In their recent paper, Chang et al. (2012) explained some inconsistencies in the results for the number of insular cortex clusters. They reported three insular clusters, one anterior dorsal, one posterior dorsal, and one middle-anterior ventral. In a previous paper (Cauda et al. 2011) on the same subject, we parcellated the insula into two clusters, one anterior ventral and one posterior dorsal. We confirmed these results in a recent meta-analytic paper (Cauda, Costa et al. 2012).

To explain the divergences between the two studies, Chang et al. claimed that we “employed a very coarse resolution (only 10 insular region of interest (ROIs) compared with our 1252) and a priori fixed the number of clusters to 2.” This is incorrect, as we actually performed two different analyses: an ROI-based analysis employing 10 ROIs and a voxelwise analysis employing a number of ROIs equal to the number of voxels in the brain. Furthermore, the optimal number of clusters was not fixed but was determined using an appropriate procedure [see Supplementary Materials of (Cauda et al. 2011)].

We would like to suggest a simple explanation for the discrepancies between the studies that parcellated the insular surface into 2 (Jakab et al. 2011; Cloutman et al. 2012) (Taylor et al. 2009; Cauda et al. 2011; Cauda, Costa et al. 2012) versus 3 (Deen et al. 2011; Chang et al. 2012) or more clusters (Kurth et al. 2010; Power et al. 2011; Yeo et al. 2011). As introduced by Chang et al., the hierarchical clustering performed on the ROI-wise analysis of our 2011 paper moderately supports a tripartite subdivision of the insula. Given a dataset, the determination of the optimal number of parcels is not a trivial task. Several methods for calculating the optimal number of clusters have been proposed; however, none is faultless (Bowman et al. 2004). Differences between the datasets in terms of data acquisition, technical and numerical aspects may lead to a different determination of the correct number of parcels. This point corroborates the results of those papers that employ methods that do not assume an a priori number of critical components, such as the one developed by Cohen et al. (2008) and used in Nelson et al. (2010). Here, the whole-brain connectivity map of each node is examined for large-scale differences between the nodes. The nodes obtained by resting state functional connectivity magnetic resonance imaging (MRI)are compared with task-evoked data in functional magnetic resonance imaging (fMRI) to corroborate the findings from the functional connectivity analyses. Using this method, the authors suggested a pentapartite subdivision of the anterior insular cortex (AI) where 3 clusters are present in the anteriormost AI and 2 in the posteriormost AI. Other papers suggested an insular clusterization involving more than the common bipartite or tripartite scheme (Cohen et al. 2008; Kurth et al. 2010; Power et al. 2011; Yeo et al. 2011; Kelly et al. 2012). It must be noted, however, that the histological and electrophysiological data on humans and monkeys support the bipartite (see, e.g. Rivier and Clarke 1997; Jezzini et al. 2012) as well as the tripartite insular parcellations (see, e.g. Mesulam and Mufson 1982).

A recent and inspiring paper by Kelly et al. (2012) clearly demonstrated this phenomenon; rather than choosing a fixed number of clusters, they reported a number of insular cortex clusters with K between 2 and 15 (cf. Fig. 4). With K = 2 the parcels are very similar to our clusters, with K = 3 then a middle-anterior ventral cluster appears. The coexistence of different parcellation schemes is understandable in the light of the intrinsically hierarchical nature of the resting state brain networks that have been evidenced in neuroimaging studies (Power et al. 2011; Yeo et al. 2011; Doucet et al. 2011), and hypothesized in the theories of Damasio (1994) and Craig (2002, 2009) that posited a hierarchical organization within the insula representing information about the body's homeostatic state. Indeed, in this view, different levels of complexity may be nested in a hierarchical structure that starts from a bipartite insular subdivision and move to more complex parcellation schemes where each of the two previous clusters is further subdivided. That hypothesis finds a partial support in several recent studies such as Nelson et al. (2010), Yeo et al. (2011), Cauda, Torta et al. (2012), Kelly et al. (2012), and Touroutoglou et al. (2012). However, Kelly et al. point out that not all numbers of parcels lead to a high cross-modal agreement and mean consensus. We have summarized these results in Table 1. The cross-modal agreement, the first criterion chosen by Kelly et al. to find the optimal number of clusters, indicates whether a certain parcellation scheme gained the agreement of all the three parcellation modalities employed (resting state fMRI, task-based fMRI, structural covariance); the second criterion is the mean consensus, which is a measure of “the stability with which pairs of voxels were placed in the same cluster, across the 20 data collection sites”: it is evident in Table 1 that in this analysis the only two solutions that have a high cross-modal agreement and mean consensus for the right and left insulae are K = 2 and 9. Given the suggested hierarchical structure of the brain networks (Leech et al. 2012), it can be hypothesized that these two solutions may constitute two different parcellation steps in the same hierarchical structure. Nonetheless, the present knowledge of the insular parcellation may be further improved or modified using higher-resolution imaging data, improved signal-to-noise ratios (Gonzalez-Castillo et al. 2012), alternative imaging modalities, and refined techniques such as the analysis of temporal functional modes described by Smith et al. (2012), which takes into account temporally dynamic relationships among regions.

