• EEG;
  • Photosensitive epilepsy;
  • PPR;
  • PDC;
  • Connectivity


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosures
  7. References

Purpose:  Photosensitive epilepsy (PSE) is the most common form of reflex epilepsy presenting with electroencephalography (EEG) paroxysms elicited by intermittent photic stimulation (IPS). To investigate whether the neuronal network undergoes dynamic changes before and during the transition to an EEG epileptic discharge, we estimated EEG connectivity patterns in photosensitive (PS) patients with idiopathic generalized epilepsy.

Methods:  EEG signals were evaluated under resting conditions and during 14 Hz IPS, a frequency that consistently induces photoparoxysmal responses (PPRs) in PS patients. Partial directed coherence (PDC), a linear measure of effective connectivity based on multivariate autoregressive models, was used in 10 PS patients and 10 controls. Anterior versus posterior (F3, F4, C3, C4, and P3, P4, O1, O2) and interhemispheric connectivity patterns (F4, C4, P4, O2, and F3, C3, P3, O1) were estimated with focus on beta and gamma band activity.

Key Findings:  PDC analysis revealed an enhanced connectivity pattern in terms of both the number and strength of outflow connections in the PS patient group. Under resting condition, the greater connectivity in the PS patients occurred in the beta band, whereas it mainly involved the gamma band during IPS (i.e., the frequencies ranging from 40–60 Hz that include the higher harmonics of the stimulus frequency). Both at rest and during IPS, the differences between the PS patients and controls were due primarily to clearly increased connectivity involving the anterior cortical regions.

Significance:  Our findings indicate that PS patients are characterised by abnormal EEG hyperconnectivity, primarily involving the anterior cortical regions under resting conditions and during IPS. This suggests that, even if the occipital cortical regions are the recipient zone of the stimulus and probably hyperexcitable, the anterior cortical areas are prominently involved in generating the hypersynchronization underlying the spike-and wave discharges elicited by IPS.

Photosensitive epilepsy (PSE), the most common reflex epilepsy in humans, characterizes genetically determined epileptic syndromes (Kasteleijn-NolstTrenité, 1998; Zifkin & Kasteleijn-NolstTrenité, 2000; Stephani et al., 2004). Patients with PSE show electroencephalography (EEG) paroxysms in response to intermittent photic stimulation (IPS), an activation procedure that is applied routinely as a functional test during EEG examinations to enhance preexisting abnormalities or induce abnormal findings. Because of this, PSE offers a highly reproducible model for investigating the dynamic changes in neuronal activity that may occur before and during the transition to an EEG epileptic discharge.

It is known that IPS elicits a physiologic “photic driving” response in normal subjects, which consists of rhythmic EEG activity that is maximal over the posterior regions, time-locked to the stimulus, and has a frequency that is identical or harmonically related to that of the stimulus. However, in photosensitive (PS) patients, IPS can induce a photoparoxysmal response (PPR), a highly inheritable EEG trait characterized by the occurrence of spikes or spike-wave complexes (Fisher et al., 2005). This PPR may occur over the posterior scalp regions, but more frequently has a generalized distribution and can evolve into a self-sustained discharge that outlasts the stimulus itself (Waltz et al., 1992). The PPR and seizures triggered by IPS may be the only epileptic event in some patients (Guerrini & Genton, 2004; Lu et al., 2008); however, in most patients, the photosensitive trait appears as an age-dependent penetrance within a picture of idiopathic generalized epilepsies (IGEs) (Waltz & Stephani, 2000).

Although photosensitive EEG characteristics have been known for a long time, little is known about the mechanisms generating them or the relationship between physiologic and pathologic responses during IPS. The findings of previous electrophysiologic studies suggest that the PPR originates in the cortex and involves the synchronization of large neuronal networks (Wilkins et al., 1979; Binnie et al., 1984; Harding & Fylan, 1999). Evidence indicates that the control of excitation and synchronization is defective in photosensitive patients as a result of their impaired control of contrast gain mechanisms (Porciatti et al., 2000) and/or enhanced synchrony in the gamma band (30–120 Hz), which is harmonically related to the IPS frequencies, before the onset of PPRs (Parra et al., 2003).

In a previous study (Visani et al., 2010) of coherence between occipital and frontal derivations, we found abnormal intrahemispheric and interhemispheric gamma band synchronization in PS patients even at rest. Coherence is a validated means of studying EEG coupling affected by two major limitations: It is a bivariate measure and consequently cannot distinguish direct from indirect linear relationships in a multivariate system, and it does not take into account the direction of the information flow.

To overcome these intrinsic limitations of bivariate and undirected measures, Baccalà and Sameshima (2001) introduced partial directed coherence (PDC), a connectivity estimator in the frequency domain that is based on multivariate autoregressive (MVAR) models and provides a linear measure of causality indicating the direction and strength of the interactions between multiple coupled variables.

We used MVAR modeling and PDC to investigate the connectivity patterns in the beta and gamma bands of healthy participants and PS patients under resting conditions and during 14 Hz IPS. The aim of the study was to evaluate whether intrinsic connectivity patterns differentiate the networks of PS patients and healthy controls, and what changes occur immediately before the appearance of PPR during IPS.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosures
  7. References


The study involved 10 PS patients (six women; mean age and SD 19.0 ± 4.5 years) and 10 healthy controls (six women; mean age 23.9 ± 3.5 years). All of the participants were right-handed.

