Causes and definition of ESGS
ESGS was first reported by Rushton and Ferry, who coined the term, “video-game epilepsy”(7,8). With the subsequent rapid dissemination of these games, increasing numbers of studies have been conducted, but the definition of ESGS has not yet been established.
No clear definition is given even in the Consensus Statement prepared by the Video-Game Epilepsy Consensus Group, which was published in 1994. The statement merely introduces the following as its causes (9): (a) a photosensitive response to the physical characteristics of a television display, i.e.; (b) a photosensitive response to the visual content of the game; (c) seizure precipitation by specific cognitive activities, decision making, hand movements, etc.; (d) seizure precipitation by nonspecific emotional factors relating to the subject's engagement in the game, such as anxiety or excitement; (e) lowering of seizure threshold by fatigue or sleep deprivation; and (f) chance occurrence of a spontaneous seizure in a person with epilepsy playing a video-game.
Accordingly, if ESGS is defined as any seizure activity that develops when one plays a video game, it may include even those that incidentally occur while playing such a game.
For the diagnosis of an epileptic seizure provoked while playing a video game, William et al. (10) emphasized the following: (a) the repeated occurrence of seizures during video games; (b) a history of previous epileptic manifestations triggered by other visual stimuli; and (c) laboratory demonstration of EEG paroxysmal responses to a variety of visual excitations, including actual video-game playing, stroboscopic stimulation, and presentation of line patterns.
Identification of photo-pattern–sensitive individuals
Based on the theories described earlier, we studied 17 patients with ESGS. Among them, three (cases 10, 13, and 14) experienced repeated spontaneous non-ESGS seizures between the onset of ESGS and the CRT-pattern test, raising the possibility that their ESGS was incidental. Fourteen others did not experience a spontaneous seizure during this period: nine (cases 1–9) were PPR-positive on the CRT-pattern test, and photosensitivity had already been proved in three of the remaining five (cases 15, 16, and 17). VPA has been shown to be the most effective agent for the treatment of photosensitivity (11). The three patients taking VPA were PPR-positive on the previous conventional IPS test before the introduction of VPA, but PPR-negative on the CRT-pattern test after the initiation of this treatment. It is plausible that these 12 patients (cases 1–9, and 15, 16, and 17) represent ESGS caused by photo-pattern sensitivity. The remaining two (cases 11 and 12) were PPR-negative, had no history of spontaneous seizures other than ESGS, and experienced repeated ESGS. Therefore, it is highly likely that in these cases, ESGS were induced by a factor other than photo-pattern sensitivity.
In a survey conducted by Quirk et al. (1) throughout England, and based on EEGs and other findings, it was believed that 47 (40%) of 118 patients who experienced ESGS for the first time represented incidental seizures. In the same survey, incidental seizures were defined as those in which there was no PPR provocation and they were not classified as types 1 through 4, according to Waltz et al. (3); those with no subsequent occurrence of other photic-induced seizures or no occurrence of seizures in association with playing ESGs with subsequent exposure; and those with no occipital epileptiform discharges in the resting EEG. Only two patients (12%, cases 10 and 14) in the present study matched these descriptions. Conversely, 64 (54%) patients in the survey by Quirk et al. were judged to represent the condition in which photosensitivity was involved in the pathogenesis of ESGS. Their study was community based and restricted to patients with a first seizure while playing ESGs. In contrast, our study is hospital based, including not only the patients experiencing their first seizure but also patients with two or more seizures. A simple comparison was believed to be inadequate, but in the present study, 13 (76%) patients were placed in this category—the aforementioned 12 in whom photo-pattern sensitivity was presumably the cause of their seizures; and one (case 13) in whom photo-pattern sensitivity was not confirmed but who had a history of seizures provoked by visual stimulation other than ESGs.
