Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory

Clinical Neurophysiology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

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

Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory Fabrizio Vecchio a,⇑, Francesca Miraglia a, Giuseppe Curcio b, Giacomo Della Marca c, Catello Vollono c, Edoardo Mazzucchi c, Placido Bramanti d, Paolo Maria Rossini a,c a

Brain Connectivity Laboratory, IRCCS San Raffaele Pisana, Rome, Italy Department of Life, Health and Environmental Sciences, L’Aquila, Italy Institute of Neurology, Department of Geriatrics, Neuroscience and Orthopedics, Catholic University, Policlinic A. Gemelli, Rome, Italy d IRCCS Centro Neurolesi Bonino-Pulejo, Messina, Italy b c

a r t i c l e

i n f o

Article history: Accepted 17 September 2014 Available online xxxx Keywords: Graph theory Fronto-temporal epilepsy Functional connectivity EEG Alpha band eLORETA

h i g h l i g h t s  Effective connectivity and optimal network structure is essential for proper information processing in

the brain.  Functional abnormalities of the brain are found to be associated with the pathological changes in con-

nectivity and network structures.  Aim of the present study, was to explore the interictal network properties of EEG signals from tem-

poral lobe structures in the context of fronto-temporal lobe epilepsy by graph analysis tools.

a b s t r a c t Objective: It is believed that effective connectivity and optimal network structure are essential for proper information processing in the brain. Indeed, functional abnormalities of the brain are found to be associated with pathological changes in connectivity and network structures. The aim of the present study was to explore the interictal network properties of EEG signals from temporal lobe structures in the context of fronto-temporal lobe epilepsy. Methods: To complete this aim, the graph characteristics of the EEG data of 17 patients suffering from focal epilepsy of the fronto-temporal type, recorded during interictal periods, were examined and compared in terms of the affected versus the unaffected hemispheres. EEG connectivity analysis was performed using eLORETA software in 15 fronto-temporal regions (Brodmann Areas BAs 8, 9, 10, 11, 20, 21, 22, 37, 38, 41, 42, 44, 45, 46, 47) on both affected and unaffected hemispheres. Results: The evaluation of the graph analysis parameters, such as ‘global’ (characteristic path length) and ‘local’ connectivity (clustering coefficient) showed a statistically significant interaction among side (affected and unaffected hemisphere) and Band (delta, theta, alpha, beta, gamma). Duncan post hoc testing showed an increase of the path length in the alpha band in the affected hemisphere with respect to the unaffected one, as evaluated by an inter-hemispheric marker. The affected hemisphere also showed higher values of local connectivity in the alpha band. In general, an increase of local and global graph theory parameters in the alpha band was found in the affected hemisphere. It was also demonstrated that these effects were more evident in drug-free patients than in those undergoing pharmacological therapy. Conclusions: The increased measures in the affected hemisphere of both functional local segregation and global integration could result from the combination of overlapping mechanisms, including reactive neuroplastic changes seeking to maintain constant integration and segregation properties. Significance: This reactive neuroplastic mechanism seeking to maintain constant integration and segregation properties seems to be more evident in the absence of antiepileptic treatment. Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

⇑ Corresponding author at: Brain Connectivity Laboratory, IRCCS San Raffaele Pisana, Via Val Cannuta, 247, 00166 Rome, Italy. Tel.: +39 06 52253767. E-mail addresses: [email protected], [email protected] (F. Vecchio). http://dx.doi.org/10.1016/j.clinph.2014.09.019 1388-2457/Ó 2014 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

2

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

1. Introduction Epilepsy is a common neurological disorder, characterized by a sudden occurrence of paroxysmal neuronal firing. It is sometimes accompanied (when several causes occur simultaneously, including paroxysmal activity that is highly synchronized, sufficiently prolonged in time and involves a critical neuronal assembly) by clinically evident epileptic attack. It is the most frequently occurring disease of the central nervous system, affecting approximately 1% of the world population. Despite enormous research efforts, the pathogenesis of epilepsy has not completely been elucidated (Timofeev and Steriade, 2004), which hampers both full understanding of the pathophysiology and subsequent treatment. The clinical diagnosis of epilepsy is based on the criteria of the International League Against Epilepsy (ILAE). A diagnostic interictal electro-encephalogram (EEG) showing ‘interictal epileptiform discharges’ (IEDs) is obtainable. Unfortunately, while visual EEG inspection is highly specific as a diagnostic tool, it has a relatively low sensitivity, since only 30–50% of patients have IEDs during their first routine EEG (King et al., 1998). Even though this percentage increases with repeated EEG recordings, between 2% and 18% of patients never display IEDs on their EEGs (Marsan and Zivin, 1970; Noachtar and Remi, 2009). To make matters worse, approximately 0.5% of the healthy population shows IEDs that never lead to a clinically evident epileptic attack (Robin et al., 1978; Gregory et al., 1993). Thus, the development of an EEG measure expanding the diagnostic yield of IEDs, whilst preserving high specificity, would be highly valuable. A relatively new concept in neuroscience is ‘‘functional connectivity’’. Functional connectivity in human neuroscience refers to the synchrony of activity in anatomically distinct but functionally collaborating brain regions. For this reason, if two neuronal assemblies are highly correlated in their rhythmic firing activity over time, they are considered functionally connected. This notion refers to the statistical interdependencies (or synchronization) between time series from different brain areas, as measured by electroencephalography (EEG), magnetoencephalography (MEG), or functional magnetic resonance imaging (fMRI). Synchronization of neuronal discharges on one side may be pivotal for optimal brain functioning (Varela et al., 2001). However, it can also reflect abnormal dynamics of hyper-synchronous firing related to epilepsy (Douw et al., 2010). Within this theoretical framework, focal epilepsy is increasingly seen as a ‘network disorder’ (Kramer and Cash, 2012; Richardson, 2012; Engel et al., 2013). During the genesis of partial seizures (particularly temporal lobe seizures), it has been observed that the EEG rhythms from involved brain networks are characterized by increased synchronization culminating at the end with a clinical seizure (Lieb et al., 1987; Duckrow and Spencer, 1992; Gotman and Levtova, 1996; Le Van et al., 1998; Bartolomei et al., 2001, 2004, 2005; Schindler et al., 2007). In contrast, few studies have investigated network properties and functioning during the interictal period. An increase of EEG synchrony has been described from cortical surface/grids recordings (Schevon et al., 2007) or from intracerebral recordings in mesial temporal lobe epilepsy (Bettus et al., 2008). In this context, an approach to the characterization of complex networks is the use of the ‘graph theory’ (Strogatz, 2001; Boccaletti and Pecora, 2006). A graph is a representation of a network, which is expressed by its nodes (‘vertices’) and connections (‘edges’). Graphs can be described by several parameters and particularly by a clustering coefficient (C) and characteristic path length (L). The clustering coefficient is a measure for the local connectedness of the graph, whereas the characteristic path length is an indicator of overall connectedness. It has been shown that graphs with many local connections and a few random long distance connections are

