Clinical Neurophysiology 128 (2017) 1–3

Contents lists available at ScienceDirect

Clinical Neurophysiology
journal homepage: www.elsevier.com/locate/clinph

Letter to the Editor
Small world brain network characteristics
during EEG Holter recording of a stroke event

1. Introduction
Several studies revealed that cerebral ischemia provokes transient or permanent disruption of functional connections both
locally and distantly from the lesion. Recently, brain connectivity
has been described using graph theory, a mathematical approach
which depicts the brain as a network in order to simplify its complex topology. The human brain consists of complex inhibitory and
excitatory circuits located in specialized areas, which are engaged
in sharing and integrating information with a time-varying interplay at resolution in the millisecond range. Brain functional activity is based on a balance between manifold processes of local
segregation and global integration, quantified by the clustering
coefficient and the path length coefficient, respectively (Bassett
and Bullmore, 2006). A connectivity pattern characterized by high
clustering and short path, known as ‘‘small world” network model
(Watts and Strogatz, 1998), reflects the need of the brain networks
to satisfy the competitive demands of local and global processing.
Recent studies demonstrated that stroke presents functional balance disruption with respect to control subjects (Yin et al., 2014).
In this unique case report, we analyzed an EEG Holter recording
before and during a stroke attack, using graph theory.
2. Materials and methods
2.1. Recordings and pre-processing
A 72-year-old male patient, with a previous history of chronic
hypertension, had been referred to our hospital’s emergency unit
because of sudden onset of confusion and speech production
difficulty. Brain-MRI documented an acute ischemic lesion in the
fronto-basal region of the left medial cerebral artery (MCA) territory, and extracranial ultrasound revealed left internal carotid
artery (ICA) occlusion.
Approximately one week after the first event, the patient came
to our attention for recurrent, transient episodes of severe weakness of the right limbs and inferior facial paresis, with almost complete recovery. A new MRI showed multiple acute and subacute
ischemic lesions in the left MCA territory (frontal lobe, putamen,
temporo-insular region, parietal lobe) with moderate mass effect.
EEG Holter (Fp1, Fp2, O1, O2, T3, C3, C4, T4) was recorded for
evaluating signs of stroke-related epilepsy. After more than 14 h
of recording, the patient presented another stroke attack during
nocturnal sleep, experiencing persistent right facio-brachio-crural
hemiplegia without subsequent recovery.

Holter EEG data were subdivided offline into three 40-minperiods of interest: Baseline (12 h before the event), Stroke I (first
40 min-period from the event), Stroke II (40 min-period immediately following Stroke I).
Data were analyzed with Matlab using scripts based on EEGLAB.
EEG recordings (band-pass 0.2–47 Hz, sampling frequency 256 Hz)
were fragmented in 2-s duration epochs. Detection and rejection of
artifacts were performed through independent component analysis.
A further brain MRI, performed after this last event, showed
enlargement of the recent ischemic lesions in left hemisphere with
extensive involvement of frontal lobe, temporo-insular area and
basal nuclei and increased compression on the left lateral ventricle.
2.2. Functional connectivity
Functional connectivity analysis was obtained by a spectral coherence algorithm of the coupling between two (EEG) signals at any
given frequency (delta (2–4 Hz), theta (4–8 Hz), alpha1 (8–10 Hz),
alpha2 (10–13 Hz), beta1 (13–20 Hz), and beta2 (20–30 Hz)). It was
computed by magnitude squared coherence (mscohere) with homemade Matlab software. In addition, total coherence was calculated
by averaging all possible links between all pairs of contacts in order
to obtain an index of global brain connectivity.
2.3. Graph analysis
A network is a mathematical representation of a real-world
complex system and is defined by a collection of nodes (vertices)
and links (edges) between pairs of nodes. In this study, weighted
and undirected networks were built. Network vertices are the electrode contacts, while edges are weighted by the mscohere value.
Weighted Clustering Cw and Weighted Characteristic Path
length Lw were used to compute the small world coefficient Sw,
defined as the ratio between normalized Cw and Lw, and used to
describe the balance between local connectedness and global integration brain processes.
2.4. Compliance with ethical standards
Ethical approval: All procedures performed in studies involving
human participants were in accordance with the ethical standards
of the institutional and national research committee and with the
1964 Helsinki declaration and its later amendments or comparable
ethical standards. Informed consent was obtained from the participant included in this study.
3. Results
Fig. 1A shows the brain imaging after the event (DWI,
FLAIR, T1, T2⁄ and MR angiography): FLAIR images and

http://dx.doi.org/10.1016/j.clinph.2016.10.090
1388-2457/Ó 2016 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

2

Letter to the Editor / Clinical Neurophysiology 128 (2017) 1–3

Fig. 1. The A part of the figure shows MRI findings after the ischemic stroke event (DWI, FLAIR, T1 and T2⁄ are illustrated). The B part illustrates the EEG signal total
coherence. The C part of the figure reports the percentage variation of Small World connectivity at the beginning and after the stroke event compared to the baseline period.