Table 1

Mean consensus and cross-modal agreement for different number of insular clusters (Kelly et al. 2012). The only two solutions that have a high cross-modal agreement and mean consensus for the right and left insulae are K = 2 and 9.

 Number of insular clusters 10 11 12 
Right insula Cross-modal agreement      
Mean consensus         
Left insula Cross-modal agreement        
Mean consensus        
 Number of insular clusters 10 11 12 
Right insula Cross-modal agreement      
Mean consensus         
Left insula Cross-modal agreement        
Mean consensus        

An interesting future development would be the evaluation of the cross-modal agreement between all the different techniques used to parcellate the insular cortex, such as neurophysiology, histology, fMRI, diffusion tensor imaging, and anatomy.

References

Bowman
FD
Patel
R
Lu
C
Methods for detecting functional classifications in neuroimaging data
Hum Brain Mapp
 , 
2004
, vol. 
23
 (pg. 
109
-
119
)
Cauda
F
Costa
T
Torta
DM
Sacco
K
D'Agata
F
Duca
S
Geminiani
G
Fox
PT
Vercelli
A
Meta-analytic clustering of the insular cortex: characterizing the meta-analytic connectivity of the insula when involved in active tasks
Neuroimage
 , 
2012a
, vol. 
62
 (pg. 
343
-
55
)
Cauda
F
D'Agata
F
Sacco
K
Duca
S
Geminiani
G
Vercelli
A
Functional connectivity of the insula in the resting brain
Neuroimage
 , 
2011
, vol. 
55
 (pg. 
8
-
23
)
Cauda
F
Torta
DM
Sacco
K
D'Agata
F
Geda
E
Duca
S
Geminiani
G
Vercelli
A
Functional anatomy of cortical areas characterized by Von Economo neurons
Brain Struct Funct
 , 
2012b
Chang
LJ
Yarkoni
T
Khaw
MW
Sanfey
AG
Decoding the role of the insula in human cognition: functional parcellation and large-scale reverse inference
Cereb Cortex
 , 
2012
 
