Short-term changes in the response characteristics of the human visual evoked potential

0042-6989(94)E0063-Q

Pergamon

Vision Res. Vol. 34, No. 21, pp. 2823-2831, 1994

Elsevier Science Ltd. Printed in Great Britain

Short-term Changes in the Response Characteristics of the Human Visual Evoked Potential NEAL S. PEACHEY,*t PAUL J. DEMARCO JR,*~" RODRIGO UBILLUZ,*t WENDY YEE*t Received 29 March 1993; in revised form 13 December 1993

The present study examined how the response characteristics of the visual evoked potential (VEP) varied during the course of trials using a sinusoidai grating stimulus that reversed contrast in a square-wave manner. To accomplish this, amplitude and phase values were derived in short segments during the course of continuous stimulation for three subjects. When stimulus spatial frequencies of 0.77 or 1.55 c/deg were used, VEP amplitude remained at a stable value throughout the trial. At 3.1 c/deg, 6-12 sec were required for VEP amplitude to increase to a stable value, which was on average 204% greater than the value noted during the first few seconds of the trial. At 6.2 and 12.4c]deg, VEP amplitude changes were more complex, first increasing and then decreasing substantially, to levels that were on average 63.8% and 38% of the peak reached earlier in the trial. In all cases, VEP phase decreased during the trial. The magnitude of this decrease ranged up to 50 deg, corresponding to an approx. 10.5 msec delay for the 6.65 Hz stimulation rate used. Prior exposure to an adapting grating diminished the changes in VEP amplitude and advanced the phase changes. Therefore, these changes appear to represent a form of contrast adaptation that is restricted to responses to high spatial frequencies. In addition, the present results provide evidence against a fundamental assumption of signal averaging--that an invariant stimulus will evoke an invariant response.

Spatial frequency Temporal frequency Visual evoked potential Amplitude Phase

INTRODUCTION The response characteristics of the visual evoked potential (VEP) are dependent upon stimulus attributes including spatial frequency (e.g. Armington, Corwin & Marsetta, 1971) and contrast (e.g. Campbell & Maffei, 1970). Our understanding of this dependence is based primarily upon studies in which averaged VEPs were obtained in a series of separate trials, across which a given stimulus parameter was varied. An implicit assumption in this type of study is that the response characteristics of the VEP are relatively stable during the course of stimulus presentation. However, there is evidence that this assumption may not hold. For example, Ho and Berkley (1988) reported that VEP amplitude underwent large changes during a 45 sec trial. Specifically, amplitude increased to a maximum early in the trial and then fell gradually to a low level as the trial continued. A similar pattern of response change has been noted in single unit recordings from primary visual cortex of cat (Maffei, Fiorentini & Bisti, 1973; Vautin &

Berkley, 1977; Albrecht, Farrar & Hamilton, 1984) and monkey (Sclar, Lennie & DePriest, 1989). While Ho and Berkeley (1988) focussed on response amplitude, the changes that may occur in VEP response phase have received relatively little attention. However, a delay is seen before a stable VEP phase value is achieved following a step increase in contrast (Seiple & Holopigian, 1989; Xin, Seiple, Holopigian & Kupersmith, 1994), indicating that phase may also change during the course of a trial. In the present study we compared VEP amplitude and phase during the course of stimulus presentation. We observed changes in both VEP amplitude and phase that were dependent upon spatial frequency. These changes were influenced by prior exposure to a grating of similar spatial frequency, suggesting that they represent a form of contrast adaptation. MATERIALS AND METHODS Subjects

VEP recordings were obtained from three male sub*I51-E, Veterans AffairsHospital, Hines, IL60141, U.S.A. tDepartment of Neurology, Stritch School of Medicine, Loyola jects, aged 30-36 yr. Each subject had a corrected Snellen visual acuity of 20/20 or better, normal color vision, University Chicago, Maywood,IL60153, U.S.A. 2823 VR 34/21-- C

2824

N E A L S. P E A C H E Y et al.

was normal on ophthalmic examination, and wore their optical correction during testing.