All of the patients underwent one or more EEG examinations at the Department of Neurophysiology of the C. Besta Neurological Institute between 2004 and 2010, were followed up by a clinical neurologist of the same Foundation, and were diagnosed as having IGE. All patients have had few convulsive generalized seizures presenting when they were 10–17 years old. Before the appearance of convulsive seizures, one patient had a single febrile seizure, another patient a simple absences, and a third patient had a photosensitive eyelids myoclonus. No patients had absences or myoclonic jerks at that moment of our evaluation. All patients had a normal neurologic picture and normal magnetic resonance imaging (MRI) findings. Four patients were treated with valproate, one with valproate and levetiracetam, one with ethosuximide and lamotrigine, and one with ethosuximide alone; three patients were untreated.

Our local institutional ethics committee approved the study. Written informed consent was obtained from all of the adult subjects and the parents of subjects aged <18 years.

EEG recordings and stimulus procedures

The EEG was recorded in a dimly lit room by means of Ag/AgCl surface electrodes placed according to the 10–20 International System, and the signals were acquired using a computerized Micromed Brain Quick system (Micromed SpA, Mogliano Veneto (TV), Italy) (sampling rate 256 Hz; band-pass filter 1–120 Hz, 12 dB/octave). All of the EEG signals were recorded using a montage with a common reference electrode that allowed off-line mathematical data reformatting.

A baseline eyes-closed EEG sample lasting >5 min was recorded in all of the patients before IPS. White flashes were delivered at 14 Hz for 10 s, using a Micromed Flash-10 photo-stimulator, with the lamp being placed 30 cm in front of the participants who kept their eyes closed. The IPS was stopped briefly after the onset of the paroxysmal response to avoid provoking ictal events.

To ensure the best rejection of all kinds of interference, the EEG machine was electrically isolated from any external power source or electrical device, including the photo stimulator; the stroboscope was controlled by the EEG software through an output trigger signal. Furthermore, EEG data were transferred from headbox (battery powered) to EEG machine using a fiberoptic connection. Finally, we checked the electrode impedances at the beginning of the EEG recording, immediately before the administration of the photic stimulation protocol, and at the end of the recording to maintain their values below 5 kΩ.

EEG analysis

The presence of artifacts on the EEG signals due to physiologic or nonphysiologic sources was accurately checked by an expert neurophysiologist, and epochs contaminated by any type of artifacts were excluded from the analysis.

As a preprocessing step, the EEG data were normalized by subtracting the mean value and dividing by the standard deviation, and a Laplacian spherical spline was applied to ensure reference-free and spatially sharpened data (Perrin et al., 1989; Babiloni et al., 1996). The issue of the EEG reference is still an open and critical point and an optimal EEG reference does not exist (Frei et al., 2010). We chose Laplacian derivations, since it has been previously shown (Srinivasan et al., 1998; Nunez & Srinivasan, 2006) that spline-Laplacian coherence estimates are more conservative with respect to common references. Signals from eight EEG derivations (F3, F4, C3, C4, P3, P4, O1, and O2) were included in the MVAR model. Given a set S = {xm(k), 1 ≤ m ≤ M} of M simultaneously observed stationary time series, an MVAR model with order p is defined as

  • image

where A1, A2,…,Ap are the coefficient matrices (dimensions M × M), with the coefficient aij(r) describing the linear interaction of xj(k−r) with xj(k), and wj(k) representing a random Gaussian white noise driving innovation.

The autoregressive (AR) model order was determined using the multichannel version of the Akaike (AIC) criterion as a guideline, and the goodness of the identification was verified by means of “portmanteau” chi-square and Anderson’s tests (Box & Jenkins, 1970; Lopes da Silva, 1999).

Once the MVAR coefficients had been adequately estimated, the PDC from channel j to i (πiJ) could be calculated as:

  • image


  • image

represents the difference between the identity matrix of dimension M × M and the Fourier transform of the model coefficient matrix.

PDC values range between 0 and 1; πiJ measures the outflow of information from channel xj to xi in relation to the total outflow of information from xj to all of the channels.

The statistical significance of nonzero PDC values at each frequency was estimated using a bootstrap approach based on phase randomization and the Theiler algorithm (Zoubir & Iskander, 2004; Baccalà et al., 2006). All of the data were preprocessed and analyzed using a custom-written toolbox in MATLAB (MathWorks Inc., Natick, MA, U.S.A.), containing modified functions from ARFIT (Neumaier & Schneider, 2001; Schneider & Neumaier, 2001) and Biosig (Schlogl & Brunner, 2008).

The PDC spectra of each participants were calculated on 1-s EEG epochs at rest (just before the beginning of IPS), and during 14 Hz IPS (the epoch immediately preceding the appearance of a PPR in the patients and 1.5 s after IPS onset in the controls). The spectra were analyzed in the beta and gamma bands, which were further divided into broader subbands. In particular, we analyzed the 13–18 Hz beta band (corresponding to the frequency range in which PPR frequently occurs in PS patients (Harding & Harding, 1999; Fisher et al., 2005), and the 26–40 and 40–60 Hz gamma subbands, each including an IPS harmonic. During IPS, PDC values were also considered in restricted bands centered on the harmonic frequencies of the 14 Hz IPS stimulus (28, 42, and 56 ± 2 Hz).