diffusion-weighted images (DWI) revealed an extensive, recent
ischemic lesion in the left MCA territory involving frontal lobe,
temporo-insular area and basal ganglia; T1-weighted images
and gradient-echo (GRE) T2⁄-weighted sequences showed signs
of partial hemorrhagic transformation in different phases of evolution. A MR Angiography (MRA), previously performed, showed
complete occlusion of the left ICA and collateral flow to the left
MCA, provided by the contralateral ICA via the anterior communicating artery and the vertebrobasilar system via the posterior

communicating artery. Fig. 1B illustrates the EEG signal total
coherence in the different frequency bands for the three different
time periods. A prominent decrease of functional coupling was
observed at the beginning of the stroke event, and a partial recovery near to the end of event, except for the delta band. Fig. 1C
reports the small world percentage change at the beginning and
at the end of the strong event compared to the baseline period.
It shows an evident reduction in the delta band and, to a lesser
extent, an increase in the alpha bands.

Letter to the Editor / Clinical Neurophysiology 128 (2017) 1–3

4. Discussion
The present results, collected in a unique case of EEG recordings
during an ischemic stroke event, confirm that stroke provokes
immediate functional changes of brain networks. The modulation
of EEG rhythms is noted to be frequency-dependent, involving
most evidently the delta and alpha EEG bands. The decrease of
SW in the delta band and increase of SW in the alpha bands are
in line with previous evidence of SW modulation in physiological
and pathological brain networks (Miraglia et al., 2016; Vecchio
et al., 2014, 2016a,b; Caliandro et al., 2016). It could be supposed
that locally reduced and globally increased connectivity in the
alpha rhythm (as a possible interpretation of the increased SW)
might reflect a balance, in the acute phase, between functional
impairment due to acute stroke and an attempt of compensation.
On the other hand, the reduction of delta SW could be seen as a
sort of more ordered network but in a pathological band, namely
an indication of physiological disconnection. In conclusion,
stroke-related acute cerebral network modulation as revealed by
analysis of EEG data could reflect the electrophysiological counterpart of disruptive and adaptive brain processes of integration and
segregation that are well described by graph theory indexes.
Conflict of interest
None.
References
Bassett DS, Bullmore E. Small-world brain networks. Neuroscientist 2006;12:
512–23.
Caliandro P, Vecchio F, Miraglia F, Reale G, Della Marca G, La Torre G, et al. Smallworld characteristics of cortical connectivity changes in acute stroke.
Neurorehabil Neural Repair 2016. pii: 1545968316662525.
Miraglia F, Vecchio F, Bramanti P, Rossini PM. EEG characteristics in ‘‘eyes-open”
versus ‘‘eyes-closed” conditions: small-world network architecture in healthy
aging and age-related brain degeneration. Clin Neurophysiol 2016;127:1261–8.
Vecchio F, Miraglia F, Marra C, Quaranta D, Vita MG, Bramanti P, Rossini PM. Human
brain networks in cognitive decline: a graph theoretical analysis of cortical
connectivity from EEG data. J Alzheimers Dis 2014;41:113–27.

3

Vecchio F, Miraglia F, Piludu F, Granata G, Romanello R, Caulo M, et al. ‘‘Small world”
architecture in brain connectivity and hippocampal volume in Alzheimer’s
disease: a study via graph theory from EEG data. Brain Imaging Behav 2016.
http://dx.doi.org/10.1007/s11682-016-9528-3.
Vecchio F, Miraglia F, Quaranta D, Granata G, Romanello R, Marra C, et al. Cortical
connectivity and memory performance in cognitive decline: a study via graph
theory from EEG data. Neuroscience 2016b;316:143–50.
Watts DJ, Strogatz SH. Collective dynamics of ‘small-world’ networks. Nature
1998;393:440–2.
Yin D, Song F, Xu D, Sun L, Men W, Zang L, et al. Altered topological properties of the
cortical motor-related network in patients with subcortical stroke revealed by
graph theoretical analysis. Hum Brain Mapp 2014;35:3343–59.

⇑

Fabrizio Vecchio
Brain Connectivity Laboratory, IRCCS San Raffaele Pisana, Rome, Italy
* Corresponding author at: Brain Connectivity Laboratory, IRCCS San
Raffaele Pisana, Via Val Cannuta, 247, 00166 Rome, Italy.
E-mail addresses: fabrizio.vecchio@uniroma1.it,
fabrizio.vecchio@sanraffaele.it
Francesca Miraglia
Brain Connectivity Laboratory, IRCCS San Raffaele Pisana, Rome, Italy
Institute of Neurology, Dept. Geriatrics, Neuroscience & Orthopedics,
Catholic University, Policlinic A. Gemelli, Rome, Italy
Angela Romano
Institute of Neurology, Dept. Geriatrics, Neuroscience & Orthopedics,
Catholic University, Policlinic A. Gemelli, Rome, Italy
Placido Bramanti
IRCCS Centro Neurolesi Bonino-Pulejo, Messina, Italy
Paolo Maria Rossini
Brain Connectivity Laboratory, IRCCS San Raffaele Pisana, Rome, Italy
Institute of Neurology, Dept. Geriatrics, Neuroscience & Orthopedics,
Catholic University, Policlinic A. Gemelli, Rome, Italy
Available online 3 November 2016