Recording and stimulation Recordings were made with Ag/AgC1 electrodes attached with collodion to the scalp along the midline; electrode impedance was 2000 fL The active electrode was placed 2.5 cm above the inion; a reference electrode was placed at Fz and a ground was placed halfway between the active and reference leads. The input was fed to a Grass preamplifier (Model P511 R) with a frequency bandpass of 1-300 Hz and a gain of 50,000. The signal was then digitized at 357 Hz with 12-bit resolution by a Data Translation card (DT2821) synchronized to the graphics display. Stimuli were generated by a digital display generator (Venus, Neuroscientific Corp.) and presented on a Princeton Graphic System monitor (Model Ultrasync). The mean luminance of the display was 29.4cd/m ~ (EG&G Electro-optics Model 550-1 Photometer). The monitor was calibrated to produce linear steps in luminance contrast. At the 2 m viewing distance, the rectangular monitor subtended 4.9 × 5.3deg of visual angle. Test stimuli consisted of vertical sinusoidal gratings that reversed contrast in a square-wave fashion; the Michelson contrast was 70%. A small spot in the center of the display served as a fixation guide. The test grating was presented immediately after exposure to an adapting field, which was either a uniform field or an adapting grating. Adapting gratings had the same spatial and temporal frequencies as the test grating, but were of lower contrast (30%). Adapting fields were presented over the entire screen; test gratings were presented within a 4.4 deg diameter circular aperture located at the center of this field. There was no difference in mean luminance between the adapting and test gratings. Subjects viewed the monitor binocularly through natural pupils. In pilot sessions, we replicated the classic finding that binocular viewing evokes larger amplitude responses than does monocular viewing (e.g. Harter, Seiple & Salmon, 1973), thereby conferring an improved signal-to-noise ratio. These sessions also showed that the pattern of results obtained was similar for binocular and monocular trials. Procedure Figure 1 presents a schematic diagram of the time-course by which stimuli were presented during each trial. Trials were divided into 13 segments. Each segment contained exactly 15 stimulus cycles (i.e. 30 contrast reversals). Therefore, at a frequency of 6.65 Hz, each segment had a duration of 2.26 sec. During the first four segments of a trial ("initial adapt" in Fig. 1), either a uniform adapting field or an adapting grating was presented. During the remaining nine segments, a contrast-reversing test grating was presented continuously. Within an experimental session, test gratings of different spatial frequency were presented in alternate trials. A 90 sec intertrial interval was used (cf. Ho & Berkley, 1988).

Using the Neuroscientific VENUS software, a discrete Fourier transform derived the amplitude and phase values for the second harmonic of the reversal rate (i.e. at 13.3 Hz) within each of the 13 segments that comprised a trial. The phase and amplitude values obtained from each trial were stored separately. Phase values were averaged using a method which takes into account the circular distribution of phase space (Zar, 1974). RESULTS Figures 2-6 present results obtained when the test grating was presented following adaptation to a uniform field using a 6.65 Hz alternation rate. Each subject is represented by a different symbol. Each plotted point represents the VEP measure for one of the 2.26 sec segments diagrammed in Fig. 1, and is the average response measure obtained across five repetitions for a given subject. Figure 2 presents VEP amplitude [Fig. 2(A)] and phase [Fig. 2(B)] results obtained when the test grating had a spatial frequency of 0.77 c/deg. The first four amplitude values, obtained when the uniform field was present, provide an index of the noise at the second harmonic for each subject. Following the presentation of the test grating, VEP amplitude increased to a level above the noise and remained there for the duration of the trial. As expected, VEP phase was unstable during presentation of the uniform field; across stimulus conditions and subjects the average standard deviation was 82 deg. Therefore, for these points, error bars have been omitted for clarity. In comparison, VEP phase was quite reproducible when the test grating was presented (error bars indicating I S E M were often smaller than the plotted points). VEP phase gradually decreased during the course of a trial. At 1.55 c/deg (Fig. 3) the results obtained were similar to those obtained at 0.77 c/deg. Following presentation of the test grating, VEP

Initial

Test grating

Adapt

20.34 sec

9.04 sec

/

\

2.26 sec F I O U R E 1. Schematic diagram of time ~ourse of stimulus presentation and analysis, At 6.65 Hz, each trial was divided into 13 segments of 2.26 sec duration. The first four segments comprised an adaptation period during which either a uniform field or a contrast reversin B vertical sinusoidal adaptin B grating was presented. The test grating for a given trial was presented continuously during the last nine segments. The VEP response recorded during each segment was analyzed separately.

RESPONSE 4-

Z

A

CHARACTERISTICS

OF VEP

A

0.77 epd

T

~.

3-

-

2-

.

