Spatial Luminance Contrast Sensitivity Measured with ... - CiteSeerX

4 downloads 0 Views 418KB Size Report
METHODS. Subjects. Six healthy adults (three men and three women; mean age, 21. 2 ... of the Health National Council of Brazil. ... Presented in part at the annual meeting of the Association for. Research ... gratings, with CIE1976 chromaticity (u , v ). (0.215 ..... Strasburger H, Scheidler W, Rentschler I. Amplitude and phase.
Spatial Luminance Contrast Sensitivity Measured with Transient VEP: Comparison with Psychophysics and Evidence of Multiple Mechanisms Givago S. Souza,1 Bruno D. Gomes,1 Ce´zar A. Saito,1 Manoel da Silva Filho,1 and Luiz Carlos L. Silveira1,2 PURPOSE. To compare the spatial luminance contrast sensitivity function (CSF) obtained with transient visual evoked potentials (VEPs) with that obtained with psychophysical measurements. METHODS. The stimuli consisted of horizontal luminance gratings. In the VEP experiments, 0.4, 0.8, 2, 4, 8, and 10 cpd of spatial frequency were used, at 1 Hz square-wave contrastreversal mode. Eight to 10 Michelson contrasts were used at each spatial frequency. Contrast thresholds were estimated from extrapolation of contrast response functions. Psychophysical sensitivities were obtained with spatial gratings of 0.4, 0.8, 1, 2, 4, 6, 8, and 10 cpd and presented at 1 Hz square-wave contrast-reversal or stationary mode (dynamic and static presentation, respectively). CSF tuning was estimated by calculating the ratio between peak sensitivity and the sensitivity at 0.4 cpd. RESULTS. In all subjects tested (n ⫽ 6), VEP contrast-response functions showed nonlinearities—namely, amplitude saturation and double-slope amplitude functions that occurred at low and medium-to-high spatial frequencies, respectively. Mean electrophysiological and psychophysical CSFs peaked at 2 cpd. CSF tuning for electrophysiology and dynamic and static psychophysics were, respectively, 1.08, 1.11, and 1.31. Correlation coefficients (r2) between electrophysiological CSF and dynamic or static psychophysical CSF were, respectively, 0.81 and 0.45. CONCLUSIONS. Electrophysiological and psychophysical CSFs correlated more positively when temporal presentation was similar. Spatial frequencies higher than 2 cpd showed that at least two visual pathways sum their activities at high contrasts. At low contrast levels, the results suggest that the transient VEP is dominated by the magnocellular (M) pathway. (Invest Ophthalmol Vis Sci. 2007;48:3396 –3404) DOI:10.1167/iovs.070018

T

he use of the visual evoked potential (VEP) to study the human spatial luminance contrast sensitivity started in the 1970s, when it was demonstrated that the steady state VEP amplitude decreases with stimulus log contrast after a straightline relationship.1 Moreover, the steady state VEP was successfully used to derive the spatial luminance contrast sensitivity function (CSF) in humans. When compared with behavioral measurements, the electrophysiological CSF was similar to the psychophysical CSF.1 It was also observed that at spatial frequencies between 1.5 and 3 cpd, the contrast response functions were describable by two regressions with different slopes.1 Further studies in humans and nonhuman primates have reported such nonlinearity of the VEP amplitude as a function of the stimulus contrast2–11: either a straight-line relation for low contrasts followed by saturation at high contrasts or a double-slope straight-line relation such as those described earlier. Originally, it was proposed that these nonlinearities indicate the contribution of different retinal regions,1 but more recent studies have suggested that the nonlinearities are related to different contrast sensitivity mechanisms from parallel visual pathways.6,10 –12 In this study, we assessed the similarity of the spatial luminance CSF obtained with transient VEP, when compared with that obtained psychophysically with similar stimuli. We also analyzed whether transient VEP amplitudes can be related to visual parallel pathway sensitivities. The transient VEP can be used as an objective tool to study spatial luminance CSF. There was good agreement between the electrophysiological and psychophysical measurements obtained with the same temporal frequency. Our results suggest that the M pathway is the main source of the VEP amplitude, at least at low-luminance contrasts. Some of these results have been presented in abstract form (Souza et al. IOVS 2006;47:ARVO E-Abstract 5379).

