Takao Yamasaki and Yoshinobu Goto contributed equally to the present paper.
Address correspondence to Takao Yamasaki, M.D., Ph.D., Department of Clinical Neurophysiology, Neurological Institute, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. E-mail: firstname.lastname@example.org
Purpose: To determine a psychophysiological basis for age visual sensitivity to chromatic and achromatic stimuli.
Methods: We investigated the effects of achromatic and four isoluminant color combinations (blue/red, blue/green, green/red, and blue/yellow), luminance ratio changes in color combinations (blue/red; 1:1, 3:4, 4:3) and contrast changes (3 to 100%) on steady-state electroretinograms (ERGs) and visual evoked potentials (VEPs) in 32 healthy teenagers and 30 young adults.
Results: We found that (1) dual peaks at 9 and 18 Hz with a dip at 12 Hz were observed in VEPs with all isoluminant color combinations, (2) VEP responses were significantly enhanced and the 12-Hz dip became unclear with luminance ratio changes between two colors with a nonantagonistic relationship (blue/red), and (3) VEP amplitudes were significantly increased when the contrast was increased. These characteristics were more evident in teenagers than young adults; however, ERGs were qualitatively similar between the two groups.
Discussion: The visual cortex is differently modulated by different color-luminance combinations, and higher sensitivity to color-luminance combinations in the visual cortex in teenagers is responsible for the high prevalence of photo/chromatic sensitivity in adolescence.
Photosensitive epilepsy (PSE) is a well-known condition characterized by seizures in patients who show photoparoxysmal responses (PPRs) on EEG elicited by intermittent photic stimulation (Harding & Jeavons, 1994; Kasteleijin-Nolst Trenité, 1998). On the other hand, photosensitivity is defined by abnormal EEG responses to light or pattern stimulation, and consisting of a PPR (Fisher et al., 2005). The estimated prevalence of seizures from light stimuli is ∼1 per 10,000, or 1 per 4,000 individuals age 5–24 years, while photosensitivity occurs in ∼0.3–3% of the population (Fisher et al., 2005). Television (TV) is the most common provocative stimulus for PSE (Parra et al., 2005). For example, a UK TV commercial film precipitated epileptic seizures in three viewers in 1993 (Fisher et al., 2005). In 1997, 685 Japanese children suffered epileptic seizures while watching a popular animated TV program (Pocket Monsters: Pokemon) (Harding, 1998; Hayashi et al., 1998; Tobimatsu et al., 1999). The key scenes in that program consisted of rapid blue/red (B/R) frame changes (temporal frequency, 12 Hz; luminance ratio, B:R = 4:3) (Harding, 1998; Tobimatsu et al., 1999). Interestingly, 76% of patients having “Pokemon seizures” had no previous history of epilepsy (Fisher et al., 2005), and 81% of these had no recurrence of epileptic seizures in a 5-year follow-up study (Okumura et al., 2004). Thus, healthy people can suffer from seizures induced by visual images. Therefore, it is important to examine the neural basis of latent color-luminance sensitivity in healthy people to prevent epileptic seizures occurring when watching TV.
In the present study, we focused on the neural basis of color-luminance sensitivity in healthy people; therefore, we investigated the effects of isoluminant color combinations, luminance ratio changes in color combinations and contrast changes on steady-state electroretinograms (ERGs) and visual evoked potentials (VEPs) in healthy teenagers and young adults. EEG was also monitored during the ERG and VEP recordings.
Thirty-two healthy teenagers (17 males and 15 females, aged 12–16 years) and 30 healthy young adults (18 males and 12 females, aged 23–40 years) were studied in different combinations in the four experiments. We defined healthy subjects as those with no color blindness by Ishihara color plates (Ishihara, 1997), no loss of visual acuity, no family history of epilepsy, no history of a febrile convulsion, and no neurological disorders. Informed consent was obtained after the nature of the experiment had been fully explained. The experimental procedures were approved by the ethics committee of the Graduate School of Medical Sciences, Kyushu University.
