease. The authors conclude
that this result supports the notion that AD is spread along brain networks via an infectious
disease vector. This work connects with studies of similar types of communication models that
are increasingly used to understand the spread of misfolded proteins across a range of neu-
rodegenerative diseases (for a review, see Carbonell et al., 2018)
Shadi et al. (2020) test a thresholding model on the mouse connectome, based on an elabora-
tion of influences of single-neuron dynamics. Their model, termed an asynchronous lin-
ear threshold model (Granovetter, 1978; Miši´c et al., 2015), includes a McCulloch–Pitts-like
threshold based on empirical connection weights (tracer-based fiber volumes) in combination
with a consideration of empirical physical distances between nodes and resultant signal de-
lays. The behavior of this intriguing model suggests that a few regions such as the claustrum
and posterior parietal cortex are instrumental in generating cascades of multimodal sensory
signals that ultimately spread throughout the brain.
A crucial component of brain network dynamics that has received little attention concerns
interactions among signals. Most models assume signals do not interact (but see Miši´c et al.,
2014). 然而, Hao and Graham (2020) argue that interactions are likely, given the extremely
short network distances between nodes in mammal brain networks. Hao and Graham (2020)
focus on collisions, which are ubiquitous in large-scale engineered communication systems.
They compare numerical simulations of two routing protocols when collisions are considered:
a standard random walk strategy and an “information spreading” scheme similar to the infec-
tious disease model of Lella et al. (2020).
In simulations on two tracer-based connectomes
of the macaque monkey cortex and one of the mouse whole brain, Hao and Graham (2020)
show that information spreading actually achieves lower overall activity and greater sparse-
ness of activity compared to a random walk model. Hao and Graham (2020) provide evidence
that the mammal brain network is well suited to generating efficient network communication
through a dynamic interplay of signal creation and destruction.
The hierarchical nature of brain networks also likely influences communication strategies
(Vázquez-Rodríguez et al., 2019). Vázquez-Rodríguez et al. (2020) investigate how hierarchies
within and across modalities guide network communication based on imaging data. Their analy-
sis shows that messages are likely to be passed to nodes nearby in the hierarchy. 此外,
they begin to broach the question of selective control of signals, which must operate on brain
networks given that network structure is fixed over the short term but yet must achieve real-
time routing of attention, decision outputs, invariances, ETC. (see Graham & Rockmore, 2011).
网络神经科学
977
我
D
哦
w
n
哦
A
d
e
d
F
r
哦
米
H
t
t
p
:
/
/
d
我
r
e
C
t
.
米
我
t
.
/
t
/
e
d
你
n
e
n
A
r
t
我
C
e
–
p
d
我
F
/
/
/
/
/
4
4
9
7
6
1
8
6
7
0
6
8
n
e
n
_
e
_
0
0
1
6
7
p
d
t
.
F
乙
y
G
你
e
s
t
t
哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3
Editorial: Network Communication in the Brain
Vázquez-Rodríguez et al. (2020) demonstrate the potential for systematic “detours” or re-
routing of messages, especially in attentional networks, which could achieve selectivity.
One of the main assumptions underlying these studies is that the network organization, 超过
its individual components, is what largely conditions the emergence of higher level commu-
nication dynamics. 换句话说, these models do not rely on top-down control of signal
flow in the brain (as there is in traditional telephone networks, 例如). 然而, 这
does not mean that a subset of nodes cannot exert strong influence on the entire network.
The Focus Feature concludes with two papers that consider the role of controllability, a notion
borrowed from the study of dynamical systems in physics. Controllability in networks captures
the degree to which network dynamics can be driven by a small subset of nodes.
Patankar et al. (2020) examine the relationship between network structure and controllability. Uti-
lizing numerical simulations over networks derived from human structural imaging data, 他们
show that the relationship between network structure and controllability is complex, 然后
it depends in part on connection weights. Specifically, Patankar et al. (2020) find that measures
of network modularity are not closely related to controllability, whereas measures that consider
connection weights and hub-like properties can succeed in predicting controllability.
最后, Srivastava et al. (2020) provide a review of control theory as applied to brain networks,
focusing on similarities and differences between frameworks based on network control and
those based on network-wide communication. Srivastava et al. (2020) show intricate connec-
tions and contrasts related to the models’ level of abstraction, dynamical complexity, 和别的
因素. They argue that the two frameworks can and should be integrated to build richer and
more insightful models of whole-brain dynamics.
This collection of studies broadens the range of communication models in brain networks and
highlights novel structural and functional demands that are likely at play. One can expect that
this work will lead to further blossoming in this area of investigation and a deeper consideration
of network communication in other areas of brain science.
我
D
哦
w
n
哦
A
d
e
d
F
r
哦
米
H
t
t
p
:
/
/
d
我
r
e
C
t
.
米
我
t
.
t
/
/
e
d
你
n
e
n
A
r
t
我
C
e
–
p
d
我
F
/
/
/
/
/
4
4
9
7
6
1
8
6
7
0
6
8
n
e
n
_
e
_
0
0
1
6
7
p
d
.
t
参考
Assaf, Y。, Bouznach, A。, Zomet, 奥。, Marom, A。, & Yovel, 是. (2020).
Conservation of brain connectivity and wiring across the mam-
malian class Nature Neuroscience, 23, 805–808.