characterized by a high clustering coefficient and a short characteristic path length (Watts and Strogatz, 1998). These networks, which acts as intermediaries between an ordered and a random organization, have been defined as ‘‘small world networks’’. Such a topology is responsible for high local and global efficiency with low energy and wiring costs (Achard and Bullmore, 2007). Neuronal networks behave as a small world phenomenon, which is also an optimal organization for time-varying dynamic synchronization of neuronal activity among different brain regions (Lago-Fernandez et al., 2000). Graph analysis of structural/anatomical (diffusion MRI and cortical thickness correlation) and functional (fMRI signals and MEG recordings) data have demonstrated a small world configuration in the healthy human brain (Sporns et al., 2000, 2004; Stam, 2004; Sporns and Zwi, 2004; Salvador et al., 2005; Achard and Bullmore, 2007; He et al., 2007; Hagmann et al., 2008; Gong et al., 2009). These small-world properties would be responsible for the high efficiency of the brain information processing, or the efficiency of such an organization being related to cognitive performance (van den Heuvel et al., 2009; Bassett et al., 2009). Along the same lines, alterations of small-world properties have been observed in several brain diseases, shedding light both on their pathophysiology and their behavioral/cognitive consequences (Reijneveld et al., 2007; Bassett et al., 2009; D’Amelio and Rossini, 2012). In the context of epilepsy, changes in network topology were first described during the ictal period (Ponten et al., 2007; Kramer et al., 2008, 2010; Schindler et al., 2008). More recently, research investigations focused on the interictal period and changes in the graph topology of EEG signals (Chavez et al., 2010; Liao et al., 2010; Horstmann et al., 2010; Bernhardt et al., 2011; Vaessen et al., 2012). Results are not homogeneous. Some studies have reported an increase in clustering and a path length shortening (Chavez et al., 2010; Horstmann et al., 2010; Bernhardt et al., 2011). Others have found a decrease in these network properties (Liao et al., 2010) or a decreased clustering and an increased path length (Vaessen et al., 2012). These discrepancies are probably related to different populations studied at different times (i.e., initial or chronic epilepsies), conditions (i.e., under antiepileptic or drug-free conditions) and with different methodological approaches. To our knowledge, no previous report has investigated the network’s properties during the interictal period in a source’s analysis from the EEG scalp recordings in patients with focal frontotemporal epilepsy. The aim of the present study was a proof of concept for the use of graph theory in investigating the interictal network properties of EEG signals—namely those from epileptic temporal lobe structures—in the context of fronto temporal lobe epilepsy. To this end, the graph characteristics of scalp EEG signals recorded during interictal periods were examined and compared in both hemispheres with respect to an inter-hemispheric marker of healthy subjects. A corollary endpoint of the present study was also to investigate whether interictal EEGs from patients suffering from MTLE had differing graph characteristics in drug-free or in chronic antiepileptic drugs treatment conditions.

2. Materials and methods 2.1. Participants A dataset of 17 patients with focal fronto-temporal epilepsy (divided according to the side of focus: 8 left and 9 right) and 48 age-matched healthy subjects was analyzed. Demographic data of the patients are reported in Table 1. To obtain a more reliable

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

3

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx Table 1 Demographic data of patients.

Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj Subj

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Age at the EEG recording

Focus hemispheric localization

Anti-epileptic therapy

Age at the first seizure

30 37 70 43 25 43 22 49 37 73 56 44 33 55 32 24 39

Right Right Right Right Right Right Right Right Right Left Left Left Left Left Left Left Left

Valproic acid, carbamazepine Levetiracetam Acid valproic No therapy Carbamazepine, topiramate, phenobarbital Valproic acid No therapy Carbamazepine Valproic acid, carbamazepine, clonazepam, acetazolamide No therapy No therapy No therapy Carbamazepine, lacosamide, lamotrigine Primidone, valproic acid No therapy Valproic acid, lacosamide, lamotrigine Lamotrigine, interrupted at the moment of EEG recording

7 36 6 43 5 39 22 35 21 73 56 44 14 15 4 16 38

statistical analysis, sex and education values were used as covariates in the following analyses to make sure that the small differences in these variables would not modify the statistical results. All subjects were right-handed, as determined by the Handedness Questionnaire (Salmaso and Longoni, 1985). Informed consent was obtained from each subject, and the study was approved by the local Ethical Committee. Experimental procedures conformed to the Declaration of Helsinki and national guidelines. 2.1.1. Inclusion and exclusion criteria Patients with focal cryptogenic epilepsy and frontotemporal localization of the epileptic focus were included. The diagnosis of focal epilepsy was based on clinical and electroencephalographic data. All patients with parenchymal alterations on brain MRI were excluded. 2.2. Data recordings and preprocessing The EEG recording was performed at rest, with closed eyes no task conditions (at least 5 min), and while the subject was seated and relaxed in a sound-attenuated and dimly lit room. Electroencephalographic signals were measured from 19 electrodes (Fp1, Fp2, F7, F8, F3, F4, T3, T4, C3, C4, T5, T6, P3, P4, O1, O2, Fz, Cz and Pz) positioned according to the International 10–20 system. The monitoring of the eye movements was obtained from two different channels: vertical and horizontal EOGs. Skin/electrode impedances were lowered below 5 KX. Data were analyzed with Matlab R2011b software (Math Works, Natick, MA) and using scripts based on the EEGLAB 11.0.5.4b toolbox (Swartz Center for Computational Neurosciences, La Jolla, CA; http://www.sccn.ucsd.edu/eeglab). The EEG recordings were band-pass filtered from 0.1 to 47 Hz using a finite impulse response (FIR) filter, and the sampling rate frequency was set up at 128 Hz. To eliminate interference caused by ocular, muscular, cardiac and other types of artifacts, imported data were fragmented in 2 s duration epochs and used two processes. First, the data were reviewed, and the epochs with aberrant waveforms or with evident epileptiform activity were manually discarded by an expert EEGer. Second, detection and rejection of artifacts were completed through independent component analysis (ICA) using the Infomax ICA algorithm (Bell and Sejnowski, 1995), as implemented in the EEGLAB. ICA is a blind source decomposition algorithm that enables the separation of statistically independent sources from multichannel data. It has been proposed as an effective method for separating ocular movements and blink artifacts from EEG data (Jung et al., 2000; Iriarte et al., 2003; Hoffmann