METHODS From the 1Departamento de Fisiologia, Universidade Federal do Para´, Brazil; and 2Nu ´ cleo de Medicina Tropical, Universidade Federal do Para´, Brazil. Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2006. Supported by Grant CNPq 471815/2004-7 from CNPq-PRONEX/ SECTAM-FUNTEC/FADESP 1079, CAPES-PROCAD 0019/01-1, and FINEP Research Grant “Rede Instituto Brasileiro de Neurocieˆncia (IBNNet)” 01.06.0842-00. GSS and BDG received CAPES-PROF fellowships for graduate students. MSF and LCLS are CNPq research fellows. Submitted for publication January 6, 2007; revised February 24, 2007; accepted May 8, 2007. Disclosure: G.S. Souza, None; B.D. Gomes, None; C.A. Saito, None; M. da Silva Filho, None; L.C.L. Silveira, None The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact. Corresponding author: Luiz Carlos L. Silveira, Universidade Federal do Para´, Nu ´ cleo de Medicina Tropical, Av. Generalı´ssimo Deodoro 92, 66055-240 Bele´m-Para´, Brazil; [email protected].

3396

Subjects Six healthy adults (three men and three women; mean age, 21 ⫾ 2 years) were monocularly tested. For each subject, only the eye with smallest dioptric error was tested. All subjects had normal or corrected acuity to 20/20, which was assessed by measuring the eye’s refractive state with an autorefractor/keratometer (Humphrey 599; Carl Zeiss Meditec, Dublin, CA). None of the participants reported previous ocular, neural, or systemic diseases that could affect the visual system. All procedures were performed according to the tenets of the Declaration of Helsinki and were approved by the Committee for Ethics in Research, Tropical Medicine Institute, Federal University of Para´ (Report 113/2004, November 25, 2004), according to Resolution 196/96 of the Health National Council of Brazil.

Visual Stimulation We used visual-stimulus software (Psycho ver. 2.6; Cambridge Research Systems, Rochester, UK) to drive a graphic card (VSG 2/3; Cambridge Research Systems). The stimuli were presented on a 20-in. Investigative Ophthalmology & Visual Science, July 2007, Vol. 48, No. 7 Copyright © Association for Research in Vision and Ophthalmology

IOVS, July 2007, Vol. 48, No. 7 color monitor (FlexScan T662, Eizo, Ishikawa, Japan) with a frame rate of 100 Hz and a spatial resolution of 800 ⫻ 600 pixels. Gamma correction was performed with a photometer (OptiCAL OP200E; Cambridge Research Systems). Visual stimuli were similar throughout the electrophysiological and psychophysical tests, consisting of achromatic horizontal sinusoidal gratings, with CIE1976 chromaticity (u⬘, v⬘) ⫽ (0.215, 0.480) and 40 cd/m2 of mean luminance and were presented in a circular patch (5° of visual field). Mean luminance and chromaticity of the background were identical with those used in the grating stimuli. The subjects were placed at a distance of 1 m from the screen monitor. A central cross (1° of visual field) was used as a reference point. For the electrophysiological experiments, the stimuli were presented in a 1-Hz square-wave contrast-reversal mode. Spatial frequencies were 0.4, 0.8, 2, 4, 8, and 10 cpd. For each spatial frequency, 8 to 10 Michelson contrasts were tested ranging from 0% to 100% after a pseudorandom presentation. For the psychophysical procedures, we added two additional spatial frequencies, 1 and 6 cpd, and performed the tests at two temporal conditions: a 1-Hz square-wave contrast reversal and a stationary stimulation mode, which we will refer to in this study as a dynamic or a static presentation, respectively.

Electrophysiological Procedure Recordings were performed by using three gold-cup electrodes placed at Oz (active electrode), Fz (reference electrode), and Fpz (ground electrode) according to the International Society of Clinical Electrophysiology of Vision.13 The bioelectric potentials were amplified by 50,000⫻ and online band-pass filtered between 0.5 and 100 Hz by a differential amplifier (MAS800; Cambridge Research Systems). The amplified potentials were digitized by a data-acquisition card (AS-1; Cambridge Research System) at 1 kHz of sample rate, with 12 bits of resolution, controlled by a third-part software (Optima ver. 1.4; Cambridge Research Systems). For each stimulus, the number of sweeps recorded over time (240 –960 of 950 ms each) varied according to the contrast and subject amplitude response, to improve the signal-to-noise ratio. Each sweep was submitted offline to a home-made Fourier transform to extract even harmonics between 0.5 and 40 Hz, which were used to synthesize the sweep. A final waveform was obtained by averaging these synthesized sweeps. The P100 amplitude of this final VEP waveform was measured peak-to-baseline and taken as the subject response at each spatial frequency and contrast condition. The contrast threshold for each spatial frequency was estimated by fitting a regression line to the amplitude data as a function of log contrast and extrapolating it to the 0 amplitude level.

