IEEE Signal Processing - May 2018 - 95

EEG/MEG source connectivity analysis. This issue is in line
with a few recent attempts to evaluate other parameters
involved in EEG source connectivity, such as inverse/connectivity measures, the number of scalp electrodes, the head
model, the toolboxes used to perform the analysis, and intra-/
intersubject reproducibility of the identified networks [14], [38],
[55]. The network measures and other issues, e.g., the preprocessing techniques, should also necessarily be the subjects of
further validation and investigation. More precisely, the field
needs studies that thoroughly evaluate graph theoretic approaches in combination with different inverse solution and
connectivity measures.

Acknowledgments

Conclusions

Mahmoud Hassan (mahmoud.hassan@univ-rennes1.fr)
received his B.S. degree in biomedical engineering from the
Islamic University of Lebanon in 2007, and his M.Sc. and
Ph.D. degrees in biomedical engineering from the University of
Technology of Compiègne, France, and Reykjavik University,
Iceland, in 2008 and 2011, respectively. He is currently a postdoctoral researcher at the Signal and Image Processing Laboratory, French Institute of Health and Medical Research,
University of Rennes. His research interests include biosignal
processing and brain network analysis.
Fabrice Wendling (fabrice.wendling@univ-rennes1.fr)
received his B.S. degree in biomedical engineering from the
University of Technology of Compiègne, France, in 1989, his
M.Sc. degree from the Georgia Institute of Technology,
Atlanta, in 1991, and his Ph.D. degree from the University of
Rennes, France, in 1996. He is currently the director of
research at the French Institute of Health and Medical
Research, and he heads the team Epileptogenic System:
Signals and Models at the Signal and Image Processing
Laboratory, Rennes, France. He has been conducting research
on brain signal processing and modeling for more than 20
years in close collaboration with clinicians. In 2012, he
received the Prix Michel Montpetit Award from the French
Academy of Science. He has coauthored approximately 120
peer-reviewed articles.

As long as there is technological progress in EEG systems and
there are advances in signal processing, there will always be
new information to extract from EEG. In this article, we presented one of the latest advances in identifying brain networks,
with high spatiotemporal resolution from dense-EEG recordings: EEG source connectivity. We provided an overview of
this approach and presented the main processing aspects of a
signal problem consisting in estimating brain networks at the
level of neuronal sources from surface EEG recordings.
We also reviewed applications of this new neuroimaging
technique within the context of normal brain functions and
brain disorders. However, this review has not been exhaustive.
The emphasis has been placed on the fundamental aspects of a
new neuroimaging technique that provides a good time/space
resolution for identifying functional brain networks. A number
of issues have not been addressed, as our intent was to provide a didactic guide for researchers interested in EEG source
connectivity. By pointing out some methodological issues, our
intent was also to help these researchers to choose/design the
methods capable of extracting relevant information from EEG
data in a given application context.
In terms of the future, the signal processing community
is directly involved with new advances mainly in the development of fully automatic preprocessing algorithms, more
realistic inverse solutions algorithms, and unbiased effective connectivity measures. Efforts will likely lead to the
development of novel signal processing methods that are
able to assess the dynamics of brain networks on short and
long timescales. At the same time, the rapid progress in the
network analysis community will certainly improve existing
methods for analyzing the brain networks identified from
dense EEG.
Recent trends in open-source neuroimaging data will
undoubtedly accelerate the validation of the current techniques,
such as the huge database of the human connectome project
(HCP) (http://www.humanconnectome.org/). The MEG HCP
data could be used to test new methods and validate existing ones. In addition, the structural connectome from the HCP
(mainly the diffusion tensor imaging data) could certainly be
used as a constraint in the inverse solutions, which could lead
to an improvement of the spatial precision of the identified
functional networks.

We would like to thank Olivier Dufor for helping to record the
real data used in Figure 4, and Isabelle Merlet for helping generate the simulations used in Figures 5 and 6. This work was
supported by a French government grant to CominLabs, a Parisdesignated excellence laboratory, and managed by the French
National Research Agency as part of the Investing for the Future
Program under reference number ANR-10-LABX-07-01. It was
also financed by the University of Rennes Hospital as part of
the COREC Projects, named conneXion (2012-2014) and
BrainGraph (2015-2017).

Authors

References

[1] D. S. Bassett and O. Sporns, "Network neuroscience," Nature Neurosci., vol. 20,
pp. 353-364, Feb. 2017.
[2] A. Fornito, A. Zalesky, and M. Breakspear, "The connectomics of brain disorders," Nature Rev. Neurosci., vol. 16, pp. 159-172, Mar. 2015.
[3] M. Hassan, P. Benquet, A. Biraben, C. Berrou, O. Dufor, and F. Wendling,
"Dynamic reorganization of functional brain networks during picture naming,"
Cortex, vol. 73, pp. 276-288, Dec. 2015.
[4] A. Kabbara, W. El-Falou, M. Khalil, F. Wendling, and M. Hassan, "The dynamic
functional core network of the human brain at rest," Scientific Rep., vol. 7, p. 2936,
June 2017.
[5] P. J. Uhlhaas and W. Singer, "Neural synchrony in brain disorders: Relevance for
cognitive dysfunctions and pathophysiology," Neuron, vol. 52, pp. 155-168, Oct.
2006.
[6] C. Brunner, M. Billinger, M. Seeber, T. Mullen, and S. Makeig, "Volume conduction influences scalp-based connectivity estimates," Frontiers Comput. Neurosci.,
vol. 10, p. 121, Nov. 2016.
[7] F. Van de Steen, L. Faes, E. Karahan, J. Songsiri, P. A. V. Sosa, and D. Marinazzo,
"Critical comments on EEG sensor space dynamical connectivity analysis," Brain
Topog., pp. 1-12, Nov. 2016.
[8] J. M. Schoffelen and J. Gross, "Source connectivity analysis with MEG and
EEG," Human Brain Mapping, vol. 30, pp. 1857-1865, June 2009.

IEEE Signal Processing Magazine

May 2018

http://www.humanconnectome.org/

Table of Contents for the Digital Edition of IEEE Signal Processing - May 2018