For each of these bands, we calculated the PDC amplitude and degree out-going from each electrode. These measures, respectively, represent the strength and the number of causal interactions originating at each electrode, which provide a measure of the source activity arising from each node in a network. These parameters were calculated for each electrode and then averaged to compare the connectivity between the anterior and posterior (F3,F4,C3,C4 and P3,P4,O1,O2), and interhemispheric (F4,C4,P4,O2 and F3,C3,P3,O1) regions.

The data were statistically analyzed using repeated-measure ANOVA at a significance level of 5%; the sphericity assumption was evaluated using Mauchley’s test, and the Greenhouse-Geisser degree of freedom correction was applied when appropriate. For each frequency band and condition (resting or IPS) two ANOVAs were made with the repeated-measure factor having two levels (anterior/posterior for the first ANOVA and right/left for the second one); the group factor had two levels (controls/PS patients). When ANOVA showed a significant main effect or interaction, a post hoc analysis within or between groups was made using Bonferroni correction for multiple comparisons, and all the reported p-values are after Bonferroni adjustment.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosures
  7. References

Resting conditions

Anterior versus posterior cortices

Comparison of the anterior and posterior electrodes showed that connectivity in the low beta band (13–18 Hz) was distributed differently in the patients and healthy controls. The controls showed significantly more out-going connectivity (both in terms of amplitude and degree) from the posterior electrodes than from the anterior ones, whereas the patients showed more homogeneous values (Fig. 1A; Table 1a). This uneven connectivity pattern was due to the significantly greater connectivity in the anterior electrodes of the patients with respect to controls (Fig. 1A; Table 1a).


Figure 1.   Upper graphs: comparison of controls and PS patients under resting conditions showing the different connectivity between the anterior and posterior regions in the beta band (13–18 Hz) (A), and between the left and right frontal regions in the low gamma band (26–40 Hz) (B). Lower graph: difference between the anterior and posterior regions comparing PS patients and controls in the medium gamma band (40–60 Hz) during IPS. Significant differences after Bonferroni corrections are indicated by horizontal lines with asterisk.

Download figure to PowerPoint

Table 1.   Resting conditions: anterior and posterior (a) and frontal and central (b) PDC amplitude and mean degree in the 13–18 Hz band
 PDC amplitudeMean degree
ControlsPatients ControlsPatients 
 Anterior0.16 ± 0.070.28 ± 0.12p = 0.0420.50 ± 0.180.8 ± 0.37NS
 Posterior0.46 ± 0.220.41 ± 0.19NS0.92 ± 0.410.9 ± 0.37NS
 p = 0.01NS p = 0.05NS 
 Frontal0.15 ± 0.080.24 ± 0.14NS0.50 ± 0.350.60 ± 0.46NS
 Central0.17 ± 0.10.32 ± 0.11p = 0.0260.50 ± 0.350.90 ± 0.46NS

An evaluation of the individual contribution of the frontal and central electrodes revealed considerable intersubject variability. However, the difference between the patients and controls was due mainly to the contribution of the central electrodes, which showed significantly higher out-going PDC in the patients. The frontal electrodes showed a similar trend, but the difference was not statistically significant (Table 1b).

In the other investigated frequency bands, including the high beta band (18–26 Hz), there were no significant differences between the patients and controls, or between the anterior and posterior regions.

Right versus left hemisphere

Because the anterior cortex of the patients and controls revealed a different connectivity pattern, we further investigated out-going connectivity from the right (F4 and C4) and left (F3 and C3) regions.

There was no difference in the 13–18 Hz band, but there were significant differences in the low gamma band (26–40 Hz), in which the controls showed significantly greater out-going connectivity in the right frontal electrodes in terms of both amplitude and degree, a difference that was not significant in the patients. Furthermore, the patients showed more connectivity in the left frontal cortex than the controls in terms of both amplitude and degree (Fig. 1B; Table 2). No statistically significant difference in the connectivity patterns was observed between right (P4 and O2) and left (P3 and O1) posterior electrodes.

Table 2.   Resting conditions: right and left anterior hemisphere PDC amplitude and mean degree in the 26–40 Hz band
 PDC amplitudeMean degree
ControlsPatients ControlsPatients 
F40.34 ± 0.110.19 ± 0.21NS1.78 ± 0.801.40 ± 1.07NS
F30.08 ± 0.140.37 ± 0.20p = 0.0040.33 ± 0.501.80 ± 1.23p = 0.028
 p = 0.002NS p = 0.002NS 
C40.36 ± 0.280.39 ± 0.5NS1.44 ± 1.331.10 ± 1.20NS
C30.22 ± 0.240.40 ± 0.27NS0.89 ± 1.100.9 ± 0.5NS

Photic stimulation

The differences in the low beta band that distinguished the patients and controls under resting conditions disappeared during 14 Hz photic stimulation, and the beta connectivity patterns of both were similar, namely being the posterior electrodes highly connected than the anterior ones in both groups.

On the contrary, there were significant differences in the medium gamma band (40–60 Hz) and in the two frequency bands centred on 42 and 56 Hz that were harmonically related to the frequency of the stimulus. Interestingly, network behavior in the 40–60 Hz gamma band nearly replicated that of the low beta band under resting conditions, with the controls showing significantly greater connectivity on the posterior than on the anterior electrodes, and the PS patients showing similar values for both (Fig. 1C, Table 3a). Moreover, the patients showed significantly greater anterior out-going connectivity than the controls (Fig. 1C. Table 3a; see Fig. 2 for a representative example). This difference was significant for the whole 40–60 Hz band, but the significance was even greater when the bands centered on 42 and 56 Hz were analyzed separately (Table 3). As regards the posterior region, we did not find any significant difference between the patients and controls.

