Table 1: Mathematical definitions of complex network measures (see main text for an informal discussion).

All binary and undirected measures are accompanied

by their weighted and directed generalizations. Generalizations which have not been previously reported (to the best of our knowledge) are marked with a star (*).
The Brain Connectivity Toolbox contains Matlab functions to compute most measures in this table.
Measure

Binary and undirected definitions

Weighted and directed definitions

Basic concepts and measures

Basic concepts
and notation

Degree: number
of links
connected to a
node
Shortest path
length: a basis
for measuring
integration

Number of
triangles: a basis
for measuring
segregation

is the set of all nodes in the network, and is the number of nodes.
is the set of all links in the network, and is number of links.
, is a link between nodes and , .
is the connection status between and : = 1 when link , exists
(when and are neighbors); = 0 otherwise ( = 0 for all ).
We compute the number of links as = , (to avoid ambiguity with
directed links we count each undirected link twice, as and as ).
Degree of a node ,
=

Shortest path length (distance), between nodes and ,

=

where is the shortest path (geodesic) between and . Note that

= for all disconnected pairs , .
Number of triangles around a node ,
1
=
2

Links , are associated with connection weights . Henceforth we

assume that weights are normalized, such that 0 1 for all and .
w is the sum of all weights in the network, computed as w = , .
Directed links , are ordered from to . Consequently, in directed
networks does not necessarily equal .
Weighted degree of , w =

(Directed) out-degree of , out = .

(Directed) in-degree of , in = .
Shortest weighted path length between and , w =
where is a map (e.g. an inverse) from weight to
shortest weighted path between and .

w

w
length and

w = 2

1/3

Number of directed triangles around ,

1
=
, + + + .

Measures of integration
Characteristic path length of the network (e.g. Watts and Strogatz, 1998),
1
1
,
=
=
,

where is the average distance between node and all other nodes.

is the

Shortest directed path length from to , = , where is

the directed shortest path from to .
(Weighted) geometric mean of triangles around ,

Characteristic
path length

Weighted characteristic path length, w =

1

Directed characteristic path length, =

w
,

Global efficiency

Global efficiency of the network (Latora and Marchiori, 2001),

1
1
1
,
=
=
,

where is the efficiency of node .

Weighted global efficiency, =

Directed global efficiency, =

1
,

Measures of segregation
Clustering
coefficient

Weighted clustering coefficient (Onnela et al., 2005),

Clustering coefficient of the network (Watts and Strogatz, 1998),

1
1
2
=
=
,

w =

2 w
1

. See Saramaki et al. (2007) for other variants.

where is the clustering coefficient of node ( = 0 for < 2).

Directed clustering coefficient (Fagiolo, 2007),

.
out + in out + in 1 2

w
2
Weighted transitivity*, w =
.
( 1)
1

=
Transitivity

Local efficiency

Transitivity of the network (e.g. Newman, 2003),

2
=
.
( 1)
Note that transitivity is not defined for individual nodes.
Local efficiency of the network (Latora and Marchiori, 2001),
1
1
,, ( )
loc =
loc, =

Modularity

Directed transitivity*, =
1

Weighted local efficiency*,

w
loc

out + in

, ,

where loc, is the local efficiency of node , and ( ) is the length of the
shortest path between and , that contains only neighbors of .

Directed local efficiency*,

Modularity of the network (Newman, 2004b),

Weighted modularity (Newman, 2004),

2

where the network is fully subdivided into a set of nonoverlapping modules

, and is the proportion of all links that connect nodes in module with
nodes in module .
An equivalent alternative formulation of the modularity (Newman, 2006) is
1

given by = , , , where is the module

containing node , and , = 1 if = , and 0 otherwise.

loc

w =

1
2
1
w

, ,

out
in
+ 1

( )

w ( )
1

+ ( )

out + in out+ in 1 2

w w
w

, .

Directed modularity (Leicht and Newman, 2008),

=

out in

, .

1 1/3

Measures of centrality
Closeness
centrality

Betweenness
centrality

Closeness centrality of node (e.g. Freeman, 1978),

1
1
.
=
,
Betweenness centrality of node (e.g. Freeman, 1978),
()
1
=
,
1 2

,
,,

Within-module
degree z-score

Participation
coefficient

where is the number of shortest paths between and , and () is the

number of shortest paths between and that pass through .
Within-module degree z-score of node (Guimera and Amaral, 2005),

=
,

where is the module containing node , is the within-module
degree of (the number of links between and all other nodes in ), and
and are the respective mean and standard deviation of the
within-module degree distribution.
Participation coefficient of node (Guimera and Amaral, 2005),
2

= 1
,

where is the set of modules (see modularity), and is the number of

links between and all nodes in module .