and Falkenstein, 2008). The components were visually inspected. If artifact contamination was found, they were manually rejected by the investigator. For the analyzed epileptic group, the average number of artifact-free trials used, with standard error, was 138 (±13.33 SE). 2.3. Functional connectivity analysis EEG connectivity analysis has been performed using the eLORETA exact low resolution electromagnetic tomography (PascualMarqui et al., 2011; Vecchio et al., 2014a,b,c) software, provided by Roberto Pascual-Marqui/The KEY Institute for Brain-Mind Research University Hospital of Psychiatry, Zurich (http:// www.uzh.ch/keyinst/NewLORETA/LORETA01.htm). The eLORETA algorithm is a linear inverse solution for EEG signals that has no localization error to point sources under ideal (noise-free) conditions (Pascual-Marqui, 2002). Based on the scalp-recorded electric potential distribution, the exact low resolution brain electromagnetic tomography (eLORETA) software (free academic software is publically available at http:// www.uzh.ch/keyinst/loreta.htm) was used to compute the cortical three-dimensional distribution of current density. The eLORETA method is a discrete, three-dimensionally (3D) distributed linear, weighted minimum norm inverse solution. The particular weights used in eLORETA endow the tomography with the property of exact localization to test point sources, yielding images of current density with exact localization, albeit with low spatial resolution (i.e., neighboring neuronal sources will be highly correlated). The description of the method and the proof of its exact zero-error localization property are described in Pascual-Marqui (2009). It is also important to emphasize that eLORETA has no localization bias, even in the presence of structured noise. It should be emphasized that the localization properties of any linear 3D inverse solution (i.e., tomography) can always be determined by localization errors to test point sources. If such a tomography has zero localization error to such point sources located anywhere in the brain, then, except in the case of low spatial resolution, the tomography will correctly localize any arbitrary 3D distribution. This is due to the principles of linearity and superposition. In the case of the standardized version sLORETA (PascualMarqui, 2002), it is worth emphasizing that two independent groups, Greenblatt et al. (2005) and Sekihara et al. (2005), showed that the method has no localization bias in the absence of measurement noise. Several results also validate eLORETA, due to its improved localization properties. It is worth emphasizing that deep structures, such as the anterior cingulate cortex (Pizzagalli

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

4

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

et al., 2001) and mesial temporal lobes (Zumsteg et al., 2006), can be correctly localized with these methods. In the current implementation of eLORETA, computations were made in a realistic head model (Fuchs et al., 2002) using the MNI152 template (Mazziotta et al., 2001), with the threedimensional solution space restricted to cortical gray matter, as determined by the probabilistic Talairach atlas (Lancaster et al., 2000). The standard electrode positions on the MNI152 scalp were taken from Oostenveld and Praamstra (2001) and Jurcak et al. (2007). The intracerebral volume is partitioned in 6239 voxels at a 5 mm spatial resolution. Thus, eLORETA images represent the electric activity at each voxel in the neuroanatomic Montreal Neurological Institute (MNI) space as the exact magnitude of the estimated current density. Anatomical labels as Brodmann areas are also reported using MNI space, with correction to Talairach space (Brett et al., 2002). To obtain a topographic view, brain connectivity was computed with sLORETA/eLORETA software in 15 fronto-temporal regions (Brodmann Areas BAs 8, 9, 10, 11, 20, 21, 22, 37, 38, 41, 42, 44, 45, 46, 47) on the affected and unaffected hemispheres (30 ROIs in total). Therefore, for each subject, the connectivity analysis was carried out once for the right hemisphere (affected or unaffected) and once for the left one (affected or unaffected). Regions of Interest (ROIs) are needed for the estimation of the electric neuronal activity that is used to analyze brain functional connectivity. No general rules for constructing the ROIs are available. In order to assess functional connectivity between all major areas, the cortex areas under the 19 head surface electrode locations Fp1/2, F7/8, F3/4, Fz, C3/4, Cz, T3/T4, T5/6, P3/4, Pz, O1/2 of the international 10/20 system were used. The signal at each cortical ROI consisted of the average electric neuronal activities of all voxels belonging to that ROI, as computed with eLORETA. For each hemisphere, among the eLORETA current density time series of the 15 ROIs, intracortical Lagged Linear Coherence, extracted using the ‘‘all nearest voxels’’ method (Pascual-Marqui, 2007a; Pascual-Marqui et al., 2011), was computed between all possible pairs of the 15 ROIs for each of the five independent EEG frequency bands of delta (2–4 Hz), theta (4–8 Hz), alpha (8– 13 Hz), beta (13–30 Hz), and gamma (30–45 Hz) for each subject. The well known definition for the complex valued coherence (Nolte et al., 2004) between time series x and y in the frequency band x is:

r xyx ¼

ReCovðx; yÞ þ iImCovðx; yÞ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi VarðxÞ  VarðyÞ

ð1Þ

which is based on the cross-spectrum given by the covariance pffiffiffiffiffiffiffi and variances of the signals, where i is the imaginary unit ( 1).The squared modulus of the coherence is:

r 2xyx ¼

½ReCovðx; yÞ2 þ ½ImCovðx; yÞ2 VarðxÞ  VarðyÞ

ð2Þ

and the lagged coherence (Pascual-Marqui, 2007b; Pascual-Marqui et al., 2011b) is:

LagR2xyx ¼

½ImCovðx; yÞ2 VarðxÞ  VarðyÞ  ½ReCovðx; yÞ2

ð3Þ

Starting with the definition of the complex valued coherence (Nolte et al., 2004; Lehmann et al., 2012) between time series x and y in the frequency band x—which is based on the cross-spectrum given by the covariance and variances of the signals—the lagged linear coherence in the frequency band x is reported in accordance with the following equation (Pascual-Marqui et al., 2011):

LagR2xyx ¼

½ImCovðx; yÞ2 VarðxÞ  VarðyÞ  ½ReCovðx; yÞ2

ð4Þ

where Var and Cov are variances and covariance of the signals. It was developed as a measure of true physiological connectivity not affected by volume conduction and low spatial resolution (Pascual-Marqui, 2007a; Pascual-Marqui et al., 2011). The values of lagged linear connectivity computing between all pairs of ROIs for each frequency band were used as weights of the networks built in the graph analysis. A weighted network was built based on the connectivity between ROIs. The nodes of the network are ROIs, and the edges of the network are weighed by the lagged linear coherence values. 2.4. Graph analysis A network is a mathematical representation of a real-world complex system. It is defined by a collection of nodes (vertices) and links (edges) between pairs of nodes. Nodes in large-scale brain networks usually represent brain regions, while links represent anatomical, functional or effective connections (Friston, 1994), depending on the dataset. Anatomical connections typically correspond to white matter fiber tracts between pairs of gray matter (cortical areas or subcortical relays) brain regions. Functional connections correspond to magnitudes of temporal correlations in activity and may occur between pairs of anatomically unconnected regions. A weighted graph is a mathematical representation of a set of elements (vertices) that may be linked through connections of variable weights (edges). In the present study, weighted and undirected networks were built. The vertices of the network are the estimated cortical sources in the BAs, and the edges are weighted by the Lagged Linear value within each pair of vertices. The software instrument used here for the graph analysis was the Brain Connectivity Toolbox (BCT, http:// www.brain-connectivity-toolbox.net/), adapted with our own Matlab scripts. The clustering (C) around vertex i is quantified by the number of triangles in which vertex i participates normalized by the maximum possible number of such triangles. C is defined as (Onnela et al., 2005; Rubinov and Sporns, 2010):

C¼

1X 1X 2t i Ci ¼ n iN n iN ki ðki  1Þ

ð5Þ

where Ci is the clustering coefficient of node i (Ci = 0 for ki < 2), ki the degree of node i and ti the number of triangles around the node i. This quantity is normalized between 0 and 1, and it characterizes the tendency of the nearest neighbors of node i to be interconnected. As triangles are one type of subgraph, the definition of C may be used to yield the weighted Clustering coefficient Cw by replacing the number of triangles ti in Eq. (5) with the sum of triangle intensities (Onnela et al., 2005; Rubinov and Sporns, 2010).

Cw ¼

1 X 2tw i n i2N ki ðki  1Þ

ð6Þ

Where tw i ,

tw i ¼

1X ðwij wih wjh Þ1=3 2 j;h2N

ð7Þ

is the geometric mean of triangles around i. wij are connection weights associated with links (i,j), assuming that weights are normalized, such that 0 6 wij 6 1 for all i and j (Onnela et al., 2005; Rubinov and Sporns, 2010). The mean clustering coefficient is computed for all nodes of the graph and is then averaged. It is a measure for the tendency of network elements to form local clusters (de Haan et al., 2009).

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

5

Fig. 1. ANOVA interaction of the inter-hemispheric characteristic path length among the factors Group (healthy subjects, left fronto-temporal epilepsy patients, and right fronto-temporal epilepsy patients) and Band (delta, theta, alpha, beta, gamma).

Fig. 2. ANOVA interaction of the inter-hemispheric clustering coefficient among the factors Group (healthy subjects, left fronto-temporal epilepsy patients, and right frontotemporal epilepsy patients) and Band (delta, theta, alpha, beta, gamma).

The Characteristic Path Length of the network is defined (Onnela et al., 2005; Rubinov and Sporns, 2010):

L¼

1X 1X Li ¼ n iN n iN

P

jN;j–i dij

n1

ð8Þ

where Li is the average of distances dij between node i and all other nodes. Weighted Characteristic Path length Lw is defined (Onnela et al., 2005; Rubinov and Sporns, 2010):

1X L ¼ n i2N w

P

w j2N;j–i dij

n1

ð9Þ 2.5. Statistical evaluation

with w

dij ¼

X

auv 2 g w i$j f ðwuv Þ

which is associated with a low level of local clustering, and regular networks or lattices, the high-level of clustering of which is accompanied by a long path length. In order to evaluate the specificity of the results, two control analyses, similar to those of the main analysis, were made, comparing affected and unaffected hemispheres. In the first one, the graph theory parameters were evaluated in the central area (Brodmann Areas BAs 1, 2, 3, 4, 6), an area very close to the localization of the epileptic focus. In the second one, we selected a region (Brodmann Areas BAs 6, 19, 30, 32, 36, 39, 40, 43) circumscribing the site of the epileptic focus.

ð10Þ

which represents the shortest weighted path length between i and j. f is a map (e.g., an inverse) from weight to length, and gw i Mj is the shortest weighted path between i and j. Cw and Lw were used to compute the small world coefficient that we used to describe brain network organization. The measure of network small-worldness (Sw) is defined (Onnela et al., 2005; Rubinov and Sporns, 2010) as the ratio between normalized C (weighted clustering coefficients) and L (weighted characteristic path lengths). The Sw coefficient is used to describe the balance between the local connectedness and the global integration of a network. When Sw is larger than 1, a network is said to have small-world properties. Small-world organization is an intermediary between random networks, the short overall path length of

eLORETA statistical evaluation was made on a graph analysis pattern extracted with sLORETA/eLORETA from the 30 ROIs (15 ROIs of the affected and 15 ROIs of the unaffected hemisphere). The normality of the transformed data was tested using the Kolmogorov–Smirnov test, and the hypothesis of Gaussianity could not be rejected. The first statistical analysis of variance (ANOVA) was computed in order to evaluate possible differences between the two hemispheres only in the healthy subjects for both characteristic path length (L) and clustering coefficient (C). It was computed between the factors Side (left and right hemisphere) and Band (delta, theta, alpha, beta and gamma). Once it was demonstrated that there was no statistical difference between the two hemispheres, an interhemispheric marker, evaluated as the differences between the left and right graph parameters (namely, L and C separately) was introduced.