Table 3.   IPS: anterior and posterior PDC amplitude and mean degree in the 40–60 Hz band, and in correspondence with the frequencies centered on higher harmonics of stimulus frequency
 PDC amplitudeMean degree
ControlsPatients ControlsPatients 
(a) 40–60 Hz      
 Anterior0.15 ± 0.110.29 ± 0.12p = 0.0320.89 ± 0.591.5 ± 0.55NS
 Posterior0.34 ± 0.070.38 ± 0.1NS1.47 ± 0.751.7 ± 0.50NS
 p = 0.016NS NSNS 
(b) 40–44 Hz      
 Anterior0.08 ± 0.060.25 ± 0.13p = 0.0040.39 ± 0.310.9 ± 0.44p = 0.042
 Posterior0.26 ± 0.080.38 ± 0.13NS1.00 ± 0.571.1 ± 0.36NS
 p = 0.002p = 0.02 p = 0.03NS 
(c) 54–58 Hz      
 Anterior0.09 ± 0.060.25 ± 0.13p = 0.010.44 ± 0.341.00 ± 0.48p = 0.02
 Posterior0.27 ± 0.100.31 ± 0.13NS0.94 ± 0.540.95 ± 0.64NS
 p = 0.008NS NSNS 

Figure 2.   EEG recordings during 14 Hz IPS in a representative healthy participant (A) and a PS patient (B), and the corresponding cortical connectivity pattern in the gamma band, showing the greater out-going connectivity of the anterior cortex of PS patients. Orange arrows represent anterior out-going connections and blue arrows the posterior ones, corresponding to PDC values significantly different from zero after Bonferroni correction.

Download figure to PowerPoint

Only the frontal electrodes contributed to the greater medium gamma band connectivity in the patients. Indeed, in all of the considered frequency bands, the out-going connectivity from the frontal electrodes had significantly higher PDC values in the patients than in the controls (Table 4).

Table 4.   IPS: frontal and central PDC amplitude and mean degree in the 40–60 Hz band, and in correspondence with the frequencies centred on the higher harmonics of stimulus frequency
 PDC amplitudeMean degree
ControlsPatients ControlsPatients 
(a) 40–60 Hz      
 Frontal0.16 ± 0.130.31 ± 0.12p = 0.040.78 ± 0.571.80 ± 1.00p = 0.028
 Central0.14 ± 0.140.29 ± 0.26NS1.00 ± 1.001.10 ± 1.00NS
(b) 40–44 Hz      
 Frontal0.09 ± 0.090.26 ± 0.13p = 0.0060.39 ± 0.421.20 ± 0.75p = 0.02
 Central0.07 ± 0.090.25 ± 0.27NS0.44 ± 0.580.60 ± 0.52NS
(c) 54–58 Hz      
 Frontal0.08 ± 0.080.25 ± 0.18p = 0.0380.33 ± 0.351.25 ± 0.90p = 0.02
 Central0.11 ± 0.120.27 ± 0.27NS0.61 ± 0.700.80 ± 0.71NS

The asymmetry in the 26–40 Hz band between the left and right hemispheres observed under resting conditions disappeared completely during IPS in all of the analyzed frequency ranges.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosures
  7. References

The results obtained in patients with photoparoxysmal EEG responses indicate an enhanced connectivity pattern mainly involving the frontocentral cortical regions. This led to significantly more and stronger outflow connections that involved beta frequencies under resting conditions and gamma activity during IPS. The different topographical patterns of connectivity between the PS patients and controls under resting conditions and during the stimulus trains indicate that specific changes in the neuronal network predispose toward the generation of synchronous paroxysmal activities and particularly paroxysmal responses to IPS.

We found that the functional alteration affecting the cortical network of PS patients under resting conditions was reflected by significantly enhanced connectivity in the 13–18 Hz beta band, the regional distribution of which was different from that observed in the controls. In the patients, the source was homogeneously distributed on the anterior and posterior electrodes as a result of enhanced interactions leaving the anterior electrodes, whereas the prominent out-going beta connectivity of the controls was on the parietooccipital electrodes. This may be in line with the findings of Kondakor et al. (2005) that, studying the Omega complexity of IGE patient’s EEG studies, indicated greater synchrony in the anterior regions of IGE patients than in those of controls.

The involvement of prominent anterior beta connectivity in the propensity to generate generalized epileptic discharges in PS patients may be due to the neuronal network (possibly inhibitory interneurons) trying to normalize cortical malfunctioning, as has been hypothesized in the case of other neurologic disorders (Chiang et al., 2009; de Haan et al., 2009). However, it is more likely that the changes in connectivity are related to the inherited cell/circuitry changes that certainly occur in photosensitive epileptic syndromes. It is well known that most epilepsies commonly associated with photosensitivity (as well as the “pure” photosensitive epilepsies) have a genetic origin, even though the specific inherited determinant is still insufficiently defined or limited to a small number of patients/families (Stephani et al., 2004; Kasteleijn-NolstTrenité, 2006; Weber & Lerche, 2008). Furthermore, genetic backgrounds are also associated with different patterns of EEG connectivity in healthy subjects, give rise to different connectivity phenotypes (Smit et al., 2010), and lead to ontogenetic changes. Using the synchronization likelihood index, Boersma et al. (2011) have recently found that connectivity in various frequency bands (including beta frequencies) decrease with age in children, possibly because of a tendency to simplify and optimize synaptic connections in order to create a less “expensive” network. This physiologic process of circuitry rearrangement may be delayed or defective in PS patients, thereby leading to different cortical coupling at a juvenile age when photosensitivity is maximally expressed.