The stimuli were generated by a VSG Three (Cambridge Research Systems, Cambridge, U.K.) and displayed on a gamma-corrected color monitor with a frame rate of 100 Hz (GDM-17SE2T, SONY, Tokyo, Japan). We used achromatic (black/white, Bk/W) stimulation and four isoluminant color combination stimuli [no antagonistic color (blue/red, B/R and blue/green, B/G), and antagonistic color (green/red, G/R and blue/yellow, B/Y)] with a circular field of 10° at a viewing distance of 114 cm (Fig. 1A). CIE coordinates (measured by a Chromameter CS 100, Konica, Minolta, Tokyo, Japan) were x = 0.391, y = 0.347 (black); x = 0.308, y = 0.342 (white); x = 0.290, y = 0.621 (green); x = 0.620, y = 0.353 (red); x = 0.166, y = 0.162 (blue); x = 0.422, y = 0.510 (yellow). Figure 1B shows the wavelength sensitivity of cones, while Fig. 1C reveals the wavelength of each color in our stimuli measured by a photometer (Colormeter III F, KONICA MINOLTA, Tokyo, Japan), demonstrating that the wavelengths of B, G, and R were fitted to the peak spectral sensitivity of each cone. Visual stimuli were surrounded by a homogeneous background containing a mixture of color combinations (Fig. 1A). The mean luminance of the visual stimuli and homogeneous background was 24 cd/m2. Before the experiment, subjects viewed each color combination stimulus alternating at 15 Hz to establish psychophysical isoluminance, and adjusted the relative luminance to minimize perception of flicker.
All visual stimuli were presented for 30 s. First, the temporal frequency was varied from 3 to 24 Hz in 3-Hz steps to elucidate the temporal frequency characteristics of VEPs in response to achromatic and isoluminant chromatic stimulation. Second, the contrast of the stimuli was varied to study the effects of contrast changes. Bk/W and B/R stimuli were used and the temporal frequency was held constant at 12 Hz. Contrast was varied as follows: 3%, 5%, 10%, 20%, 30%, 40%, 60%, 80%, 90%, and 100%, with the mean luminance held constant at 24 cd/m2. Achromatic contrast was defined as the Michaelson contrast: , where Lmax is the maximum luminance and Lmin is the minimum luminance. Chromatic contrast was defined as , and C1 = (r1, g1, b1) and C2 = (r2, g2, b2), where C1 and C2 represent each color in color combinations, and r, g, b are the elements of color. Third, the luminance ratio of the stimuli was varied to study the effect of luminance changes on color combinations. B/R and G/R stimuli were used, and their ratios were as follows: B (or G):R = 1:1, 4:3, 3:4. Subjects sat on a chair in a dark room, and fixated on a fixation point (visual angle 0.2°) in the center of the monitor. To prevent the triggering of a PPR during the experiment, only one eye was stimulated (Harding & Jeavons, 1994; Takahashi, 2002).
ERG, VEP, and EEG recordings
Steady-state ERGs were recorded from a surface skin electrode under the stimulated eye, referring to the electrode of the lateral orbital rim on the stimulated eye (Marmor et al., 2004). Pupils were not dilated. ERG recording was performed following the method of the International Society for Clinical Electrophysiology of Vision (Marmor et al., 2004). Steady-state VEPs were recorded from scalp electrodes placed over Oz, O1 and O2, referring to an electrode at Cz (International 10–20 system). The ground electrode was placed at Fz. Electrode impedance was maintained below 5 Kohm. The EEG signals were analog filtered between 0.5 and 200 Hz. EEG was also monitored during the ERG and VEP recordings using Neurofax EEG-1100 (Nihon Kohden, Tokyo, Japan).
The analog data were digitized at a sampling rate of 1000 Hz/channel and 15 samples of 2000 ms epoch were averaged using customized software (ERP average, Medical Try System, Tokyo, Japan). Epochs containing EEG deviations from the baseline greater than 100 μV were automatically rejected. The averaged responses were subjected to fast Fourier transform (FFT). The FFTs yielded the amplitude (square root of the power) of several harmonic components, but the first harmonic response (1F) was the most predominant. Thus, we mainly analyzed the 1F component. A three-way analysis of variance (ANOVA) with repeated measures was performed. Multiple comparisons with Bonferroni correction were also conducted for paired comparisons. For the contrast changes, analysis of covariance was performed.
In both groups, EEG showed well-organized posterior dominant alpha activity (9–13 Hz) and no asymmetry in background activity. No spontaneous paroxysmal activity was observed. EEG showed photic-driving responses time-locked to the visual stimuli, however, no PPRs were found.