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

6

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

Fig. 3. eLORETA connectivity maps for alpha band. Each red tract among the 15 ROIs for each hemisphere strictly refers to the connectivity values higher than the cut-off threshold (0.2) for healthy subjects, left fronto-temporal epilepsy patients, and right fronto-temporal epilepsy patients. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The second set of ANOVA was based on this inter-hemispheric marker for both C and L. It included all subjects and was computed between the factors Group (healthy subjects, left fronto-temporal epilepsy patients, and right fronto-temporal epilepsy patients) and Band (delta, theta, alpha, beta and gamma). ANOVA was chosen, since it is known to be robust with respect to the departure of normality and homoscedasticity of the data being treated. A Greenhouse and Geisser correction was used for protection against the violation of the sphericity assumption in the repeated measure ANOVA. Additionally, a post hoc analysis, with the Duncan’s test and significance level at 0.05, was performed. The statistical analysis was performed with the software Statistica v.7 (StatSoft Inc., www.statsoft.com). In previous studies, the test–retest stability of the graph characteristics was confirmed when repeated measurements in different epochs were carried out in subjects in stable clinical and behavioral conditions.

3. Results 3.1. Graph theory of EEG cortical sources connectivity The first two ANOVAs on the healthy subjects showed no statistical differences between the two hemispheres for both characteristic path length (L) and clustering coefficient (C). Comparing healthy subjects and the two groups of patients, the ANOVA for the evaluation of the global connectivity (characteristic path length L) showed a statistically significant interaction

(F(8,248) = 2.83; p < 0.0050) among Group (healthy subjects, left fronto-temporal epilepsy patients and right fronto-temporal epilepsy patients) and Band (delta, theta, alpha, beta, gamma). Fig. 1 shows the L values relative to this statistical interaction. The ANOVA for the evaluation of the local connectivity (clustering coefficient C) showed a statistically significant interaction (F(8,248) = 2.76; p < 0.0061) among the factors Group (healthy subjects, left fronto-temporal epilepsy patients and right frontotemporal epilepsy patients) and Band (delta, theta, alpha, beta, gamma). Fig. 2 shows the C values relative to this statistical interaction. In both ANOVAs, while healthy subjects present values close to zero, the Duncan post hoc testing showed that left fronto-temporal epilepsy patients presented significantly (p < 0.002 for L and p < 0.019 for C) higher positive values, and the right fronto-temporal epilepsy patients presented significantly (p < 0.008 for L and p < 0.014 for C) more negative ones in alpha band with respect to healthy subjects. Of note, as inter-hemispheric values, these results mean that an increase of both local and global graph theory parameters was found in the alpha band in the affected hemisphere with respect to the unaffected one. It is worth mentioning that such changes were significantly less evident in patients with a long history of epilepsy under pharmacological treatment, when compared do those with a more recent history of epilepsy and drug-free status. In order to evaluate this point, we also evaluated all patients (independently, by the side of the affected hemisphere) after dividing them into two groups: drug-free and chronic treatment. In both ANOVAs, the Duncan post hoc testing showed that fronto-temporal drug-free patients (patients #4,7,10,11,12,15,17) presented significantly—for characteristic path length (p < 0.005 and p < 0.0004 for (L; F(4,56) = 2.96; p < 0.0274) and for clustering coefficient (C; F(4,56) = 2.96; p < 0.0274))—higher inter-hemispheric values than patients undergoing long time therapy. 3.2. EEG cortical sources connectivity as estimated by eLORETA For illustrative purposes, we report in Fig. 3 the eLORETA connectivity maps for the significant alpha frequency for the three groups of subjects (healthy, left and right fronto-temporal epilepsy). Maps illustrate only the connections (among the mentioned 15 ROIs of the fronto-temporal regions for both left and right hemispheres) that resulted in higher than an arbitrary threshold (0.2). This enhances graphic differences, in line with our previous papers. From a visual inspection, it is almost evident an increase of connectivity in the affected hemisphere in the patients (as it is possible to observe in the higher number of red tract connections). Otherwise, the healthy subjects do not present evident inter-hemispheric differences. 3.3. Control analysis results The same ANOVA statistical analysis was performed on the two groups of patients in the two control tests. The results showed that no differences (p = 0.88) between the two hemispheres were present in the central areas (BAs 1, 2, 3, 4, 6). In the same way, the statistical comparison (p = 0.92) between the two hemispheres has suggested that no differences were found when investigating the ‘‘perifocal region’’ (Bas 6, 19, 30, 32, 36, 39, 40, 43). Furthermore, in order to evaluate the measured values (instead of the lateralization differences) two ANOVAs (both illustrated in Fig. 4), one for the Clustering (F(8,36) = 4.78; p < 0.00001) and one for the Path Length (F(8,36) = 3.61; p < 0.0005), were performed. They included the factors Group (healthy—averaging left and right hemisphere—affected and unaffected hemisphere) and Band (delta, theta, alpha, beta and gamma). Both analyses showed

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

7

Fig. 4. Two ANOVAs, one for the Clustering and one for the Path Length, both including the factors Group (healthy – averaging left and right hemisphere-, affected and unaffected hemisphere) and Band (delta, theta, alpha, beta, and gamma). Both analyses showed that the affected hemisphere presented values higher than unaffected and that unaffected showed values higher than healthy hemispheres in alpha band (p < 0.02 to p < 0.003).