Psychophysical Procedure Both dynamic and static conditions were tested. The spatial frequencies were presented in a pseudorandom sequence set up by software. The subjects controlled the Michelson contrast with a step of 1.15% by using a CB1 interface (Cambridge Research System) connected to a computer. The system allowed us to find their contrast threshold using the method of adjustment. The contrast value in the beginning of each trial was always suprathreshold and randomly selected by the software to avoid intertrial adaptation. For each spatial frequency, subject’s contrast sensitivity was estimated from the averaged contrast threshold obtained from six individual trials. In the dynamic condition, subjects were instructed to consider the flicker detection as an indication of visibility. In the static condition, they were instructed to consider pattern detection as a positive signal of visibility.

RESULTS Electrophysiology Reliable transient VEP recordings were obtained from all six subjects in the range of spatial frequencies tested. Stimulus

Contrast Sensitivity Measured with Transient VEP

3397

contrast could be considerably lowered for these spatial frequencies, and the response was still present and robust enough to be analyzed, except at 0% of Michelson contrast, as expected. In this latter condition, in all subjects, the VEP waveforms were different from those obtained with higher contrasts. Figures 1, 2 and 3 illustrate transient VEP recordings showing complete sets of responses obtained from three different subjects. In these subjects, we averaged 240 sweeps for contrast between 20% and 100% and 480 sweeps for contrast below 20%, to achieve a similar signal-to-noise ratio among the range of contrasts tested. Some subjects presented a low signalto-noise ratio, and final waveforms were obtained with 720 to 960 sweeps for low contrasts (⬍ 5%) and medium-to-low spatial frequencies (⬍4 cpd). As previously found by Plant et al.14 the transient VEP waveform changed substantially along the range of spatial frequencies tested. The VEP at 0.4 cpd had a broad waveform, which became increasingly sharp when spatial frequency was increased (Figs. 1–3). This can also be described as a differential contribution of different ranges of even harmonics to the VEP waveform (Fig. 4). We have quantified this phenomenon by comparing the normalized mean energy of each even harmonic at different spatial frequencies for the six subjects tested (one-way ANOVA, ␣ ⫽ 0.05, and post hoc comparisons, Tukey test). Two ranges of even harmonics were identified by using this approach. When the spatial frequency was increased, there was a progressively smaller contribution of lower harmonics and a larger contribution of higher harmonics (see also Refs. 15,16). The even harmonics between 2 and 10 Hz had more energy at low spatial frequencies (0.4 and 0.8 cpd) than at intermediate and high spatial frequencies (4, 8, and 10 cpd) (P ⬍ 0.01). Conversely, the contribution of even harmonics to the VEP between 16 and 30 Hz was smaller at low spatial frequencies and larger at intermediate and high spatial frequencies, reaching an energy peak at 8 cpd (P ⬍ 0.05). The 12- to 14-Hz even harmonics exhibited an intermediate behavior between 2 and 10 Hz and 16 and 32 Hz, whereas the higher harmonics, 34 to 40 Hz, showed a negligible contribution to the VEP waveform that changed little with spatial frequency (P ⬎ 0.05). The P100 implicit time peaked between 100 and 130 ms from the grating phase reversal in the range of spatial frequency and contrast that we studied. As the contrast decreased, the P100 amplitude decreased and the amplitude peak occurred at a longer implicit time (Figs. 1–3). The P100 contrast–response function has a distinct behavior along the spatial frequency domain. For the lowest spatial frequency tested, 0.4 cpd, the P100 amplitude saturated at high contrasts (Fig. 5A). This was also observed for 0.8 cpd in most of the individuals (n ⫽ 4, not illustrated). At intermediate spatial frequencies, 2- and 4-cpd, double-slope straight-lines were the best descriptors for the P100 contrast–response function, so that the data points corresponding to the high and low contrasts could be best fitted by two different regression lines (Figs. 5C, 5D). At the highest spatial frequencies tested, 8 and 10 cpd, double-slope functions were also good descriptors, but the differential behavior of the P100 amplitude at low and high contrasts was less marked than at the intermediate spatial frequencies (Figs. 5E, 5F). To estimate contrast thresholds, we excluded those data points corresponding to amplitude saturation for 0.4 and 0.8 cpd, as well as those belonging to the high-contrast limb for spatial frequencies between 2 and 10 cpd (Fig. 5). The remaining data points were then fitted by regression functions which were extrapolated to the zero amplitude level to find the contrast threshold for each spatial frequency (Fig. 5). The inverse of the contrast threshold, the contrast sensitivity, was calculated for each spatial frequency, and used to obtain the

3398

Souza et al.