[Epub ahead of print] PubMed PMID: 22437053
Cloutman
LL
Binney
RJ
Drakesmith
M
Parker
GJ
Lambon Ralph
MA
The variation of function across the human insula mirrors its patterns of structural connectivity: evidence from in vivo probabilistic tractography
Neuroimage
 , 
2012
, vol. 
59
 (pg. 
3514
-
3521
)
Cohen
AL
Fair
DA
Dosenbach
NU
Miezin
FM
Dierker
D
Van Essen
DC
Schlaggar
BL
Petersen
SE
Defining functional areas in individual human brains using resting functional connectivity MRI
Neuroimage
 , 
2008
, vol. 
41
 (pg. 
45
-
57
)
Craig
AD
How do you feel? Interoception: the sense of the physiological condition of the body
Nat Rev Neurosci
 , 
2002
, vol. 
3
 (pg. 
655
-
666
)
Craig
AD
How do you feel now? The anterior insula and human awareness
Nat Rev Neurosci
 , 
2009
, vol. 
10
 (pg. 
59
-
70
)
Damasio
AR
Descartes’ error: emotion, reason, and the human brain
 , 
1994
New York
Putnam
Deen
B
Pitskel
NB
Pelphrey
KA
Three systems of insular functional connectivity identified with cluster analysis
Cereb Cortex
 , 
2011
, vol. 
21
 (pg. 
1498
-
1506
)
Doucet
G
Naveau
M
Petit
L
Delcroix
N
Zago
L
Crivello
F
Jobard
G
Tzourio-Mazoyer
N
Mazoyer
B
Mellet
E
, et al.  . 
Brain activity at rest: a multiscale hierarchical functional organization
J Neurophysiol
 , 
2011
, vol. 
105
 (pg. 
2753
-
2763
)
Gonzalez-Castillo
J
Saad
ZS
Handwerker
DA
Inati
SJ
Brenowitz
N
Bandettini
PA
Whole-brain, time-locked activation with simple tasks revealed using massive averaging and model-free analysis
Proc Natl Acad Sci USA
 , 
2012
, vol. 
109
 (pg. 
5487
-
5492
)
Jakab
A
Molnár
PP
Bogner
P
Béres
M
Berényi
EL
Connectivity-based parcellation reveals interhemispheric differences in the insula
Brain Topogr
 , 
2011
, vol. 
25
 (pg. 
264
-
271
)
Jezzini
A
Caruana
F
Stoianov
I
Gallese
V
Rizzolatti
G
Functional organization of the insula and inner perisylvian regions
Proc Natl Acad Sci USA
 , 
2012
, vol. 
109
 (pg. 
10077
-
10082
)
Kelly
C
Toro
R
Di Martino
A
Cox
CL
Bellec
P
Castellanos
FX
Milham
MP
A convergent functional architecture of the insula emerges across imaging modalities
Neuroimage
 , 
2012
, vol. 
61
 (pg. 
1129
-
1142
)
Kurth
F
Zilles
K
Fox
PT
Laird
AR
Eickhoff
SB
A link between the systems: functional differentiation and integration within the human insula revealed by meta-analysis
Brain Struct Funct
 , 
2010
, vol. 
214
 (pg. 
519
-
534
)
Leech
R
Braga
R
Sharp
DJ
Echoes of the brain within the posterior cingulate cortex
J Neurosci
 , 
2012
, vol. 
32
 (pg. 
215
-
222
)
Mesulam
MM
Mufson
EJ
Insula of the old world monkey. I. Architectonics in the insulo-orbito-temporal component of the paralimbic brain
J Comp Neurol
 , 
1982
, vol. 
212
 (pg. 
1
-
22
)
Nelson
SM
Dosenbach
NU
Cohen
AL
Wheeler
ME
Schlaggar
BL
Petersen
SE
Role of the anterior insula in task-level control and focal attention
Brain Struct Funct
 , 
2010
, vol. 
214
 (pg. 
669
-
680
)
Power
JD
Cohen
AL
Nelson
SM
Wig
GS
Barnes
KA
Church
JA
Vogel
AC
Laumann
TO
Miezin
FM
Schlaggar
BL,
, et al.  . 
Functional network organization of the human brain
Neuron
 , 
2011
, vol. 
72
 (pg. 
665
-
678
)
Rivier
F
Clarke
S
Cytochrome oxidase, acetylcholinesterase, and NADPH-diaphorase staining in human supratemporal and insular cortex: evidence for multiple auditory areas
Neuroimage
 , 
1997
, vol. 
6
 (pg. 
288
-
304
)
Smith
SM
Miller
KL
Moeller
S
Xu
J
Auerbach
EJ
Woolrich
MW
Beckmann
CF
Jenkinson
M
Andersson
J
Glasser
MF
, et al.  . 
Temporally-independent functional modes of spontaneous brain activity
Proc Natl Acad Sci USA
 , 
2012
, vol. 
109
 (pg. 
3131
-
3136
)
Taylor
KS
Seminowicz
DA
Davis
KD
Two systems of resting state connectivity between the insula and cingulate cortex
Hum Brain Mapp
 , 
2009
, vol. 
30
 (pg. 
2731
-
2745
)
Touroutoglou
A
Hollenbeck
M
Dickerson
BC
Feldman Barrett
L
Dissociable large-scale networks anchored in the right anterior insula subserve affective experience and attention
Neuroimage
 , 
2012
, vol. 
60
 (pg. 
1947
-
1958
)
Yeo
BT
Krienen
FM
Sepulcre
J
Sabuncu
MR
Lashkari
D
Hollinshead
M
Roffman
JL
Smoller
JW
Zollei
L
Polimeni
JR,
, et al.  . 
The organization of the human cerebral cortex estimated by intrinsic functional connectivity
J Neurophysiol
 , 
2011
, vol. 
106
 (pg. 
1125
-
1165
)