EEG Diagnostic Procedures and Special Investigations in the Assessment of Photosensitivity
Guido Rubboli, Jaime Parra, Stefano Seri, Takeo Takahashi, and Pierre Thomas
From Epilepsia, Volume 45 Issue s1 Page 35 - January 2004
doi:10.1111/j.0013-9580.2004.451002.x
URL:
http://www.blackwell-synergy.com/doi/full/10.1111/j.0013-9580.2004.451002.x
Summary:
Photosensitivity can be assessed in laboratory conditions with different methods. The most common procedure is intermittent photic stimulation (IPS), whose effectiveness in detecting photosensitivity depends largely on methodologic aspects. Although IPS is a widespread and routinely used procedure in EEG laboratories, only recently has a standardization of the IPS method been proposed. Furthermore, other modalities of visual stimulation, including pattern stimulation and low-luminance visual stimulation (LLVS), have proven their usefulness in detecting photosensitivity. We provide an overview of the methodologic aspects and clinical implications of these procedures, resulting from recent consensus meetings, and the diagnostic usefulness of the LLVS technique in photosensitive individuals whose seizures are triggered particularly by television images. Finally, we briefly illustrate the potential of advanced neurophysiological (magnetoencephalography and high-density EEG) and functional imaging techniques in the investigation of the pathophysiologic mechanisms underlying photosensitivity.
Intermittent photic stimulation (IPS) is a procedure performed in the EEG laboratory to detect abnormal sensitivity to light stimuli (i.e., photosensitivity), thus providing information on seizure susceptibility of the individual exposed to intermittent lights. More specific purposes would include quantifying the degree of photosensitivity and relating this phenomenon to definite epileptic syndromes. The effectiveness of IPS is indicated by the capability of inducing an abnormal EEG response in the highest number of patients in whom visual stimulation can potentially precipitate a seizure, reducing to a minimum the chance of obtaining such responses in normal subjects. Methodologic aspects are crucial when assessing photosensitivity in laboratory conditions. Unfortunately, no agreement on IPS procedure exists; therefore, considerably different protocols and photic stimulators are used, rendering difficult the evaluation and comparison of IPS effects across laboratories. In recent years, efforts have been made to standardize the methods and interpretation of the results of IPS.
We provide an overview of the diagnostic aspects of IPS procedure, based on data presented at Consensus Meetings held in Heemstede in 1996 and in Aix-en-Provence in 1999, and of other methods of visual stimulation. In addition, we briefly illustrate the potential usefulness of advanced neurophysiological (magnetoencephalography and high-density EEG) and functional imaging techniques in the investigation of the pathophysiologic mechanisms underlying photosensitivity.
EEG DIAGNOSTIC PROCEDURES
Intermittent photic stimulation
Photostimulator
At present, the commercially available device that best fulfills these requirements is the Grass PS22 (Grass-Telefactor; Astro-Med, Inc., West Warwick, Rhode Island, U.S.A.). It can deliver stimuli at a constant and relatively high intensity throughout all frequencies, with the shortest flash duration that does not increase with increasing intensity. It is equipped with a granular diffuser that increases the sensitivity of the assessment of photosensitivity and minimizes differences related to different types of flash bulbs or lamp glasses. The capability of delivering stimuli 
60 Hz allows testing upper frequency limits to assess potential television sensitivity.
IPS procedure
The procedure should be performed in a dimly lit environment (possibly artificially lighted at a constant level), 
3 min after hyperventilation. Recommended distance between the stimulator and the patient's nasion is 30 cm. This distance provides a sufficiently large visual field with the 13-cm circular Grass lamp, and enables the examiner to detect subtle clinical phenomena (for instance, eyelid myoclonia). Ten-second trains of flashes for each frequency should be delivered, at intervals of 
7 s. Eyes should be kept open for the first 5 s, fixating the center of the lamp. The patient should then close the eyes and remain in the eyes-closed condition for the remaining 5 s of stimulation. Recommended frequencies and their order of delivery are 1, 2, 3, 4, 6, 8, 10, 12, 14, 16, 18, and 20 Hz. Should a generalized epileptiform discharge occur, the stimulator must be switched off, and the procedure is stopped. A second sequence with frequencies 60, 50, 40, 30, and 20 Hz is then delivered with the same precaution (i.e., stopping the procedure if generalized paroxysms appear). For clinical purposes, it may be important to report (a) if the EEG response can outlast the stimulus and, if yes, whether it is blocked by monocular occlusion; (b) whether clinical signs/symptoms are observed; (c) alertness and eventual sleep deprivation of the patient; and (d) medications.