IOVS, July 2007, Vol. 48, No. 7

FIGURE 1. VEP recordings at different contrast levels obtained from subject MGL041104 at 0.4 (A), 0.8 (B), 2 (C), 4 (D), 8 (E), and 10 (F) cpd. Average from 240 sweeps between 20% and 100% contrast and average from 480 sweeps at 16% contrast or less. The waveforms at 0.4 cpd are broad and became increasingly sharp at 8 to 10 cpd.

electrophysiological CSF for each individual. The mean contrast sensitivity in log units for the six individuals tested was 1.94 ⫾ 0.18 at 0.4 cpd, 2 ⫾ 0.22 at 0.8 cpd, 2.11 ⫾ 0.13 at 2 cpd, 2.05 ⫾ 0.18 at 4 cpd, 1.26 ⫾ 0.16 at 8 cpd, and 1.07 ⫾ 0.12 at 10 cpd. These values are plotted in Figure 6 as filled circles that represent the mean electrophysiological CSF. The function has a band-pass shape, with a slight increase from 0.4 to 2 cpd and then decreasing toward high spatial frequencies. The function tuning (peak contrast sensitivity/contrast sensitivity at 0.4 cpd) was 1.08.

Psychophysics The results obtained with dynamic or static visual stimulation were different in some aspects. For dynamic stimulation (Fig. 6A, empty squares) the mean contrast sensitivity in log units for the six subjects at the spatial frequencies tested was: 1.97 ⫾ 0.12 at 0.4 cpd, 2.1 ⫾ 0.11 at 0.8 cpd, 2.14 ⫾ 0.09 at 1 cpd, 2.19 ⫾ 0.11 at 2 cpd, 2.19 ⫾ 0.1 at 4 cpd, 2.08 ⫾ 0.1 at 6 cpd, 1.93 ⫾ 0.14 at 8 cpd, and 1.62 ⫾ 0.21 at 10 cpd. For static stimulation (Fig. 6B, empty triangles) the mean contrast sensitivity was: 1.68 ⫾ 0.13 at 0.4 cpd, 1.93 ⫾ 0.21 at 0.8 cpd, 2.01 ⫾ 0.2 at 1 cpd, 2.2 ⫾ 0.18 at 2 cpd, 2.20 ⫾ 0.15 at 4 cpd, 2.05 ⫾ 0.13 at 6 cpd, 1.9 ⫾ 0.17 at 8 cpd, and 1.56 ⫾ 0.28 at 10 cpd. The mean dynamic psychophysical CSF had a function tuning of 1.11, whereas the mean static psychophysical CSF had a function tuning of 1.31.

Comparison of Electrophysiological and Psychophysical Results Statistical analysis was performed using one-way ANOVA (␣ ⫽ 0.05) and post hoc comparisons (Tukey test) to compare the spatial luminance contrast sensitivity estimated for each spatial

frequency by using electrophysiological or psychophysical procedures. There was no difference between electrophysiological contrast sensitivity and dynamic psychophysical contrast sensitivity at 0.4 cpd, but both were significantly different from the static psychophysical contrast sensitivity for this spatial frequency (P ⬍ 0.05). At 0.8, 2, and 4 cpd, there was no difference between contrast sensitivity estimated by the three procedures. At 8 and 10 cpd, the psychophysical contrast sensitivity estimated by both dynamic and static procedures were similar (P ⬎ 0.05), whereas the electrophysiological contrast sensitivity was significantly lower than both psychophysical contrast sensitivities (P ⬍ 0.01). The correlation coefficient (r2) between mean electrophysiological CSF and mean dynamic psychophysical CSF was 0.81, whereas the r2 between mean electrophysiological CSF and mean static psychophysical CSF was 0.45 (Figs. 6A, 6B, respectively).

DISCUSSION When the mean electrophysiological CSF was compared with the mean psychophysical CSF, we observed a good correlation between data obtained with the same dynamic condition, independent of the technique that was used. This observation is in agreement with those previously made by other authors, who have indicated the importance of using the same stimulus configuration in the electrophysiological and psychophysical tests to obtain a better correspondence between the results derived from these two procedures.1,7,17,18 A comparison of contrast sensitivity obtained with steady state VEP elicited by sinusoidal spatial gratings at 8 Hz squarewave contrast-reversal with that obtained using psychophysical

IOVS, July 2007, Vol. 48, No. 7

Contrast Sensitivity Measured with Transient VEP

3399

FIGURE 2. VEP recordings in different contrast levels obtained from subject SPG041127 at 0.4 (A), 0.8 (B), 2 (C), 4 (D), 8 (E), and 10 (F) cpd. Average from 240 sweeps between 20% and 100% contrast, and average from 480 sweeps at 16% contrast or less. Similar to Figure 1, when spatial frequency was increased, the VEP waveform became sharper.