The specific role of beta frequencies in connecting the anterior cortical regions of PS patients under resting conditions does not have an obvious explanation. Both beta and gamma frequencies have been attributed to cortical or thalamocortical networks of inhibitory interneurons (Porjesz et al., 2002), and so the enhanced beta band connectivity may reflect an attempt by the system to control the intrinsic hyperexcitability of cortical circuitry prone to generate paroxysmal activity. However, cortical circuitries including both inhibitory and excitatory neurons may also generate a wide range of oscillations (Kopell et al., 2000; van Aerde et al., 2009). Interestingly, the beta band showing enhanced connectivity in our PS patients included the most epileptogenic frequencies of photic stimuli that are consistently capable of evoking paroxysmal EEG activity (Fisher et al., 2005), and may predispose the generation of a pathologic response to IPS. The marked involvement of the central region in the enhanced connectivity of PS patients may be due to the magnification of the natural aptitude of the sensorimotor cortex to generate beta rhythms (Niedermayer, 1996).

Further evaluations of the signal recorded on the anterior regions revealed that there was also a different pattern of connectivity between the left and right anterior hemispheres in the 26–40 Hz gamma band, which corresponds to twice the frequency of the beta band responsible for the stronger connectivity pattern in PS patients. The outflow connections of the patients differed from those of controls because of the left hyperconnectivity observed in patients. Studies comparing the two hemispheres have previously detected differences in EEG coherence between patients with psychiatric disorders or mental impairment and healthy controls at rest and during photic driving (Lazarev et al., 2010; Sankari et al., 2010), which suggest that the intrahemispheric connections undergo compensatory adjustments. Moreover, asymmetrical connectivity unevenly develops and declines in healthy subjects of different ages, and the left anterior regions undergo the main changes (Zhu et al., 2011). The peculiar left “dominance” found in the anterior cortical region of PS patients may also be attributable to the same distorted (genetic and developmental) mechanism that explains the differences between the anterior and posterior cortical areas. The greater propensity of the dominant hemisphere to generate hyperexcitability phenomena has been demonstrated in large populations of patients with focal epilepsies (Gatzonis et al., 2002; Aurlien et al., 2007), and we have found that patients with juvenile myoclonus epilepsy are more likely to generate jerk-locked spikes in the dominant hemisphere (Panzica et al., 2001).

The pattern of beta band connectivity differentiating PS patients and healthy controls under resting conditions did not survive IPS administration, and the same was true of the asymmetrical gamma connectivity between the right and left anterior electrodes in the low gamma band.

The clearest and most significant finding distinguishing PS patients from healthy controls during the EEG epochs heralding the occurrence of paroxysmal EEG discharges occurred in a “medium” gamma band (40–60 Hz), predominantly in the frequencies centered on higher stimulus harmonics. The difference in this frequency window replicated the topography of the beta band under resting conditions, which again suggests that the same neuronal networks sustain both forms of connectivity and are influenced by IPS to generate faster (gamma) oscillations. The main differences in gamma band connectivity related to the frontal electrodes, which are known to be particularly involved in generating high frequency oscillations (Rosanova et al., 2009).

The issue of the EEG contamination by muscular activity has been discussed in various articles dealing with gamma EEG activities (see Dalal et al., 2011 for a review); therefore, we accurately excluded signals with obvious or suspected artifacts. Moreover, because we compared the connectivity patterns obtained in PS patients and controls who underwent identical procedures, it can be assumed that possible contamination would be similarly distributed. Eventually, most the differences observed in PDC during IPS were in frequency bands centered on harmonics of the stimulus frequency, whereas muscular activity would produce rather more distributed background noise.

The involvement of gamma activities in epileptic events has been investigated previously using different techniques in patients with severe focal epilepsies (Guggisberg et al., 2008) and attributed to decreased GABA inhibition at the dendrite level (Wendling et al., 2002). Parra et al. (2003) found significantly enhanced gamma oscillations in PS patients during IPS and suggested that a high degree of synchronization in the gamma band plays a pathogenic role in generating PPR. In line with this, we found a greater trend toward the generation of synchronous gamma activities in PS patients in a previous study based on coherence analysis (Visani et al., 2010), which suggests that gamma activity plays a role in the interhemispheric and intrahemispheric pathologic synchronization heralding PPR.

The present PDC study confirms the role of gamma activity in the generation of PPR, and we suggest that the activation of a network that generates gamma activity capable of coupling the anterior cortices of both hemisphere heralds the generation of paroxysmal activity.

There were no significant differences between the PS patients and healthy controls in terms of the pattern of connectivity in the occipital and parietal cortices (the specific recipient areas of light stimuli) except for a narrow frequency band centered on the IPS harmonic at 42 Hz. This seems to conflict with the hypothesis that the occipital cortex plays a role in generating PPR. Neurophysiologic evaluations such as transcranial magnetic stimulation (TMS) (Siniatchkin et al., 2007; Shepherd & Siniatchkin, 2009), evoked potentials (Guerrini et al., 1998; Porciatti et al., 2000), or fMRI (Chiappa et al., 1999; Moeller et al., 2009a,b) have indicated the increased involvement of the primary recipient occipital cortices of light stimuli in PS patients. However, most of these works (and other hemodynamic and neurophysiologic studies) found that extraoccipital cortices and subcortical structures are also involved in the hyperexcitability mechanism (Kapucu et al., 1996; Chiappa et al., 1999; Inoue et al., 1999; Holmes et al., 2010). Frontal or premotor cortices are certainly involved in the direct generation of spike and wave discharges, including those evoked by PS in patients with generalized epilepsies, as has been directly demonstrated by inspectional EEG analyses (Niedermeyer, 1996; Kasteleijn-NolstTrenité, 1998) or quantitative methods (Gotman, 1981; Takasaka et al., 1989; Visani et al., 2010).