Weighted closeness centrality, w

Directed closeness centrality,

w
,

Betweenness centrality is computed equivalently on weighted and directed

networks, provided that path lengths are computed on respective weighted or
directed paths.

Weighted within-module degree z-score, w =

w w

out out
out

in
in

Within-module out-degree z-score, out =

Within-module in-degree z-score, in =

Weighted participation coefficient, w = 1

In-degree participation coefficient, in = 1

in

Out-degree participation coefficient, out = 1

w
w

out

out
2
in

.

in

Network motifs
Anatomical and
functional
motifs

Motif z-score

is the number of occurrences of motif in all subsets of the network

(subnetworks). is an node, link, directed connected pattern. will
occur as an anatomical motif in an node, link subnetwork, if links in
the subnetwork match links in (Milo et al., 2002). will occur (possibly
more than once) as a functional motif in an node, link
subnetwork, if at least one combination of links in the subnetwork matches
links in (Sporns and Ktter, 2004).
z-score of motif (Milo, 2002),
rand,
=
,
rand,
rand,
where rand, and
are the respective mean and standard deviation for
the number of occurrences of in an ensemble of random networks.

(Weighted) intensity of (Onnela et al., 2005), =

, ,
where the sum is over all occurrences of in the network, and is the set
of links in the th occurrence of .

Note that motifs are directed by definition.

Intensity z-score of motif (Onnela et al., 2005), =

rand,
rand,

where rand, and rand, are the respective mean and standard deviation for
the intensity of in an ensemble of random networks.

Motif
fingerprint

node motif fingerprint of the network (Sporns and Ktter, 2004),

=

, =

, ,

where is any node motif, , is the node motif fingerprint for

node , and , is the number of occurrences of motif around node .

node motif intensity fingerprint of the network,

= , = , ,
where is any node motif, , is the node motif intensity
fingerprint for node , and , is the intensity of motif around node .

Measures of resilience
Degree
distribution

Cumulative degree distribution of the network (Barabasi and Albert, 1999),

( ),

Average
neighbor degree

Assortativity
coefficient

where ( ) is the probability of a node having degree .

Average degree of neighbors of node (Pastor-Sattoras et al., 2001),

nn, =
.

Assortativity coefficient of the network (Newman, 2002),

2
1
1 , 1 , 2 +
=
1
1
1 , 2 2 + 2 1 , 2 +

Cumulative weighted degree distribution, w =

w (

),

Cumulative out-degree distribution, out = out ( ).

Cumulative in-degree distribution, in = in ( ).
Average weighted neighbor degree (modified from Barrat et al., 2004),
w
nn,
=

Average directed neighbor degree*, nn,

=

out + in

2 out + in

Weighted assortativity coefficient (modified from Leung and Chau, 2007),

2.

2
1
w w
w
w
1
,
, 2 +
2
1
1
w 2+ w
1 , w + w
, 2

Directed assortativity coefficient (Newman, 2002),

=

2
1 out
out in
in
1
,
, 2 +
2
1
1
out 2 + in
1 , out + in
, 2

Other concepts
Degree
distribution
preserving
network
randomization.

Measure of
network smallworldness.

Degree-distribution preserving randomization is implemented by iteratively

choosing four distinct nodes 1 , 1 , 2 , 2 at random, such that links
1 , 1 , (2 , 2 ) , while links 1 , 2 , 2 , 1 . The links are then
rewired such that 1 , 2 , 2 , 1 and 1 , 1 , (2 , 2 ) (Maslov and
Sneppen, 2002). Latticization (a lattice-like topology) results if an
additional constraint is imposed, 1 + 2 + 2 + 1 < 1 + 1 + 2 + 2
(Sporns and Ktter, 2004).
Network small-worldness (Humphries et al., 2008),
rand
=
,
rand
where and rand are the clustering coefficients, and and rand are the
characteristic path lengths of the respective tested network and a random
network. Small-world networks often have 1.

The algorithm is equivalent for weighted and directed networks. In weighted

networks, weights may be switched together with links; in this case the
weighted degree distribution is not preserved, but may be subsequently
approximated on the topologically randomized graph with a heuristic weight
reshuffling scheme.

Weighted network small-worldness, w =

Directed network small-worldness, =

w
w /rand
.
w w
rand

/rand
.

rand

In both cases, small-world networks often have 1.