procedures and the same stimulus conditions showed similar results at all spatial frequencies tested from 1.5 to 35 cpd.1 The CSFs obtained had high peak values reaching approximately 2.5 log units at 2 cpd. Also, a good correlation was found between steady state VEPs (10 Hz) and psychophysics using a similar procedure.17 CSF peak values ranged between 2 and 2.5 log units at 2 cpd. Both studies1,17 estimated contrast thresholds in the VEP experiments by regressions fitted to the data point and extrapolated to zero amplitude level. In another study the visual system was stimulated with three different temporal frequencies: 6.6, 10, and 20 Hz presented either in square-wave contrast-reversal or onset– offset modes.19 Electrophysiological results were compared with the psychophysical ones by using the same stimulus conditions, but with two psychophysical threshold criteria, flicker detection, and pattern detection. It was found that in contrastreversal mode, VEP recording provided lower contrast sensitivity values than did psychophysical procedures, using both threshold criteria (for instance, 2 vs. 2.5 log units at 2 cpd, respectively), but both CSFs had similar shape with very little attenuation in the low-spatial-frequency range. Electrophysiological CSF obtained with contrast–reversal mode was also similar in shape with onset– offset psychophysical CSF when flicker detection criterion was used. However, when using onset– offset mode, the electrophysiological CSF had a different shape, with strong attenuation at low spatial frequency, similar to the psychophysical CSF obtained by using pattern detection criterion. Contrast-reversal grating was also used at a temporal frequency of 6.75 Hz to estimate electrophysiological and psychophysical CSF with 2°, 4°, and 8° grating stimuli centered on the fovea.20 Electrophysiological CSF was obtained with steady

state VEP. Moreover, 8° gratings centered on the fovea with a 2° and 4° central occlusion were used to obtain parafoveal electrophysiological and psychophysical CSFs. In general, the electrophysiological contrast sensitivity was lower than the psychophysical one, and the fall-off at low and high spatial frequencies was steeper when VEP was used than when psychophysics was used. It should be noted that these studies19,20 estimated contrast threshold with VEP by using a different approach than a straight line fitted to the data points (e.g.: Refs. 1,17, present study). First, the signal-to-noise ratio was determined by the amplitude at the expected response frequency divided by the averaged spectral amplitude across a 3-Hz wide band centered on the second harmonic of the stimulation frequency. In recordings without stimulation, this ratio was ⬍1.25 in 95% of attempts. The VEP was then steadily recorded changing the stimulus contrast until the signal-to-noise ratio crossed the criterion level, the threshold being defined as the contrast 2 or 4 dB below the lowest contrast that exceeded the signal-tonoise ratio criterion level. Another technique, the so-called swept contrast VEP, combines the use of steady state VEP, visual system stimulation with a constant spatial frequency, and a quick lowering of contrast to measure contrast sensitivity; it minimizes the recording periods to retrieve responses near threshold.21–26 Generally, this technique provides lower estimates of contrast sensitivity than do psychophysical and other electrophysiological procedures. A possible explanation for that is the way the contrast–response function is extrapolated to an estimated noise level to find the contrast threshold. Despite several studies in which transient VEPs were used to measure contrast sensitivity in mammals (e.g., Refs. 8,27– 29), the study that more directly can be compared with the

3400

Souza et al.

IOVS, July 2007, Vol. 48, No. 7

FIGURE 3. VEP recordings in different contrast levels obtained from subject ACB041104 at 0.4 (A), 0.8 (B), 2 (C), 4 (D), 8 (E), and 10 (F) cpd. Average from 240 sweeps between 20% and 100%, and average from 480 sweeps at 16% contrast or less. Similar to Figure 1, when spatial frequency was increased the VEP waveform became sharper.

present one is that by Katsumi et al.7 They measured the electrophysiological CSF by recording transient VEP at 2 Hz contrast-reversal stimulation and compared with psychophysical CSF obtained with static gratings. Psychophysical contrast sensitivity was considerably higher than electrophysiological contrast sensitivity at all spatial frequencies. The psychophysical CSF peak was 2.5 log units at 3 to 6 cpd, whereas the peak of electrophysiological CSF was 1.7 log units at the same spatial frequency range. These differences may be due to the different stimulation methods used by Katsumi et al. in electrophysiology and psychophysics (dynamic versus static stimulation in this study, respectively), as well as the smaller number of averaged sweeps used by them (64 sweeps). Some of our results support this interpretation: The difference in the CSF tuning values was smallest when comparing CSF tuning values obtained with VEP and dynamic psychophysics. The correlation coefficient (r2) between electrophysiological and psychophysical CSF was higher when the same temporal condition was used and, at least at low spatial frequencies (⬍ 0.8 cpd), there were no significant differences between the VEP and the dynamic psychophysical CSF. However, one aspect seems to be common to most of the studies that compared psychophysics and electrophysiology for assessing human CSF: Most of the time electrophysiology returns lower values for contrast sensitivity than does psychophysics, never the opposite. We partially found this at high spatial frequencies (⬎8 cpd). One possible explanation for the psychophysical methods’ being more sensitive than the electrophysiological ones, might be found in Barlow’s Neuron Doctrine.30 At threshold, perception might rely on the activity of a few neurons, whereas a large number of active neurons may be needed to exceed the bioelectrical noise and then be recorded in the scalp.