Different sensitivities to isoluminant color combinations in teenagers
Fig. 2 shows representative waveforms of VEPs to Bk/W stimulation from a teenager and a young adult, and their FFTs at each temporal frequency. VEPs were characterized by quasisinusoidal waveforms that corresponded to each temporal frequency. FFTs revealed that the first harmonic component (1F) was predominant. Fig. 3 shows the temporal frequency characteristics of the 1F amplitude to achromatic and color combination stimuli with no antagonistic relation (Fig. 3A, B) and with an antagonistic relation (Figs. 3C, D). Mean 1F amplitude in the teenagers was larger than that in young adults at all temporal frequencies in any color combination [F (1,18) = 7.82, p < 0.01]. The main effect of color combination stimuli was determined [F (4,72) = 2.51, p < 0.05]. In the teenagers, two amplitude peaks at 9 and 18 Hz with a dip at 12 Hz were observed in all color combinations (Figs. 3A, C). The amplitude of the 9-Hz peak was different among VEPs to color combinations (p < 0.01). The peak amplitude of Bk/W stimulation was the greatest and those of B/R and G/R were the lowest among the five color combinations. The 18-Hz peak of Bk/W and B/G stimuli tended to be larger than that of the other color combinations (p = 0.061). In contrast, young adults showed a 9-Hz peak, with a less defined 18-Hz peak due to low amplitudes (Figs. 3B, D). At the 9-Hz peak, the amplitude was greater with Bk/W stimulation than with the other color combinations (p < 0.01).
Higher sensitivities to contrast changes in teenagers
Fig. 4 shows the contrast modulation functions of VEP amplitudes to B/R and Bk/W stimuli in teenagers (Fig. 4A) and young adults (Fig. 4B). In both groups, VEPs showed a gradual increase in amplitude as a function of contrast in both stimuli [F (1,396) = 112.2, p < 0.001]; however, their slopes were much steeper in teenagers [F (1,396) = 84.2, p < 0.001]. These slopes tended to be saturated at about 40% with Bk/W stimulation and at about 80% with B/R stimulation in both groups. There was no significant difference in VEP amplitudes between the two stimuli in either group.
Luminance ratio changes in color combinations alter visual cortical responses
Figs. 5A and B show the response characteristics of VEP amplitudes to isoluminant and anisoluminant B/R stimuli. Mean VEP amplitude in teenagers was larger than that in young adults at all temporal frequencies of any color combination stimuli [F (1,21) = 16.51, p < 0.005]. Surprisingly, B/R with a low luminance of red (B > R) stimulation resulted in a larger amplitude and a less marked dip at 12 Hz compared with isoluminant B/R (B = R) stimulation in both teenagers (p < 0.05) and young adults (p < 0.05). In contrast, the VEP response to B/R with a high luminance of red (B < R) stimulation was lower than that of B = R stimulation in teenagers (p < 0.005). In young adults, there was no difference in the VEP amplitude between B < R and B = R stimuli.
Figures 5C and D show the temporal frequency characteristics of VEP amplitudes to isoluminant and anisoluminant G/R stimuli. Mean VEP amplitude in teenagers was larger than that of young adults at all temporal frequencies of all color combinations [F (1,21) = 11.07, p < 0.005]. In the isoluminant G/R (G = R) condition, the dip at 12 Hz was less marked in both groups compared with the isoluminant B/R condition. In teenagers, G/R with a low luminance of red (G > R) stimulation resulted in a larger amplitude (p < 0.005) and no dip at 12 Hz compared with G = R stimulation, while there were no differences in amplitude between G > R and G = R stimuli in young adults. In contrast, there were no differences in amplitude between G/R with a high luminance of red (G < R) and G = R stimuli in both teenagers and young adults.
Differential effects of isoluminant color combinations on the retina and visual cortex
Mean 1F amplitudes of VEPs at 9 Hz were significantly larger than those of ERGs in both teenagers (n = 9) and young adults (n = 10) [F (1,17) = 144.2, p < 0.001]. VEP amplitudes in teenagers were again larger than those of young adults in any color combination (p < 0.001). The difference in VEP magnitude between color combinations was also found in teenagers (p < 0.01) and showed a trend in young adults (p = 0.07). In contrast, the difference in ERG amplitudes between teenagers and young adults was not significant. Furthermore, there was no difference in the order of ERG magnitude between color combinations in either teenagers or young adults.
The visual sensitivities to color combinations, luminance ratio changes, and contrast changes are drastically changed around the age of 20 years. VEP differences were most pronounced in the isoluminant condition at 9 Hz, while there was no difference in ERG amplitudes in any color combinations between the two groups. It is likely that distinct visual sensitivities in teenagers are mainly related to the activities of the visual cortex but not the retinal cells. Although the lateral geniculate nucleus (LGN) also plays an important role in color processing (Derrington et al., 1984; Horwitz et al., 2005), LGN neurons integrate cone inputs linearly, while the responses of neurons in the visual cortex are nonlinear (Horwitz et al., 2005). Thus, the LGN may not contribute to VEP differences under the isoluminant condition.