It is also worth noting that the lack of any significant difference in the PDC from the EEG signals recorded in the posterior cortical regions of PS patients and controls is also probably due to the particular characteristics of this measure. Unlike coherence and other linear and nonlinear bivariate methods, PDC allows a multichannel analysis that is capable of detecting the direction of the interactions between paired derivations after removing the contribution of all the other signals, thus revealing the direct connections (Baccalà & Sameshima, 2001; Astolfi et al., 2006; Baccalà et al., 2006). The topographical patterns generated by PDC can, therefore, probably detect the connections more specifically and reveal the causal interdependences within the neuronal network.

In this scenario, it can be hypothesized that PPR requires a hyperexcitable occipital recipient zone, whereas the network originating spike-and wave discharges involves hyperconnected frontocentral cortices. Physiologically, beta frequencies are involved in the coordination of distributed neural activities over longer-range distances (Kopell et al., 2000), whereas gamma oscillations play a major role in coupling local neuronal assemblies and enable synchronization over relatively short distances (Fries et al., 2007). The abnormally synchronized beta bands observed at rest in the anterior cortical regions of PS patients may reflect defectively controlled high-frequency oscillatory processes that couple local and distant neuronal populations in an extended network and pathologically predispose toward generalized epileptic discharges. Under these conditions, IPS may act as an appropriate stimulus to engage the network in generating the connected (gamma) frequencies, ultimately leading to PPR as result of widespread intrahemispheric and interhemispheric transfers, possibly involving cortico-thalamo-cortical pathways or structurally defective connections due to subtle maldevelopments similar to that found in juvenile myoclonic epilepsy (Vulliemoz et al., 2011).