The present study was extended to identify a diagnostic test that would be suitable for detecting PPR positivity. The four patients with negative responses to conventional IPS reacted positively to the CRT-pattern test used in this study. Specifically, cases 1 and 7 showed no response to either IPS or the strobe-pattern test, but both were PPR positive on the CRT-pattern test. Case 2 showed a negative response to IPS but responded positively to the strobe-pattern test and the conditions of the CRT-pattern test. Case 8, who did not undergo strobe-pattern testing, reacted negatively to IPS but showed a positive response to the CRT-pattern test. Moreover, cases 7 and 8 responded positively to the CRT-pattern test even while receiving VPA treatment. These observations indicated that all four patients had only pattern sensitivity (i.e., not photosensitivity), or that the CRT-pattern test, followed by the strobe-pattern test, and finally the IPS test, was useful for detecting PPR-positivity. In Japan, the Nihon-Kohden photic stimulator is the most common means of delivering IPS. Although we did not compare the PPR-eliciting effectiveness of this device with those of other stimulators, such as the Grass and Nihon-Kohden types, or the former and the CRT monitor, the present results suggest the CRT-pattern test to be more useful than the conventional IPS test. Thus, by introducing special stimulation tests (such as the CRT-pattern test), it may be possible to discover patients with photo-pattern sensitivity among those who have been considered to represent incidental seizure or nonphotosensitive seizure, based on the results of conventional IPS testing.
There has been a paucity of reports on the incidence of ESGS not caused by photosensitivity. According to the survey by Quirk et al. (1), seven (6%) patients had recurrent seizures on repeated exposure to ESGs without electroclinical evidence of photosensitivity. By introducing this special stimulation test, it may be possible to obtain more accurate data on this incidence. In this study, the incidence of non-photosensitive ESGS was 12% (two of the 17 cases).
Pathophysiology of ESGS
In this study, pattern stimulation provided by the CRT-pattern test was divided according to three elements, spatial resolution, brightness perception, and pattern-movement recognition. These three elements are examined when processing information on visual perception. According to Livingstone and Hubel (12), the visual information-processing pathway, which begins at the retinal level and reaches the cerebral cortex—including the primary and succeeding fields of visual perception—can be roughly divided into two parts, each of which is involved in selective information-processing tasks.
The first is the P-cell pathway represented by the parvocellular layers of the lateral geniculate body. It receives information projected from the small cells of the retina and ends at the 4Cβ layer of the primary visual field. This pathway then divides, and the information is transmitted to blobs, which exist in layers 2 and 3, and are darkly stained by cytochrome oxidase (CO), or to an interblob. From the blobs, the information is further transmitted to the CO-positive thin stria of the secondary visual field, and then to the higher V4 field, which is involved mainly in processing color perception (partly for brightness and form). From the interblob, information is projected to the CO-negative sections between the stria of the secondary visual field, and then probably to the V4 field, which is involved mainly in processing information on form and its edges.
The second is the M-cell pathway represented by magnocellular layers of the lateral geniculate body. It receives information projected from the large cells of the retina and ends at the 4Cα layer of the primary visual field. From there, the information is transmitted directly to the higher MT area (middle temporal area) through the 4B layer; or it is projected from the 4B layer to the MT area through the CO-positive thick-stria section of the secondary visual field. The MT area is involved in processing information concerning form, movement, and stereoscopic characteristics.
Processing through the blobs in the P-cell system excels in color selectivity and brightness perception but is inferior in spatial resolution or speed perception. The pathway through the interblobs is color selective and outstanding in spatial resolution but is less sensitive in brightness perception or speed perception. The M-cell pathway, on the other hand, excels in brightness perception or speed perception; but it lacks color selectivity and is inferior in spatial resolution (Table 4).
In an isoluminescence test for the study of visual perception, the M-cell pathway, which transmits information on differences in brightness including distance perception, stereoscopic view, and perception of pattern movements, is ineffective when patterns of equal brightness are compared. In a test designed to recognize fine patterns with high spatial frequencies, perception by the M-cell pathway or the P-cell pathway, especially via blobs, which are inferior in spatial resolution, become invalid (i.e., perception of pattern movements and color perception).
Zeki (13) proposed two similar systems in regard to visual information processing. Specifically, the P-cell pathway pertains to color vision and the perception of form that is related to color, whereas the M-cell pathway is related to perception of the form and position of moving objects (movement perception). However, this theory suggests that these pathways are not independent: instead they advance to a higher level while exchanging information and jointly compiling the visual information thus obtained.
According to the studies by Wilkins and Binnie et al. (14–16), oriented lines or oscillating patterns are considered to be more epileptogenic than are checkerboards or static patterns. A spatial frequency of two to four cycles per degree and a reversal frequency of 10–20 Hz have also been shown to be most epileptogenic. Thus the results of the present study are not inconsistent with those of their experiments.