Interpretation of the results
IPS can elicit several types of EEG responses, with different clinical significance. A recent classification proposal identified the following types: photic following, a normal response that ends with termination of the stimulation; orbitofrontal photomyoclonus, a mainly electromyographic (although a frontal cortical component has been hypothesized) response at the flash frequency, terminating at the end of stimulation; posteriorstimulus
dependent response, which can be observed after suppression of IPS-induced generalized epileptiform activity because of medication, or in neuronal ceroid lipofuscinosis; posterior stimulus
independent response, which includes theta
delta activity, and clear-cut epileptiform patterns, not following the flash frequency or its harmonics, with two subtypes: (a) limited to the stimulus train, or (b) self-sustained; this latter may evolve to an overt seizure.
Generalized photoparoxysmal response (PPR), represented by multiple spikes or by spike
wave activity, apparently generalized, with anterior predominance. This can be limited to the stimulus train or be self-sustaining (i.e., continuing after stimulation). This latter response is considered to have a relation to epilepsy and visually induced seizures. When a photoparoxysmal response is obtained, it may be useful to test the effect of monocular occlusion (for example, for 10 s at 25 and 50 Hz to establish if this maneuver has a protective effect on the patient). A correct technique involves pressing the palm of the hand firmly into the orbit, with the lateral gaze up against the nose.
The activation of preexisting epileptogenic areas refers to the rare possibility of the activation of a spontaneously, posteriorly located, epileptogenic cortex. Its significance as a PPR is still debated.
Striped pattern testing
Although not standardized, striped pattern testing is a useful technique in photosensitive patients. It can be performed with eyes open, in ambient lighting with a TV monitor connected to an electronic grating generator, positioned 1 m from the patient's eyes. High-contrast black-and-white bars are successively presented for 10 s, either stationary or oscillating at spatial frequencies between 0.5 and 6 c/degree. A less sophisticated technique uses high-contrast black-and-white bars on circular pieces of cardboard of various diameters.
Low-luminance visual stimulation
"Full field" visual stimuli
EEG activation by low-luminance visual stimuli (LLVS) is performed only in the eyes-open condition with luminance ranging between 10 and 30 cd/m2. Among LLVS types, pattern flicker and red flicker are particularly potent in inducing PPR, and the rate of PPR activation by LLVS is higher than that obtained by stroboscopic IPS.
Flicker, pattern, and color are three fundamental parameters that determine efficacies of PPR activation by LLVS. For flicker, an 18-Hz stimulation is used. For pattern flicker by dot pattern with spatial frequencies of 0.5
4.9 c/degree, those in the range of 1.5
2.1 c/degree with a flicker frequency of 15
20 Hz were most effective in eliciting generalized PPRs. Stimuli within those parameters resulted in using three different patterns (dot pattern, vertical grating, and horizontal grating) with spatial frequency of 2 c/degree. To produce red flicker, of which the long-wavelength red (>600 nm) is a crucial factor, either a red glass filter in the case of visual stimulator SLS-5100 (Nihon Kohden) or a red plastic filter in the case of strobo-filter (Nihon Kohden) is used. The screen size of the SLS-5100 is 25 
25 cm (57 
57). Three strobo-filter sizes are as follows: (a) name plate type, 3 
13 cm; (b) square type, 13 
13 cm; and (c) circular type, 13 cm in diameter. Inserting each of the strobo-filters into the filter holder on the front of the flash lamp greatly reduces luminance of stroboscopic IPS to the degree of LLVS (
20 cd/m2), which enables subjects to look at the stimuli directly. Spatial frequency of each pattern of square-type strobo-filters is 2 c/degree when the distance between the flash lamp and the eyes is maintained at 30 cm. When square-type strobo-filters are used, some skills to adjust equal luminance in each are required, whereas only a simple manipulation is needed in the case of circular strobo-filters; inserting each filter into the filter holder yields nearly equal luminance of 30 cd/m2.