Our results raise an important question about the mechanism of the significant changes in photo/chromatic sensitivity around the age of 20 years. There is a significant relationship between the development and aging of the visual cortex and visual function. In the visual cortex, the synaptic density is greatest at 8 months (Regan, 1988), and myelination is complete before 6 years (Benes, 1989; Sowell et al., 2003). Moreover, cell loss is not massive and does not occur rapidly with aging (Spear, 1993; Sowell et al., 2003). Psychophysical and electrophysiological studies also show a gradual decline of visual abilities with aging (Porciatti et al., 1992; Spear, 1993; Fiorentini et al., 1996; Knoblauch et al., 2001; Crognale, 2002). Thus, a drastic change in photo/chromatic sensitivity at around 20 years of age cannot be explained by either continued development or the sudden aging of the visual cortex. On the other hand, lower levels of inhibitory amino acids (GABA, taurine) and higher levels of excitatory amino acids in the cerebrospinal fluid are correlated with increased photosensitivity in photosensitive baboons (Lloyd et al., 1986), and saturation of VEP responses to contrast changes are abolished by local application of bicuculline to the cat visual cortex (Morrone et al., 1987). Considering these results together, we hypothesized that the functional development of inhibitory and excitatory neural networks in the visual cortex may suppress photo/chromatic sensitivity after 20 years of age.
Surprisingly, when the luminance ratio in B/R or G/R stimulation was changed, VEP amplitudes were significantly modulated and the 12-Hz dip became unclear (Fig. 5). These findings suggest that a difference in the luminance of the two colors can influence the activating effect for V1. Color information is perceived by three types of cone in the retina: long (L or red), middle (M or green), and short (S or blue) wavelength-sensitive cones. In monkeys, the color tuning of a population of S cone-driven V1 neurons has been studied (Horwitz et al., 2005). The study suggested the presence of color-luminance interaction in V1; blue (S-cone)-yellow (a combination of L and M cones) color opponent signals are enhanced by a nonopponent signal nonlinearly. Therefore, our finding that VEP amplitudes were modulated by luminance ratio changes may be related to a color-luminance interaction. Moreover, responses to the B/R combination were more enhanced than the G/R combination by luminance contrast in this study. This difference probably results from the absence of an antagonistic effect with B/R but not with G/R stimulation (Livingstone & Hubel, 1984).
VEP amplitudes in both B/R and Bk/W stimuli were significantly increased when contrast was increased. This result suggests that the responses of the visual cortex to achromatic and chromatic stimuli are modulated by contrast change, and contrast gain control (Spekreijse et al., 1973; Porciatti et al., 2000) is driven in both stimuli. Interestingly, the slopes for teenagers were much steeper than those for the young adults. Thus, achromatic and chromatic contrast sensitivities are higher in teenagers.
Overall, our results suggested latent photo/chromatic sensitivity in adolescence. Is there any link between our findings and PSE? Recent studies (Porciatti et al., 2000; Parra et al., 2003) suggest that there are different mechanisms active in PSE patients compared to healthy subjects, although color sensitivity is an important factor in the generation of PSE (Harding, 1998; Tobimatsu et al., 1999; Parra et al., 2007). Therefore, our results in healthy subjects may not fully explain the neural mechanism of PSE. However, many of those with “Pokemon seizures” had no previous history of epilepsy, and had no recurrence of epileptic seizures in a 5-year follow-up study (Okumura et al., 2004). These facts suggest that the 12-Hz B/R stimulation (luminance ratio; B:R = 4:3) may be a provocative stimulus for the visual cortex to induce seizures in healthy teenagers. This hypersensitivity of the visual cortex to 12-Hz B/R stimulation (luminance ratio; B:R = 4:3) may explain the VEP response characteristics (an unclear 12-Hz dip in the 12 Hz B > R stimulation) in healthy teenagers (Fig. 5).
In conclusion, color-luminance combinations can significantly influence the excitability of the visual cortex. In particular, higher sensitivities to these visual stimuli in teenagers play an important role in photo/chromatic sensitivity in adolescence.
This work was supported in part by a grant from the Japan Epilepsy Research Foundation. This study was also supported in part by Grant-in-Aid for the 21st Century COE Program and Grant-in-Aid for Scientists, No. 16390253 and No. 16200005 from the Ministry of Education, Culture, Sports, Science and Technology in Japan.
Conflict of interest: 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 declare.