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosures
  7. References

We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. None of the authors has any conflict of interest to disclose.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosures
  7. References
  • Astolfi L, Cincotti F, Mattia D, Marciani MG, Baccalà LA, de Vico Fallani F, Salinari S, Ursino M, Zavaglia M, Babiloni F. (2006) Assessing cortical functional connectivity by partial directed coherence: simulations and application to real data. IEEE Trans Biomed Eng 53:18021812.
  • Aurlien H, Aarseth JH, Gjerde IO, Karlsen B, Skeidsvoll H, Gilhus NE. (2007) Focal epileptiform activity described by a large computerised EEG database. Clin Neurophysiol 118:13691376.
  • Babiloni F, Babiloni C, Carducci F, Fattorini L, Onorati P, Urbano A. (1996) Spline Laplacian estimate of EEG potentials over a realistic magnetic resonance-constructed scalp surface model. Electroencephalogr Clin Neurophysiol 98:363373.
  • Baccalà LA, Sameshima K. (2001) Partial directed coherence: a new concept in neural structure determination. Biol Cybern 84:463474.
  • Baccalà LA, Takahashi DY, Sameshima K (2006) Computer intensive testing for the influence between time series. In Shelter B, Winterhalder M, Timmer J (Eds) Handbook of time series analysis. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, pp. 411433.
  • Binnie CD, Aarts JH, Houtkooper MA, Laxminarayan R, Martins da Silva A, Meinardi H, Nagelkerke N, Overweg J. (1984) Temporal characteristics of seizures and epileptiform discharges. Electroencephalogr Clin Neurophysiol 58:498505.
  • Boersma M, Smit DJ, de Bie HM, Van Baal GC, Boomsma DI, de Geus EJ, Delemarre-van de Waal HA, Stam CJ. (2011) Network analysis of resting state EEG in the developing young brain: structure comes with maturation. Hum Brain Mapp 32:413425.
  • Box GEP, Jenkins GM (1970) Time series analysis forecasting and controls. Holden-Day, San Francisco.
  • Chiang J, Wang Z, McKeown MJ. (2009) EEG source extraction by autoregressive source separation reveals abnormal synchronization in Parkinson’s disease. Conf Proc IEEE Eng Med Biol Soc 2009:18681872.
  • Chiappa KH, Hill RA, Huang-Hellinger F, Jenkins BG. (1999) Photosensitive epilepsy studied by functional magnetic resonance imaging and magnetic resonance spectroscopy. Epilepsia 40(Suppl. 4):S3S7.
  • Dalal SS, Vidal JR, Hamamé CM, Ossandón T, Bertrand O, Lachaux JP, Jerbi K. (2011) Spanning the rich spectrum of the human brain: slow waves to gamma and beyond. Brain Struct Funct 216:7784.
  • de Haan W, Pijnenburg YA, Strijers RL, van der Made Y, van der Flier WM, Scheltens P, Stam CJ. (2009) Functional neural network analysis in frontotemporal dementia and Alzheimer’s disease using EEG and graph theory. BMC Neurosci 10:101.
  • Fisher RS, Harding G, Erba G, Barkley GL, Wilkins A. (2005) Photic- and pattern-induced seizures: a review for the Epilepsy Foundation of America Working Group. Epilepsia 46:14261441.
  • Frei MG, Zaveri HP, Arthurs S, Bergey GK, Jouny CC, Lehnertz K, Gotman J, Osorio I, Netoff TI, Freeman WJ, Jefferys J, Worrell G, Quyen Mle V, Schiff SJ, Mormann F. (2010) Controversies in epilepsy: debates held during the Fourth International Workshop on Seizure Prediction. Epilepsy Behav 19:416.
  • Fries P, Nikolic′ D, Singer W. (2007) The gamma cycle. Trends Neurosci 30:309316.
  • Gatzonis SD, Roupakiotis S, Kambayianni E, Politi A, Triantafyllou N, Mantouvalos V, Chioni A, Zournas Ch, Siafakas A. (2002) Hemispheric predominance of abnormal findings in electroencephalogram (EEG). Seizure 11:442444.
  • Gotman J. (1981) Interhemispheric relations during bilateral spike-and-wave activity. Epilepsia 22:453466.
  • Guerrini R, Genton P. (2004) Epileptic syndromes and visually induced seizures. Epilepsia 45(Suppl. 1):S14S18.
  • Guerrini R, Bonanni P, Parmeggiani L, Thomas P, Mattia D, Harvey AS, Duchowny MS. (1998) Induction of partial seizures by visual stimulation. Clinical and electroencephalographic features and evoked potential studies. Adv Neurol 75:159178.
  • Guggisberg AG, Kirsch HE, Mantle MM, Barbaro NM, Nagarajan SS. (2008) Fast oscillations associated with interictal spikes localize the epileptogenic zone in patients with partial epilepsy. Neuroimage 39:661668.
  • Harding GF, Fylan F. (1999) Two visual mechanisms of photosensitivity. Epilepsia 40:14461451.
  • Harding GFA, Harding PF. (1999) Televised material and photosensitive epilepsy. Clin Neurophysiol 40:6569.
  • Holmes MD, Quiring J, Tucker DM. (2010) Evidence that juvenile myoclonic epilepsy is a disorder of frontotemporal corticothalamic networks. Neuroimage 49:8093.
  • Inoue Y, Fukao K, Araki T, Yamamoto S, Kubota H, Watanabe Y. (1999) Photosensitive and nonphotosensitive electronic screen game-induced seizures. Epilepsia 40(Suppl. 4):S8S16.
  • Kapucu LO, Gücüyener K, Vural G, Köse G, Tokçaer AB, Turgut B, Unlü M. (1996) Brain SPECT evaluation of patients with pure photosensitive epilepsy. J Nucl Med 37:17551759.
  • Kasteleijn-NolstTrenité DG. (1998) Reflex seizures induced by intermittent light stimulation. Adv Neurol 75:99121.
  • Kasteleijn-NolstTrenité DG. (2006) Photosensitivity, visually sensitive seizures and epilepsies. Epilepsy Res 70(Suppl. 1):S269S279.
  • Kondakor I, Toth M, Wackermann J, Gyimesi C, Czopf J, Clemens B. (2005) Distribution of spatial complexity of EEG in idiopathic generalized epilepsy and its change after chronic valproate therapy. Brain Topogr 18:115123.
  • Kopell N, Ermentrout GB, Whittington MA, Traub RD. (2000) Gamma rhythms and beta rhythms have different synchronization properties. Proc Natl Acad Sci U S A 97:18671872.
  • Lazarev VV, Pontes A, Mitrofanov AA, deAzevedo LC. (2010) Interhemispheric asymmetry in EEG photic driving coherence in childhood autism. Clin Neurophysiol 121:145152.