Comparisons of the rate of PPR activation by use of square and circular types of strobo-filters in 42 photosensitive epilepsy patients were performed. Stimuli used were: pattern flicker (18 Hz) stimuli by dot pattern, vertical grating, horizontal grating; and 18-Hz red flicker. Results were identical in both, suggesting that LLVSs by circular strobo-filters are able to provide sufficient data for diagnosis of photosensitive epilepsy.
LLVS to the center, periphery, and hemifield
With SLS-5100 and square-type strobo-filters, EEG changes in response to LLVS to different visual fields were investigated in photosensitive epilepsy patients. Compared with red-flicker stimulation to the periphery, that to the center (6 
6 degrees) promptly elicited a generalized PPR, suggesting that the macular areas are very sensitive to red-flicker stimulation. Producing a 20-Hz flickering dot pattern stimulation by dot patterns (0.5, 1, 2, 4, and 6 mm in diameter), EEG changes in response to that stimuli to the center (11 
11 degrees) and periphery (11
30 
11
30 degrees) were studied in 20 photosensitive epilepsy patients: a higher spatial frequency dot-pattern flicker stimulation to the center elicited PPR more strongly than that to the periphery, whereas a lower-spatial-frequency one showed the opposite result, with peripheral stimulation being more effective than central stimulation. PPR responses to LLVS to the lateral hemifield are as follows: (a) equal PPRs, sensitive to each lateral hemifield stimulation; (b) unequal PPRs, sensitive to either right or left hemifield stimulation; and (c) no PPRs, sensitive to neither right nor left hemifield stimulation (sensitive to full-field stimulation alone). Analyses of those data from an etiologic viewpoint revealed that equal PPRs and no PPRs had a closer link to genetic factors, whereas unequal PPRs related to acquired factors, such as head trauma.
ADVANCED NEUROPHYSIOLOGICAL AND FUNCTIONAL IMAGING TECHNIQUES IN THE STUDY OF PHOTOSENSITIVITY
MEG studies in photosensitive epilepsies
Magnetoencephalography (MEG) is a neurophysiologic technique, complementary to EEG, that measures the magnetic fields generated by the electrical currents in the brain. MEG has potential advantages over EEG. Whole-head MEG systems provide a higher spatial density of recording points than does the routine 10
20 system scalp EEG. Moreover, MEG offers theoretical advantages in studying synchronization or coherence because it does not require a reference sensor. Finally, magnetic field fluxes are less distorted than electric potentials by the smearing effect of the skull. Although clinical applications of MEG are still under development and must be properly validated, its introduction in clinical practice is increasing.
Few MEG studies have attempted to study the neuromagnetic origin of the PPR in epilepsy patients. Ricci et al. performed neuromagnetic measurements by using four sensors with simultaneous two-channel EEG recording in 12 patients with PPR identified in routine studies. Their results supported a cortical origin of the PPR, with a regional sensitivity and occasional asymmetries. They concluded that the erratic cortical involvement suggested the existence of a general instability of cortical excitability.
Inoue et al. studied 15 photo- and pattern-sensitive epilepsy (PSE) patients with seizures induced by electronic screen games and compared with 14 nonphotosensitive (non-PSE) patients, while they were playing videogames in the MEG environment. They estimated equivalent current dipoles of the MEG spikes and found that, whereas in PSE patients, dipoles had a posterior predominance, in non-PSE patients, the estimated sources of these epileptiform spikes tended to cluster over the anterior part of the brain. They concluded that factors involving functions of the anterior regions of the brain other than photo- or pattern sensitivity may play a role in the induction of seizures during the playing of electronic videogames. Furthermore, the changes in spike frequency in specific brain areas may correspond to their involvement in praxic activity and emotional changes during these games.