Avena-Koenigsberger, A。, Miši´c, B., & 斯波恩斯, 氧. (2018). Commu-
nication dynamics in complex brain networks. 自然评论
神经科学, 19(1), 17.
Bassett, D. S。, & 斯波恩斯, 氧. (2017). Network neuroscience. 自然
神经科学, 20(3), 353–364.
Carbonell, F。, Iturria-Medina, Y。, & 埃文斯, A. C. (2018). Mathemat-
ical modeling of protein misfolding mechanisms in neurological
疾病: A historical overview. Frontiers in Neurology, 9, 37.
Crofts, J. J。, & Higham, D. J. (2009). A weighted communicability
measure applied to complex brain networks Journal of the Royal
Society Interface, 6(33), 411–414.
Estrada, E., & Hatano, 氮. (2008). Communicability in complex net-
作品. Physical Review E, 77(3), 036111.
Gidon, A。, Zolnik, 时间. A。, Fidzinski, P。, Bolduan, F。, Papoutsi, A。,
Poirazi, P。, & Larkum, 中号. E., (2020). Dendritic action potentials
and computation in human layer 2/3 cortical neurons. 科学,
367(6473), 83–87.
F
乙
y
G
你
e
s
t
t
哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3
Graham, D. J。, & Rockmore, D. 氮. (2011). The packet switching brain.
认知神经科学杂志, 23(2), 267–276.
Granovetter, 中号.
(1978). Threshold models of collective behavior.
American Journal of Sociology, 83(6), 1420–1443.
Hao, Y。, & Graham, D. (2020). Creative destruction: Collisions and
redundancy generate emergent sparseness on the mammal connec-
tome. 网络神经科学, 4(4), 1055–1071. https://doi.org
/10.1162/netn_a_00165
Knoblauch, K., 肯尼迪, H。, & Toroczkai, Z. (2016). The brain in
空间. In Micro-, meso-and macro-connectomics of the brain
(PP. 45–74). 施普林格: 剑桥, 嘛.
Lella, E., & Estrada, 乙.
(2020). Communicability distance reveals
hidden patterns of Alzheimer’s disease. 网络神经科学,
4(4), 1007–1029. https://doi.org/10.1162/netn_a_00143
Miši´c, B., 贝策尔, 右. F。, Nematzadeh, A。, Goni,
J。, Griffa, A。,
哈格曼, P。, . . . 斯波恩斯, 氧. (2015). Cooperative and competitive
spreading dynamics on the human connectome. 神经元, 86(6),
1518–1529.
Miši´c, B., 斯波恩斯, 奥。, & McIntosh, A. 右. (2014). Communication ef-
ficiency and congestion of signal traffic in large-scale brain net-
作品. 公共科学图书馆计算生物学, 10(1), e1003427.
网络神经科学
978
Editorial: Network Communication in the Brain
Patankar, S. P。, Kim, J. Z。, Pasqualetti, F。, & Bassett, D. S.
(2020).
Path-dependent connectivity, not modularity, consistently pre-
dicts controllability of structural brain networks. Network Neuro-
科学, 4(4), 1091–1121. https://doi.org/10.1162/netn_a_00157
Sardi, S。, Vardi, R。, Sheinin, A。, Goldental, A。, & Kanter, 我. (2017).
New types of experiments reveal that a neuron functions as multi-
ple independent threshold units. 科学报告, 7(1), 1–7.
Seguin, C。, Tian, Y。, & 扎莱斯基, A. (2020). Network communication
models improve the behavioral and functional predictive utility of
the human structural connectome. 网络神经科学, 4(4),
980–1006. https://doi.org/10.1162/netn_a_00161
Shadi, K., Dyer, E., & Dovrolis, C.
(2020). Multi-sensory integra-
tion in the mouse cortical connectome 1 using a network diffu-
sion model. 网络神经科学, 4(4), 1030–1054. https://土井
.org/10.1162/netn_a_00164
Srivastava, P。, Erfan, N。, Kim, J. Z。, Ju, H。, 周, D ., Becker, C。,
(2020). 楷模
Pasqualetti, F。, Pappas, G. J。, & Bassett, D. S.
of communication and control for brain networks: Distinctions,
convergence, and future outlook. 网络神经科学, 4(4),
1122–1159. https://doi.org/10.1162/netn_a_00158
Vázquez-Rodríguez, B., Sárez, L. E., Markello, 右. D ., Shafiei, G。,
Paquola, C。, 哈格曼, P。, . . . Miši´c, 乙.
(2019). Gradients of
structure–function tethering across neocortex. 诉讼程序
美国国家科学院, 116(42), 21219–21227.
Vázquez-Rodríguez, B., 刘, Z. Q., 哈格曼, P。, & Miši´c, 乙. (2020).
Signal propagation via cortical hierarchies. Network Neuro-
科学, 4(4), 1072–1090. https://doi.org/10.1162/netn_a_00153
我
D
哦
w
n
哦
A
d
e
d
F
r
哦
米
H
t
t
p
:
/
/
d
我
r
e
C
t
.
米
我
t
.
t
/
/
e
d
你
n
e
n
A
r
t
我
C
e
–
p
d
我
F
/
/
/
/
/
4
4
9
7
6
1
8
6
7
0
6
8
n
e
n
_
e
_
0
0
1
6
7
p
d
.
t
F
乙
y
G
你
e
s
t
t
哦
n
0
7
S
e
p
e
米
乙
e
r
2
0
2
3
网络神经科学
979