that the affected hemisphere presented values higher than the unaffected one and that the unaffected one showed values higher than healthy hemispheres in the alpha band (p < 0.02 to p < 0.003). 4. Discussion The human brain consists of complex inhibitory and excitatory circuits consisting of functionally specialized regions that continuously interact to acquire, share and integrate information. The white-matter (axonal) fibers provide the structural and anatomical basis for signal transfer and communication between the brain subunits. Furthermore, the connections between brain areas are not random but balance segregation and integration characteristics, as revealed by local clustering (segregation) and path length (integration). The brain is in a constant state of dynamic change, for example switching between cognitive and behavioral tasks, movement and rest, wakefulness and sleep, elimination and acquisition of new information. The brains of people with epilepsy display additional features in their dynamic repertoire, particularly regarding the paroxysmal occurrence of firing within networks linking neuronal assemblies. The proof of concept, which triggered the present study, could be considered an extension of a previous study (Bernhardt et al., 2011), in which it was demonstrated that in epilepsy patients both the characteristic path length and clustering increase with respect to healthy, age/sex matched subjects. The present results also shed light on previous EEG pivotal works of our research group, in which we demonstrated that both physiological and pathological aging processes modulate the network configuration of brain connectivity. Here, an increase in graph theory parameters has been observed limitedly for the alpha rhythm in the affected hemisphere with

respect to the non-affected one. Such a circumscribed effect should be discussed in the light of the physiological role of alpha rhythm. Alpha frequencies constitute the leading characteristic of normal EEG activity at waking rest, which is usually defined as the ‘‘idling rhythms’’ of the adult brain (Niedermeyer and da Silva, 2005). Recently, several empirical papers supported the hypothesis that spontaneous alpha activity on the scalp EEG is not simply a resting rhythm. Rather, it is more probably a deterministic chaotic signal with several functional correlates ranging from memory to sensory-motor processing (Schurmann and Basar, 2001). In healthy individuals, alpha rhythm does work as an oscillatory component of brain activity and thus can be interpreted as a basic form of information transmission in the brain (Klimesch, 1999). Moreover, event-related activity studies have shown a positive correlation between alpha frequency and the speed of information processing and between this rhythm and a good cognitive performance (Klimesch, 1999). In sum, a well-designed anatomical network could combine the occurrence of functionally specialized (segregated) modules with a robust number of intermodular (integrating) links. Such a design is commonly termed small-world and indeed appears to be a ubiquitous facet of anatomical connectivity. It is commonly thought that such an organization reflects an optimal balance of functional integration and segregation (Sporns and Honey, 2006). Here, the increased measures in the affected hemisphere of both functional segregation (evaluated by the clustering coefficient as a mathematical representation of the ability for specialized processing to occur within densely interconnected groups of brain regions), and of functional integration (based on the concept of a path as the brain’s ability to rapidly combine specialized information from distributed brain regions), could result from the combination of overlapping mechanisms including reactive neuroplastic changes aiming to

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

8

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx

maintain integration and segregation of the interictal condition, as opposed to hypersynchronization of the spiking state. This hypothesis was also supported by the observation that an effective antiepileptic treatment is combined with a stabilization of the inter-hemispheric parameters when comparing patients under therapy to pharmacologically naïve patients.

5. Conclusions Altogether, both ‘‘global’’ (average shortest path length of the network representing an index of how efficient the information transfer is from one part of the network to another) and ‘‘local’’ (amount of local interconnectedness and network segregation) measures can reveal the cortical network features distinguishing epileptic affected and non-affected hemispheres. This study opens interesting avenues for future researches investigating eventual modifications of brain connectivity in the epochs approximating a clinical seizure, as well as the impact of antiepileptic drugs on pharmacologically naïve patients. Acknowledgements Dr. Francesca Miraglia participated to this study in the framework of her Ph.D. program at the Doctoral School in Neuroscience, Department of Neuroscience, Catholic University of Rome, Italy. The article is partially funded by the Italian Ministry of Instruction, University and Research MIUR (‘‘Approccio integrato clinico e sperimentale allo studio dell’invecchiamento cerebrale e delle malattie neurodegenerative: basi molecolari, epidemiologia genetic, neuroimagnig multimodale e farmacogenetica’’ and ‘‘Functional connectivity and neuroplasticity in physiological and pathological aging’’ PRIN project). Confilct of interest: None of the authors have potential conflicts of interest to be disclosed. References Achard S, Bullmore E. Efficiency and cost of economical brain functional networks. PLoS Comput Biol 2007;3:e17. Bartolomei F, Wendling F, Bellanger JJ, Regis J, Chauvel P. Neural networks involving the medial temporal structures in temporal lobe epilepsy. Clin Neurophysiol 2001;112:1746–60. Bartolomei F, Wendling F, Regis J, Gavaret M, Guye M, Chauvel P. Pre-ictal synchronicity in limbic networks of mesial temporal lobe epilepsy. Epilepsy Res 2004;61:89–104. Bartolomei F, Khalil M, Wendling F, Sontheimer A, Regis J, Ranjeva JP, et al. Entorhinal cortex involvement in human mesial temporal lobe epilepsy: an electrophysiologic and volumetric study. Epilepsia 2005;46:677–87. Bassett DS, Bullmore ET, Meyer-Lindenberg A, Apud JA, Weinberger DR, Coppola R. Cognitive fitness of cost-efficient brain functional networks. Proc Natl Acad Sci USA 2009;106:11747–52. Bell AJ, Sejnowski TJ. An information-maximization approach to blind separation and blind deconvolution. Neural Comput 1995;7:1129–59. Bernhardt BC, Chen Z, He Y, Evans AC, Bernasconi N. Graph-theoretical analysis reveals disrupted small-world organization of cortical thickness correlation networks in temporal lobe epilepsy. Cereb Cortex 2011;21:2147–57. Bettus G, Wendling F, Guye M, Valton L, Regis J, Chauvel P, et al. Enhanced EEG functional connectivity in mesial temporal lobe epilepsy. Epilepsy Res 2008;81:58–68. Boccaletti S, Pecora LM. Introduction: stability and pattern formation in networks of dynamical systems. Chaos 2006;16:015101. Brett M, Johnsrude IS, Owen AM. The problem of functional localization in the human brain. Nat Rev Neurosci 2002;3:243–9. Chavez M, Valencia M, Navarro V, Latora V, Martinerie J. Functional modularity of background activities in normal and epileptic brain networks. Phys Rev Lett 2010;104:118701. D’Amelio M, Rossini PM. Brain excitability and connectivity of neuronal assemblies in Alzheimer’s disease: from animal models to human findings. Prog Neurobiol 2012;99:42–60. de Haan W, Pijnenburg YA, Strijers RL, van der Made Y, van der Flier WM, Scheltens P, et al. Functional neural network analysis in frontotemporal dementia and Alzheimer’s disease using EEG and graph theory. BMC Neurosci 2009;10:101.