Also, our findings support the hypothesis that the VEP results from the activity of two or more parallel pathways extending from the retina to V1. The cortical response changes in two important aspects when either the spatial frequency or contrast is changed. First, we found that different harmonic bandwidths contribute differentially to the cortical response in the range of spatial frequency tested (Fig. 4). This result indicates the existence of at least two different response mechanisms, partially overlapping and partially differentiated between 0.4 and 0.8 and 2 and 10 cpd. The possibility of different visual mechanisms contributing to the VEP has been explored.15 Nine consecutive wavelets have been found superimposed on the conventional VEP by using high-pass filtering (cutoff at 17 Hz). These wavelets were time locked with reversing patterns and tuned at 5 cpd. Later studies indicate that this oscillatory response (⬎17 Hz) and conventional VEP (⬍17 Hz) may have different functional roles.16,31,32 Second, we found some nonlinearities in the VEP contrast– response functions that can be explained by the contribution of retinogeniculocortical pathways with different contrast sensitivities.33–36 Our results are consistent with a visual pathway highly sensitive to stimulus contrast, such as the magnocellular (M) pathway dominating the VEP at low contrast. Together with the M pathway, the parvocellular (P) and koniocellular (K) pathways are active at high contrasts and may cooperate to generate the VEP at high contrast. At low spatial frequencies, 0.4 to 0.8 cpd, only the M pathway seems to be relevant for the cortical response that saturates at high contrast in the same way as observed in the recording of M cells at the retina and thalamus. The presence of two limbs in the VEP contrast–response function was interpreted as being due to differential contribu-

IOVS, July 2007, Vol. 48, No. 7

Contrast Sensitivity Measured with Transient VEP

3401

FIGURE 4. Normalized power spectra of even harmonics that compose the transient VEP at 0.4, 0.8, 2, 4, 8, and 10 cpd (from top to bottom) and at four contrast levels, 16% to 100% (right to left). Histograms and error bars are the mean power ⫾ SD, from our sample (n ⫽ 6). Odd harmonics and DC components are not shown. The band-width between 2 and 10 Hz decreased in energy when the spatial frequency was increased and the band-width between 16 and 30 Hz showed the opposite behavior.

3402

Souza et al.

IOVS, July 2007, Vol. 48, No. 7

FIGURE 5. P100 amplitude as a function of log contrast, subject MGL041104. At 0.4 cpd (A), the amplitude saturated at high contrast (E). At 2 to 10 cpd (C–F), a double-slope function was used to describe the amplitude response along the contrast levels: one limb to high contrasts (E) and another limb to medium and low contrasts (F). Dashed line: the regression used to estimate the contrast threshold. In the presence of nonlinearity, regressions were fitted on the data obtained at medium to low contrasts.

tion of cells located in the foveal and parafoveal retinal regions along the range of spatial frequencies tested.1 According to this study, at low spatial frequencies (1.5–3 cpd) and especially at low contrasts, the VEP arises from the parafoveal region. In the present study, we found double-slope functions in a similar spatial frequency range (2– 4 cpd), but this finding was absent at lower spatial frequencies, 0.4 to 0.8 cpd, where we found a saturation at high contrast instead. The double-slope, amplitude-contrast function was described as a prominent and general feature of the monkey steady state VEP (4.4 –54.6 Hz of temporal frequency; 0.23–19

FIGURE 6. Comparison between the electrophysiological and dynamic psychophysical CSF (A; F and 䡺, respectively) and between the electrophysiological and static psychophysical CSF (B; F and ‚, respectively). Symbols: mean contrast sensitivities (n ⫽ 6); vertical lines: SD. There was a higher positive correlation for (A) when compared with (B).