  • Lopes da Silva FH (1999) EEG analysis: theory and practice. In Niedermeyer E, Lopes da Silva FH (Eds) Electroencephalography: basic principles, clinical application and related fields. Urban & Schwarzenberg, Baltimore, MD/Munich, pp. 11351163.
  • Lu Y, Waltz S, Stenzel K, Muhle H, Stephani U. (2008) Photosensitivity in epileptic syndromes of childhood and adolescence. Epileptic Disord 10:136143.
  • Moeller F, Siebner HR, Wolff S, Muhle H, Granert O, Jansen O, Stephani U, Siniatchkin M. (2009a) Mapping brain activity on the verge of a photically induced generalized tonic-clonic seizure. Epilepsia 50:16321637.
  • Moeller F, Siebner HR, Ahlgrimm N, Wolff S, Muhle H, Granert O, Boor R, Jansen O, Gotman J, Stephani U, Siniatchkin M. (2009b) fMRI activation during spike and wave discharges evoked by photic stimulation. Neuroimage 48:682695.
  • Neumaier A, Schneider T. (2001) Estimation of parameters and eigenmodes of multivariate autoregressive models. ACM Trans Math Softw 27:2757.
  • Niedermeyer E. (1996) Primary (idiopathic) generalized epilepsy and underlying mechanisms. Clin Electroencephalogr 27:121.
  • Nunez PL, Srinivasan R (2006) Electric fields of the brain: the neurophysics of EEG. Oxford University Press, New York.
  • Panzica F, Rubboli G, Franceschetti S, Avanzini G, Meletti S, Pozzi A, Tassinari CA. (2001) Cortical myoclonus in Janz syndrome. Clin Neurophysiol 112:18031809.
  • Parra J, Kalitzin SN, Iriarte J, Blanes W, Velis DN, Lopes da Silva FH. (2003) Gamma-band phase clustering and photosensitivity: is there an underlying mechanism common to photosensitive epilepsy and visual perception? Brain 126:11621172.
  • Perrin F, Pernier J, Bertrand O, Echallier JF. (1989) Spherical splines for scalp potential and current density mapping. Electroencephalogr Clin Neurophysiol 72:184187.
  • Porciatti V, Bonanni P, Fiorentini A, Guerrini R. (2000) Lack of cortical contrast gain control in human photosensitive epilepsy. Nat Neurosci 3:259263.
  • Porjesz B, Almasy L, Edenberg HJ, Wang K, Chorlian DB, Foroud T, Goate A, Rice JP, O’Connor SJ, Rohrbaugh J, Kuperman S, Bauer LO, Crowe RR, Schuckit MA, Hesselbrock V, Conneally PM, Tischfield JA, Li TK, Reich T, Begleiter H. (2002) Linkage disequilibrium between the beta frequency of the human EEG and a GABAA receptor gene locus. Proc Natl Acad Sci U S A 99:37293733.
  • Rosanova M, Casali A, Bellina V, Resta F, Mariotti M, Massimini M. (2009) Natural frequencies of human corticothalamic circuits. J Neurosci 29:76797685.
  • Sankari Z, Adeli H, Adeli A. (2010) Intrahemispheric, interhemispheric, and distal EEG coherence in Alzheimer’s disease. Clin Neurophysiol 122:897906.
  • Schlogl A, Brunner C. (2008) BioSig: a free and open source software library for BCI research. Computer 41:4450.
  • Schneider T, Neumaier A. (2001) Algorithm 808: ARfit – A Matlab package for the estimation of parameters and eigenmodes of multivariate autoregressive models. ACM Trans Math Softw 27:5865.
  • Shepherd AJ, Siniatchkin M. (2009) Visual pattern adaptation in subjects with photoparoxysmal EEG response: evidence for increased visual cortical excitability. Invest Ophthalmol Vis Sci 50:14701476.
  • Siniatchkin M, Groppa S, Jerosch B, Muhle H, Kurth C, Shepherd AJ, Siebner H, Stephani U. (2007) Spreading photoparoxysmal EEG response is associated with an abnormal cortical excitability pattern. Brain 130:7887.
  • Smit DJ, Boersma M, van Beijsterveldt CE, Posthuma D, Boomsma DI, Stam CJ, deGeus EJ. (2010) Endophenotypes in a dynamically connected brain. Behav Genet 40:167177.
  • Srinivasan R, Nunez PL, Silberstein RB. (1998) Spatial filtering and neocortical dynamics: estimates of EEG coherence. IEEE Trans Biomed Eng 45:814826.
  • Stephani U, Tauer U, Koeleman B, Pinto D, Neubauer BA, Lindhout D. (2004) Genetics of photosensitivity (photoparoxysmal response): a review. Epilepsia 45(Suppl. 1):1923.
  • Takasaka Y, Takamatsu K, Nakagawara M. (1989) Anterior–posterior relationships of EEG in photosensitive subjects: coherence and cross-phase-spectral analysis. Jpn J Psychiatry Neurol 43:651663.
  • van Aerde KI, Mann EO, Canto CB, Heistek TS, Linkenkaer-Hansen K, Mulder AB, van der Roest M, Paulsen O, Brussaard AB, Mansvelder HD. (2009) Flexible spike timing of layer 5 neurons during dynamic beta oscillation shifts in rat prefrontal cortex. J Physiol 587:51775196.
  • Visani E, Varotto G, Binelli S, Fratello L, Franceschetti S, Avanzini G, Panzica F. (2010) Photosensitive epilepsy: spectral and coherence analyses of EEG using 14Hz intermittent photic stimulation. Clin Neurophysiol 121:318324.
  • Vulliemoz S, Vollmar C, Koepp MJ, Yogarajah M, O’Muircheartaigh J, Carmichael DW, Stretton J, Richardson MP, Symms MR, Duncan JS. (2011) Connectivity of the supplementary motor area in juvenile myoclonic epilepsy and frontal lobe epilepsy. Epilepsia 52:507514.
  • Waltz S, Stephani U. (2000) Inheritance of photosensitivity. Neuropediatrics 31:8285.
  • Waltz S, Christen HJ, Doose H. (1992) The different patterns of the photoparoxysmal response: a genetic study. Electroencephalogr Clin Neurophysiol 83:138145.
  • Weber YG, Lerche H. (2008) Genetic mechanisms in idiopathic epilepsies. Dev Med Child Neurol 50:648654.
  • Wendling F, Bartolomei F, Bellanger JJ, Chauvel P. (2002) Epileptic fast activity can be explained by a model of impaired GABAergic dendritic inhibition. Eur J Neurosci 15:14991508.
  • Wilkins AJ, Darby CE, Binnie CD. (1979) Neurophysiological aspects of pattern-sensitive epilepsy. Brain 102:125.
  • Zhu C, Guo X, Jin Z, Sun J, Qiu Y, Zhu Y, Tong S. (2011) Influences of brain development and ageing on cortical interactive networks. Clin Neurophysiol 122:278283.
  • Zifkin BG, Kasteleijn-NolstTrenité D. (2000) Reflex epilepsy and reflex seizures of the visual system: a clinical review. Epileptic Disord 2:129136.
  • Zoubir AM, Iskander DR (2004) Bootstrap techniques for signal processing. Cambridge University Press, Cambridge.