Douw L, de GM, van DE, Heimans JJ, Ronner HE, Stam CJ, et al. ‘Functional connectivity’ is a sensitive predictor of epilepsy diagnosis after the first seizure. PLoS One 2010;5:e10839. Duckrow RB, Spencer SS. Regional coherence and the transfer of ictal activity during seizure onset in the medial temporal lobe. Electroencephalogr Clin Neurophysiol 1992;82:415–22. Engel Jr J, Thompson PM, Stern JM, Staba RJ, Bragin A, Mody I. Connectomics and epilepsy. Curr Opin Neurol 2013;26:186–94. Friston KJ. Functional and effective connectivity in neuroimaging: a synthesis. Hum Brain Mapp 1994;2:56–78. Fuchs M, Kastner J, Wagner M, Hawes S, Ebersole JS. A standardized boundary element method volume conductor model. Clin Neurophysiol 2002;113:702–12. Gong G, He Y, Concha L, Lebel C, Gross DW, Evans AC, et al. Mapping anatomical connectivity patterns of human cerebral cortex using in vivo diffusion tensor imaging tractography. Cereb Cortex 2009;19:524–36. Gotman J, Levtova V. Amygdala-hippocampus relationships in temporal lobe seizures: a phase-coherence study. Epilepsy Res 1996;25:51–7. Greenblatt RE, Ossadtchi A, Pflieger ME. Local linear estimators for the bioelectromagnetic inverse problem. IEEE Trans Signal Process 2005;53:3403–12. Gregory RP, Oates T, Merry RT. Electroencephalogram epileptiform abnormalities in candidates for aircrew training. Electroencephalogr Clin Neurophysiol 1993;86:75–7. Hagmann P, Cammoun L, Gigandet X, Meuli R, Honey CJ, Wedeen VJ, et al. Mapping the structural core of human cerebral cortex. PLoS Biol 2008;6:e159. He Y, Chen ZJ, Evans AC. Small-world anatomical networks in the human brain revealed by cortical thickness from MRI. Cereb Cortex 2007;17:2407–19. Hoffmann S, Falkenstein M. The correction of eye blink artefacts in the EEG: a comparison of two prominent methods. PLoS One 2008;3:e3004. Horstmann MT, Bialonski S, Noennig N, Mai H, Prusseit J, Wellmer J, et al. State dependent properties of epileptic brain networks: comparative graphtheoretical analyses of simultaneously recorded EEG and MEG. Clin Neurophysiol 2010;121:172–85. Iriarte J, Urrestarazu E, Valencia M, Alegre M, Malanda A, Viteri C, et al. Independent component analysis as a tool to eliminate artifacts in EEG: a quantitative study. J Clin Neurophysiol 2003;20:249–57. Jung TP, Makeig S, Humphries C, Lee TW, McKeown MJ, Iragui V, et al. Removing electroencephalographic artifacts by blind source separation. Psychophysiology 2000;37:163–78. Jurcak V, Tsuzuki D, Dan I. 10/20, 10/10, and 10/5 systems revisited: their validity as relative head-surface-based positioning systems. Neuroimage 2007;34:1600–11. King MA, Newton MR, Jackson GD, Fitt GJ, Mitchell LA, Silvapulle MJ, et al. Epileptology of the first-seizure presentation: a clinical, electroencephalographic, and magnetic resonance imaging study of 300 consecutive patients. Lancet 1998;352:1007–11. Klimesch W. EEG alpha and theta oscillations reflect cognitive and memory performance: a review and analysis. Brain Res Brain Res Rev 1999;29:169–95. Kramer MA, Cash SS. Epilepsy as a disorder of cortical network organization. Neuroscientist 2012;18:360–72. Kramer MA, Kolaczyk ED, Kirsch HE. Emergent network topology at seizure onset in humans. Epilepsy Res 2008;79:173–86. Kramer MA, Eden UT, Kolaczyk ED, Zepeda R, Eskandar EN, Cash SS. Coalescence and fragmentation of cortical networks during focal seizures. J Neurosci 2010;30:10076–85. Lago-Fernandez LF, Huerta R, Corbacho F, Siguenza JA. Fast response and temporal coherent oscillations in small-world networks. Phys Rev Lett 2000;84:2758–61. Lancaster JL, Woldorff MG, Parsons LM, Liotti M, Freitas CS, Rainey L, et al. Automated Talairach atlas labels for functional brain mapping. Hum Brain Mapp 2000;10:120–31. Le Van QM, Adam C, Baulac M, Martinerie J, Varela FJ. Nonlinear interdependencies of EEG signals in human intracranially recorded temporal lobe seizures. Brain Res 1998;792:24–40. Lehmann D, Faber PL, Tei S, Pascual-Marqui RD, Milz P, Kochi K. Reduced functional connectivity between cortical sources in five meditation traditions detected with lagged coherence using EEG tomography. Neuroimage 2012;60:1574–86. Liao W, Zhang Z, Pan Z, Mantini D, Ding J, Duan X, et al. Altered functional connectivity and small-world in mesial temporal lobe epilepsy. PLoS One 2010;5:e8525. Lieb JP, Hoque K, Skomer CE, Song XW. Inter-hemispheric propagation of human mesial temporal lobe seizures: a coherence/phase analysis. Electroencephalogr Clin Neurophysiol 1987;67:101–19. Marsan CA, Zivin LS. Factors related to the occurrence of typical paroxysmal abnormalities in the EEG records of epileptic patients. Epilepsia 1970;11:361–81. Mazziotta J, Toga A, Evans A, Fox P, Lancaster J, Zilles K, et al. A probabilistic atlas and reference system for the human brain: international consortium for brain mapping (ICBM). Philos Trans R Soc Lond B Biol Sci 2001;356:1293–322. Niedermeyer E, da Silva FL. Electroencephalography: basic principles, clinical applications, and related fields. Lippincott Williams & Wilkins; 2005. Noachtar S, Remi J. The role of EEG in epilepsy: a critical review. Epilepsy Behav 2009;15:22–33. Nolte G, Bai O, Wheaton L, Mari Z, Vorbach S, Hallett M. Identifying true brain interaction from EEG data using the imaginary part of coherency. Clin Neurophysiol 2004;115:2292–307.