cpd).6 Also, the double-slope amplitude-contrast function was found in the first and second harmonics when steady state VEPs were obtained with a 6 cpd grating onset– offset modulated at 7.5 Hz.12 These results were explained by the differential contribution of two visual mechanisms to the VEP at different contrast levels.6,12 We found these nonlinearities in all subjects tested. Figure 7 shows the normalized mean contrast–response functions from our sample (n ⫽ 6) along the spatial frequency and contrast range that we used. It shows some of the information displayed in the Figure 5, but this time the values represent the normalized mean amplitude for the six subjects tested, instead of the absolute amplitudes as in Figure 5, to stress the VEP behavior when the visual system is stimulated at different spatial frequencies.14 At low spatial frequencies, 0.4 and 0.8 cpd, P100 amplitude initially shows a slow decrease when contrast is lowered and then decreases more sharply in the low-contrast range. At spatial frequencies between 2 and 10 cpd, P100 amplitude declines quickly at high contrast and then decreases more modestly at low contrast. However, it should be appreciated that the decrease rate at low contrast is very similar for all the spatial frequencies between 0.4 to 8 cpd, with the possible exception of 10 cpd, which shows a higher rate of amplitude decrease than the others. The different behavior of P100 amplitude in high and low contrasts, which was observed for all spatial frequencies tested, justifies the use of only the amplitude values obtained at low contrast to estimate contrast threshold. Single-unit recordings performed along the visual pathway detected at least two populations of cells mediating high- and low-contrast spatial mechanisms.33–36 Valberg and Rudvin10 and Rudvin et al.11 suggested that VEP to high luminance contrasts comprises the contribution of the M and P pathways, whereas only the M pathway responds to low luminance con-

IOVS, July 2007, Vol. 48, No. 7

Contrast Sensitivity Measured with Transient VEP

3403

FIGURE 7. Mean relative P100 amplitude as a function of contrast and spatial frequency for the six subjects tested. The shapes of contrast-response functions at 0.4 and 0.8 cpd were similar, but different from those at 2 to 10 cpd. At 0.4 and 0.8 cpd there was a tendency of saturation at high contrasts, whereas at 2 to 10 cpd the double-slope response was more predominant.

trasts. The more recent understanding that other visual pathways, which have been loosely grouped in the koniocellular domain,37 provide input to the primate primary visual cortex, may complicate the interpretation of VEP waveform at high stimulus contrast. Our results reinforce those of previous studies showing that the transient VEP is an objective and noninvasive tool for the study of human CSF. We suggest that gratings with medium to low luminance contrasts (⬍25% of Michelson contrast) can be used to find contrast thresholds along the spatial frequency domain, avoiding the effect of high-contrast nonlinearity.

References 1. Campbell FW, Maffei L. Electrophysiological evidence for the existence of orientation and size detectors in the human visual system. J Physiol (Lond). 1970;207:635– 652. 2. Kulikowski JJ. Visual evoked potentials as a measure of visibility. In: Desmedt JE, ed. Visual Evoked Potentials in Man. New Developments. Oxford: Clarendon Press; 1977;168 –183. 3. Spekreijse H, Este´vez O, Reis D. Visual evoked potentials and the physiological analysis of visual processes in man. In: Desmedt JE, ed. Visual Evoked Potentials in Man: New Developments. Oxford: Clarendon Press; 1977;16 – 89. 4. Jones R, Keck MJ. Visual evoked response as a function of grating spatial frequency. Invest Ophthalmol Vis Sci. 1978;17:652– 659. 5. Tyler CW, Apkarian P, Nakayama K. Multiple spatial-frequency tuning of electrical responses from human visual cortex. Exp Brain Res. 1978;15:535–550. 6. Nakayama K, Mackeben M. Steady state visual evoked potentials in the alert primate. Vision Res. 1982;22:1261–1271. 7. Katsumi O, Tanino T, Hirose T. Measurement of contrast sensitivity function using pattern-reversal visual evoked responses. Graefes Arch Clin Exp Ophthalmol. 1985;223:190 –195. 8. Bonds AB, Casagrande VA, Norton TT, Debruyn EJ. Visual resolution and sensitivity in a nocturnal primate (galago) measured with visual evoked potentials. Vision Res. 1987;27:845– 857. 9. Strasburger H, Scheidler W, Rentschler I. Amplitude and phase characteristics of the steady-state visual evoked potential. Appl Opt. 1988;27:1069 –1088. 10. Valberg A, Rudvin I. Possible contributions of magnocellular- and parvocellular-pathway cells to transient VEPs. Vis Neurosci. 1997; 14:1–11.