Binnie and Wilkins (16) also stated that the M-cell pathway plays an important role in the two visual information-processing systems described earlier in relation to the pathophysiology of pattern-sensitive epilepsy. They cited the following factors to illustrate the importance of this pathway: (a) the patterns of stripes that differ in brightness are far more epileptogenic than are those with stripes that differ only in color, strongly suggesting the participation by the M-cell pathway without color information; (b) movement perception and binocular vision pertain to information that is processed by the M-cell pathway; (c) flickering pattern stimuli with low spatial frequencies and frequencies >20 Hz are likely to provoke EEG abnormalities, and inferior spatial resolution and detecting speed are characteristic of the M-cell pathway; and (d) EEG abnormality provoked by pattern stimulation, when focal, is usually dominant at the parietal region, which is consistent with the M-cell pathway projecting to the parietal lobe.
Conversely, Harding and Fylan (17) emphasized linear luminance contrast dependency in PPR, in addition to color sensitivity, because they believed that PPR is generated by the P-cell pathway. Porciatti et al. (18) studied patients with photosensitive epilepsy by recording visually evoked potentials in response to temporally modulated patterns of different contrast. Their results also indicated that pattern stimuli of relatively low temporal frequency and high luminance contrast may operate in the cortical hyperexcitability of photosensitive epilepsy (18).
According to the results of the present study, the circumstances for PPR induction by the CRT-pattern test can be described as follows: (a) percentage PPR induction differed according to pattern, with the vertical stripe pattern provoking this response most frequently; (b) percentage PPR provocation also differs depending on spatial frequency (highest at 2.0 to 4.0 cpd); (c) percentage PPR differs depending on pattern-reversal frequencies (highest at 20 Hz); (d) the clarity of the edges of a pattern has no bearing on percentage PPR induction; (e) changes in the brightness of pattern elements do not affect percentage PPR induction; and (f) with the exception of the two subjects who showed dominance of the frontal region, the stimulation caused by pattern movement is not effective in eliciting PPR.
When these findings are superimposed on the characteristics of the two visual information-processing systems described earlier, the P-cell pathway, especially the one traversing the interblob, rather than the M-cell pathway, appears to correspond more closely to the process of PPR provocation. In particular, the lack of significance of brightness perception and the importance of high spatial resolution in provoking PPR differ from the findings of Binnie et al (16). In the present study, spatial frequencies produced the most significant difference in provoking PPR. In other words, PPR was induced more readily with stimulation by finer patterns, and pattern types altered percentage PPR induction. These responses illustrate the characteristics of the P-cell (interblob) pathway that are outstanding in spatial resolution and are essential for “form” perception. Lack of involvement of changes in the brightness and the relative insignificance of pattern movement noted in the test results further illustrate the characteristics of the P-cell (interblob) pathway. However, the observation that the definition of the edges of a pattern is not important in PPR provocation is not consistent with the known characteristics of this pathway. This discrepancy can be explained as follows: when the spatial frequency exceeds 2.0 cpd, it becomes difficult to perceive a difference in the definition of an edge because the pattern is too fine. Another finding, the correlation between pattern-reversal frequency and percentage PPR induction, also is contrary to the characteristics of the P-cell pathway, which is not sensitive to speed. However, pattern-reversal velocity is not believed to be a stimulus related to the speed of pattern movement; rather, like a dot-flickering stimulus, it is related to the synchronicity of the stimulus (i.e., the P-cell pathway is stimulated repeatedly and at regular intervals) and the activation of neurons is synchronized at the cerebral level and induces PPR. In fact, percentage PPR induction is significantly higher at 20 Hz than at lower frequencies, although there was no significant difference between 30 Hz and the lower frequencies. This finding is in agreement with the results of an earlier study indicating that PPR is readily provoked with flickering-flash stimulation at 15 to 20 Hz (19). In other studies, when the number of PPR-positive patients was compared in terms of pattern-reversal frequencies, no significant difference was found. Therefore, as compared with spatial frequencies and pattern types (forms), pattern-reversal frequencies do not constitute an important element in PPR induction.