Parra et al. studied another visual reflex phenomenon, closely related to photosensitive epilepsy: the fixation-off phenomenon, and in particular its relations with the alpha rhythm. This condition denotes a variety of EEG abnormalities elicited by elimination of central vision and fixation, which include high-amplitude runs of repetitive occipital spike
wave or rhythmical slow activity in the EEG or even generalized discharges. Their findings indicated that abnormalities related to fixation-off sensitivity could emerge in thalamocortical networks, with larger and more anterior cortical distribution than those that generate alpha rhythm. Thus, the transition in the type of oscillation appears not only to depend on a change in cellular dynamics but also to be reflected in a different spatial distribution of the underlying neuronal networks.
Recently MEG has been used to study the phase-synchronization properties of the driving response preceding the onset of the PPR. They found an enhanced synchrony in the gamma band harmonics when IPS triggered a PPR. This was not seen in controls or in subjects in whom no PPR was elicited. Thus, it seems that properties of the phase spectrum, especially in MEG recordings, can characterize the dynamics of underlying neuronal networks more specifically than do amplitude spectra. These findings may reflect a loss of control of the brain over a fast-frequency oscillatory process that normally operates transiently to connect neural assemblies involved (e.g., in perceptual mechanisms). The possibility of anticipating the onset of a PPR may lead to new diagnostic evaluation of photosensitive patients, eventually resulting in an improvement of the safety of the IPS procedure, decreasing the risk of unnecessarily provoking seizures.
High-resolution EEG and functional imaging
Functional imaging techniques [positron emission tomography (PET); functional magnetic resonance imaging (fMRI)] with high spatial resolution can contribute to the understanding of metabolic and perfusion changes associated with both ictal and interictal EEG discharges in patients with epilepsy. Patients with photosensitive epilepsies and/or PPR on the EEG are ideal candidates for in vivo measurements, as their electroclinical phenomena usually have a localised onset and occur with minimal head motion. The millisecond time resolution of electrophysiologic techniques, coupled with indirect nature of the blood oxygenation level
dependent (BOLD) changes detected by fMRI and the underlying neural activity, give EEG and MEG a significant edge in deconvoluting fast changes associated with epileptogenic activity. The advent of high-resolution EEG acquisition systems, ranging from 128 to 256 channels, can overcome the limited spatial resolution of conventional EEG. In parallel with these technologic changes, novel analysis techniques have been developed, such as digital filtering, small time-lag estimation, and three-dimensional source modeling and reconstruction. As a result of these technologic advances, we are now able to infer the localisation of the underlying sources from scalp-recorded electrical and magnetic fields (the inverse problem). Furthermore, image fusioning with the anatomic MRI of the subject's head is now used to circumvent the inaccuracy of earlier studies, which relied on spherical head models. A number of studies on patients with symptomatic epilepsy have confirmed the high spatial correlation between the localization of reconstructed sources and that of the active brain regions identified from intracranial recordings or strongly correlating with anatomic lesions seen on the MRI. The application of these techniques to the understanding of cortical dynamics of the PPR is still in its infancy. Unpublished data from high-resolution EEG-MRI fusioning in conjunction with short time-lag estimation techniques seem to suggest a different electrophysiologic PPR pattern in juvenile myoclonic epilepsy (JME) versus progressive myoclonic epilepsy, with occipitally localised discharges occurring asynchronously in the latter.
Novel methods of simultaneous acquisition of EEG and fMRI have recently been developed. They capitalise on the relative merits of strength of the individual methods, achieving an unprecedented time and spatial resolution. Some of the initial technologic difficulties were related to the effect of strong magnetic fields on the EEG apparatus. Artifact-correction algorithms and specific hardware design have now overcome these initial limitations, allowing clinical studies to be undertaken. A recent study using simultaneous EEG and fMRI acquisition during pattern stimulation was able to localise sources of visual evoked responses "in vivo" in the calcarine area with unprecedented spatiotemporal resolution, demonstrating that the largest cortical activation corresponds to the N75 and P100 components of the visually evoked potential (VEP). As these modern integrated functional imaging techniques become more readily available, cross-fertilisation between different research areas will deepen the understanding of the dynamics and topography of cortical networks involved in photosensitive epilepsies, contributing to their improved electroclinical classification.