11. Rudvin I, Valberg A, Kilavik BE. Visual evoked potentials and magnocellular and parvocellular segregation. Vis Neurosci. 2000; 17:579 –590. 12. Bobak P, Bodis-Wollner I, Harnois C, Thornton J. VEPs in humans reveal high and low spatial contrast mechanisms. Invest Ophthalmol Vis Sci. 1984;25:980 –983. 13. Odom JV, Bach M, Barber C, et al. Visual evoked potentials standard. Doc Ophthalmol. 2004;108:115–123. 14. Plant GT, Zimmern RL, Durden K. Transient visually evoked potentials to the pattern reversal and onset of sinusoidal gratings. Electroencephalogr Clin Neurophysiol. 1983;56:147–158. 15. Sannita WG, Lopez L, Piras C, Di Bon G. Scalp-recorded oscillatory potentials evoked by transient pattern-reversal visual stimulation in man. Electroencephalogr Clin Neurophysiol. 1995;96:206 –218. 16. Tzelepi A, Bezerianos T, Bodis-Wollner I. Functional properties of subbands of oscillatory brain waves to pattern visual stimulation in man. Clin Neurophysiol. 2000;111:259 –269. 17. Harris L, Atkinson J, Braddick O. Visual contrast sensitivity of a 6 month old infant measured by the evoked potential. Nature. 1976; 264:570 –571. 18. Strasburger H, Remky A, Murray I, Hadjizenonos C, Rentschler I. Objective measurement of contrast sensitivity and visual acuity with the steady-state visual evoked potential. Ger J Ophthalmol. 1996;5:42–52. 19. Cannon MW Jr. Contrast sensitivity: psychophysical and evoked potential methods compared. Vision Res. 1983a;23:87–95. 20. Cannon MW Jr. Evoked potential contrast sensitivity in the parafovea: spatial organization. Vision Res. 1983;23:1441–1449. 21. Tyler CW, Apkarian P, Levi DM, Nakayama K. Rapid assessment of visual function: an electronic sweep technique for the pattern visual evoked potential. Invest Ophthalmol Vis Sci. 1979;18:703– 713. 22. Seiple WH, Kupersmith MJ, Nelson JI, Carr RE. The assessment of evoked potential contrast thresholds using real-time retrieval. Invest Ophthalmol Vis Sci. 1984;25:627– 631. 23. Orban GA, Orban LC. Rapid contrast sweeping of steady-state visual evoked potentials: a new method for investigation of the visual system. Doc Ophthalmol. 1985;61:7–16. 24. Norcia AM, Tyler CW, Hamer RD, Wesemann W. Measurement of spatial contrast sensitivity with the swept contrast VEP. Vision Res. 1989;29:627– 637. 25. Chen S, Wu L, Wu D. Objective measurement of contrast sensitivity using the steady-state visual evoked potential. Doc Ophthalmol. 1990;75:145–153.

3404

Souza et al.

26. Lopes de Faria JM, Katsumi O, Arai M, Hirose T. Objective measurement of contrast sensitivity function using contrast sweep visual evoked responses. Br J Ophthalmol. 1998;82:168 –173. 27. Silveira LCL, Picanc¸o-Diniz CW, Oswaldo-Cruz E. Contrast sensitivity function and visual acuity of opossum. Vision Res. 1982;22: 1371–1377. 28. Silveira LCL, Heywood CA, Cowey A. Contrast sensitivity and visual acuity of the pigmented rat determined electrophysiologically. Vision Res. 1987;27:1719 –1731. 29. Heywood CA, Silveira LCL, Cowey A. Contrast sensitivity in rats with increased or decreased numbers of retinal ganglion cells. Exp Brain Res. 1988;70:513–526. 30. Barlow HB. Single units and sensation: a neuron doctrine for perceptual psychology? Perception. 1972;1:371–394. 31. Sannita WG, Conforto S, Lopez L, Narici L. Synchronized 15.0 – 35.0 Hz oscillatory response to spatially modulated visual patterns in man. Neuroscience. 1999;89:619 – 623.

IOVS, July 2007, Vol. 48, No. 7 32. Sannita WG, Bandini F, Beelke M, et al. Time dynamics of stimulusand event-related gamma band activity: contrast-VEPs and the visual P300 in man. Clin Neurophysiol. 2001;112:2241–2249. 33. Kaplan E, Shapley RM. X and Y cells in the lateral geniculate nucleus of macaque monkeys. J Physiol (Lond). 1982;330:125– 143. 34. Kaplan E, Shapley RM. The primate retina contains two types of ganglion cells, with high and low contrast sensitivity. Proc Natl Acad Sci USA. 1986;83:2755–2757. 35. Lee BB, Silveira LCL, Yamada ES, et al. Visual responses of ganglion cells of a New World primate, the capuchin monkey, Cebus apella. J Physiol (Lond). 2000;528.3:573–590. 36. Kremers J, Silveira LC, Kilavik BE. Influence of contrast on the responses of marmoset lateral geniculate cells to drifting gratings. J Neurophysiol. 2001;85:235–246. 37. Hendry SHC, Reid RC. The koniocellular pathway in primate vision. Ann Rev Neurosci. 2000;23:127–153.