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019

F. Vecchio et al. / Clinical Neurophysiology xxx (2014) xxx–xxx Onnela JP, Saramaki J, Kertesz J, Kaski K. Intensity and coherence of motifs in weighted complex networks. Phys Rev E Stat Nonlin Soft Matter Phys 2005;71:065103. Oostenveld R, Praamstra P. The five percent electrode system for high-resolution EEG and ERP measurements. Clin Neurophysiol 2001;112:713–9. Pascual RD. Standardized low-resolution brain electromagnetic tomography (sLORETA): technical details. Methods Find Exp Clin Pharmacol 2002;24(Suppl. D):5–12. Pascual-Marqui RD. Instantaneous and lagged measurements of linear and nonlinear dependence between groups of multivariate time series: frequency decomposition. arXiv preprint arXiv:0711 1455; 2007a. Pascual-Marqui RD. Discrete, 3D distributed, linear imaging methods of electric neuronal activity. Part 1: exact, zero error localization. arXiv preprint arXiv:0710.3341; 2007b. Pascual-Marqui RD. Theory of the EEG inverse problem. In: Quantitative EEG analysis: methods and clinical applications. Artech House; 2009. p. 121–40. Pascual-Marqui RD, Lehmann D, Koukkou M, Kochi K, Anderer P, Saletu B, et al. Assessing interactions in the brain with exact low-resolution electromagnetic tomography. Philos Trans A Math Phys Eng Sci 2011;369:3768–84. Pizzagalli D, Pascual-Marqui RD, Nitschke JB, Oakes TR, Larson CL, Abercrombie HC, et al. Anterior cingulate activity as a predictor of degree of treatment response in major depression: evidence from brain electrical tomography analysis. Am J Psychiatry 2001;158:405–15. Ponten SC, Bartolomei F, Stam CJ. Small-world networks and epilepsy: graph theoretical analysis of intracerebrally recorded mesial temporal lobe seizures. Clin Neurophysiol 2007;118:918–27. Reijneveld JC, Ponten SC, Berendse HW, Stam CJ. The application of graph theoretical analysis to complex networks in the brain. Clin Neurophysiol 2007;118:2317–31. Richardson MP. Large scale brain models of epilepsy: dynamics meets connectomics. J Neurol Neurosurg Psychiatry 2012;83:1238–48. Robin JJ, Tolan GD, Arnold JW. Ten-year experience with abnormal EEGs in asymptomatic adult males. Aviat Space Environ Med 1978;49:732–6. Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and interpretations. Neuroimage 2010;52:1059–69. Salmaso D, Longoni AM. Problems in the assessment of hand preference. Cortex 1985;21:533–49. Salvador R, Suckling J, Coleman MR, Pickard JD, Menon D, Bullmore E. Neurophysiological architecture of functional magnetic resonance images of human brain. Cereb Cortex 2005;15:1332–42. Schevon CA, Cappell J, Emerson R, Isler J, Grieve P, Goodman R, et al. Cortical abnormalities in epilepsy revealed by local EEG synchrony. Neuroimage 2007;35:140–8. Schindler K, Elger CE, Lehnertz K. Increasing synchronization may promote seizure termination: evidence from status epilepticus. Clin Neurophysiol 2007;118:1955–68.

9

Schindler KA, Bialonski S, Horstmann MT, Elger CE, Lehnertz K. Evolving functional network properties and synchronizability during human epileptic seizures. Chaos 2008;18:033119. Schurmann M, Basar E. Functional aspects of alpha oscillations in the EEG. Int J Psychophysiol 2001;39:151–8. Sekihara K, Sahani M, Nagarajan SS. Localization bias and spatial resolution of adaptive and non-adaptive spatial filters for MEG source reconstruction. Neuroimage 2005;25:1056–67. Sporns O, Honey CJ. Small worlds inside big brains. Proc Natl Acad Sci USA 2006;103:19219–20. Sporns O, Zwi JD. The small world of the cerebral cortex. Neuroinformatics 2004;2:145–62. Sporns O, Tononi G, Edelman GM. Connectivity and complexity: the relationship between neuroanatomy and brain dynamics. Neural Netw 2000;13:909–22. Sporns O, Chialvo DR, Kaiser M, Hilgetag CC. Organization, development and function of complex brain networks. Trends Cogn Sci 2004;8:418–25. Stam CJ. Functional connectivity patterns of human magnetoencephalographic recordings: a ‘small-world’ network? Neurosci Lett 2004;355:25–8. Strogatz SH. Exploring complex networks. Nature 2001;410:268–76. Timofeev I, Steriade M. Neocortical seizures: initiation, development and cessation. Neuroscience 2004;123:299–336. Vaessen MJ, Jansen JF, Vlooswijk MC, Hofman PA, Majoie HJ, Aldenkamp AP, et al. White matter network abnormalities are associated with cognitive decline in chronic epilepsy. Cereb Cortex 2012;22:2139–47. van den Heuvel MP, Stam CJ, Kahn RS, Hulshoff Pol HE. Efficiency of functional brain networks and intellectual performance. J Neurosci 2009;29:7619–24. Varela F, Lachaux JP, Rodriguez E, Martinerie J. The brainweb: phase synchronization and large-scale integration. Nat Rev Neurosci 2001;2:229–39. Vecchio F, Miraglia F, Bramanti P, Rossini PM. Human brain networks in physiological aging: a graph theoretical analysis of cortical connectivity from EEG data. J Alzheimers Dis 2014a;41:1239–49. Vecchio F, Miraglia F, Marra C, Quaranta D, Vita MG, Bramanti P, et al. Human brain networks in cognitive decline: a graph theoretical analysis of cortical connectivity from EEG data. J Alzheimers Dis 2014b;41:113–27. Vecchio F, Miraglia F, Valeriani L, Scarpellini MG, Bramanti P, Mecarelli O, et al. Cortical brain connectivity and B-type natriuretic peptide in patients with congestive heart failure. Clin EEG Neurosci 2014. pii: 1550059414529765. [Epub ahead of print]. Watts DJ, Strogatz SH. Collective dynamics of ‘small-world’ networks. Nature 1998;393:440–2. Zumsteg D, Friedman A, Wieser HG, Wennberg RA. Propagation of interictal discharges in temporal lobe epilepsy: correlation of spatiotemporal mapping with intracranial foramen ovale electrode recordings. Clin Neurophysiol 2006;117:2615–26.

Please cite this article in press as: Vecchio F et al. Cortical connectivity in fronto-temporal focal epilepsy from EEG analysis: A study via graph theory. Clin Neurophysiol (2014), http://dx.doi.org/10.1016/j.clinph.2014.09.019