The present cases showing PPR on the CRT-pattern test were divided into two groups, one with occipital dominance, comprising the majority, and the other with frontal dominance. The two groups reacted differently to pattern stimulation. The latter group included cases in which stimulation with a greater change in the brightness evidently resulted in a higher frequency of PPR or localized spike-and-wave complexes, suggesting the importance of sensitivity to changes in the brightness. When the pattern types were compared, the latter group also included cases in which stimulation caused by a black-and-white pattern or a yellow–blue checkered pattern with little change in the brightness did not provoke PPR; only stimulation by a yellow–black checkered pattern with a greater change in the brightness provoked PPR. This finding suggests that sensitivity to changes in the brightness or color elements (not studied here) has some effect on PPR induction in the latter group. The stimulation caused by pattern movements provoked PPR or localized spike-and-wave complexes in only those cases belonging to the group with frontal dominance, which casts some doubt on the correlation between PPR and pattern movements in this group. Binnie et al. (15) emphasized that the drifting pattern is less epileptogenic than the oscillating and phase-reversing patterns (15). The spatial resolution and brightness-perception tests in the present study used a phase-reversal stimulation method, whereas the pattern movement test used strictly the drifting pattern. The poor activation rate in our pattern-movement test does not contradict that reported by Binnie et al. However, the pattern-movement test, despite producing a less epileptogenic result, might be related to the frontal dominant EEG activation shown in our frontal dominant group.
In examining the significance of spatial frequencies and pattern types in provoking PPR in each individual, most of the cases belonging to the group with occipital dominance showed a significant difference. The correlation obtained between PPR and spatial resolution in this group is probably of importance.
Because the number of patients is small, hasty conclusions are unwarranted. However, the results of this study allow us to hypothesize that patients with ESGS caused by photosensitivity can be divided into two groups: one with occipital dominance for PPR provocation in whom the P-cell (interblob) pathway, which is outstanding in spatial resolution among visual perceptions, is involved; and the other group with frontal dominance for PPR provocation, in whom the M-cell pathway, which is outstanding in brightness perception or perceiving pattern movements among visual perceptions, or the P-cell (blob) that is outstanding in brightness perception or perceiving colors, is involved. Mishkin and Ungerleider (20) stated that the M-cell pathway projects to the parietal cortex. Binnie et al. (16) also emphasized that PPR provoked by pattern stimulation is usually dominant in the parietal region, which is consistent with their observations. However, Goldman-Rakic (21) also suggested that there is a further projection from the parietal to the frontal cortex. Thus our speculation that the M-cell pathway is involved in the group with frontal dominance may not contradict the Mishkin concept. As the number of patients is still too limited to confirm this theory, and our interpretation of the results is inconclusive, especially for the frontal dominant group, it will be necessary to amass data from a larger patient population.
Prevention of ESGS
To prevent ESGS, it is necessary to pay sufficient attention to the physical properties characteristic of the television screen as well as the nature of the visual stimulation associated with the type of game, if the condition is caused by photo-pattern sensitivity. It is difficult to determine who among these 17 patients might have had seizures due to exposure to the television screen itself and how many would have the potential for seizures only in the setting of very specific additional material. If one experienced recurrent seizures induced by a specific game, the cause would be not the television screen itself, but the nature of the visual stimulation associated with the game. In the present study, all nine patients with a positive CRT-pattern test experienced only one episode of ESGS. However, in cases 4, 6, and 7, their ESGS were induced by games on portable LCDs. Generally, LCD shows the effect of afterimage, such that it seems to be less flickering. Furthermore, the LCD itself does not radiate, but rather reflects, the surrounding room light. Thus the cause of ESGS in these three patients might be not the inherent flickering of the television screen itself, but rather the nature of the visual stimulation associated with the game.
In Japan, television sets are manufactured under the NTSC (National Television System Committee) system, with a flickering component of 60 Hz and 30 frames/s. Fylan and Harding (22) described the use of a high frame-rate television as possibly being beneficial in reducing the risk of ESGS caused by photosensitivity. However, it will be more important for patients, who are sensitive to the specific pattern itself within the video game, to avoid the following: (a) geometric patterns (especially vertical stripes) that occupy most of the display; (b) fine patterns with a spatial frequency exceeding 2.0 cpd; (c) rapid (around 20 Hz) pattern-reversal stimulation; (d) patterns with large differences in the brightness; and (e) rapid pattern movements. The establishment of universal guidelines with these recommendations in mind, when creating software for video games, is anticipated to be effective in preventing ESGS caused by photo-pattern sensitivity.