/,,~

/A__T

-

E

<:

:

:

:

:

:

:

:

:

,

:

0

-5

[3

180

:

:

:

:

:

:

:

5

:

:

10

:

:

:

:

I

15

:

:

:

:

Z~

cn O '(3

90 o

o

o

o~.~_o.~O_..o,

o~z___o_~ '

• z w j. : O: : : :O" : : , : : : : ', : : : : : : : : :.t : : : : ,

-10

-5

0

5

10

15

20

Time (seconds)

ampltude increased above the noise level and stabilized whereas VEP phase showed a gradual decline from the value obtained following onset of the test grating. As a measure of amplitude stability, we compared the responses obtained during the first and last segments of a

A 3.

"O :D

2-

1 . 5 5 cpd

T

~0

T

T

T

CL

E

<

8~Me

0 , : : : : ~ : : : : , : : : : ~ : : : : ~ : : : : ~ : : : : , -10 -5 0 5 10 15 20 180 B (21 -U

•

90-

•

~ 03 0 cn

A

O•

O O

0 O

-9O -180

r

l

i/' 9

~

B°

o o ° o-o_o

i

20

-o-o-;-;-;

1 : : : : : : : : : : : : : : : : : : : : : : -10 -5 0 5 10

Time

0o O t-

' -go-.

n

A •

-180

•

. . . . . .

-10

/

~'~-

-5

,~

6 .

• .

.

0

• x-".

A : - - - , A

5

z~

z~...,.,..z~_~

. . . . . . . . . . . . .

10

15

20

Time (seconds)

FIGURE 2. VEP amplitude (A) and phase (B) obtained during the course o f stimulus trials using a 0 . 7 7 c/deg test grating. For Figs 2 ~ , the first four points were obtained in response to a u n i f o r m adapting field; subsequent points represent the second h a r m o n i c response to contrast reversal o f the test grating. N o t e that amplitude remained stable during the trial while phase decreased. Each subject is represented by a different symbol. Each point represents the average o f five trials for each subject. Error bars indicate -I-1 S E M and are omitted w h e n smaller than the plotted point. Error bars for the first four phase points were obtained w h e n amplitude was at a noise level and are omitted for clarity.

Z

.

0-

*

-180,

180 90-

0 0

-9o

ZX

3.1 pd : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : -10 -5 0 5 10 15

,

20

...

zx

0 03 0 cEl_

1

< 0,' : -10

cn (D

1

CL

1

E

$~T---.o ~ °

;,

6

6. 4~- ,

Q..

r.,.-A

/0~0

O/-0"-"0~/0"-0~

-0 -1

2825

i : : : : : : , 15

20

(seconds)

FIGURE 3. V E P amplitude (A) and phase (B) obtained during the course o f stimulus trials using a 1.55 c/deg test grating. C o n v e n t i o n s used are the same as in F i g . 2. N o t e that amplitude remained stable during the trial while phase decreased.

FIGURE 4. V E P amplitude (A) and phase (B) obtained during the course o f stimulus trials using a 3.1 c/deg test grating. C o n v e n t i o n s used are the same as in F i g . 2. N o t e that amplitude increased during the trial while phase decreased.

trial. Across subjects, amplitudes measured during the first segment averaged 98.8% of those measured during the last segment, indicating the stability of response amplitudes to these low spatial frequencies. Figure 4 presents the results obtained when the spatial frequency of the test grating was 3.1 c/deg. The first amplitude measure obtained following presentation of the test grating was relatively close to the noise level. During the course of the trial, VEP amplitude only gradually increased to a maximum level that averaged 204% of the amplitude of the initial measure. The time required to reach this maximum ranged from 6 to 12 sec for the three subjects. This slow increase was a reproducible feature at 3.1 c/deg and persisted when a larger (8.8 deg) stimulus field was used (not shown). In addition, VEP phase showed a gradual decrease that persisted after VEP amplitude appeared to stabilize. Figures 5 and 6 present the results obtained when the two highest spatial frequencies were used (6.2 and 12.4 c/deg, respectively). In both cases, VEP amplitude reached a maximum soon after test grating presentation and then decreased during the remainder of the trial. The amplitude measured at the end of the trial averaged 63.8% of the peak value at 6.2c/deg and 38% of the peak value at 12.4 c/deg. VEP phase decreased throughout the course of the trial [Fig. 5(B)] or until VEP amplitude had fallen near or to the noise level [Fig. 6(B)]. In additional sessions, we examined how the amplitude changes described above for the three higher spatial frequencies changed with stimulus temporal frequency. The results obtained showed that the pattern of amplitude change noted at 6.65 Hz were similar when temporal frequencies of 5, 8.5 and 10 Hz were used (data not shown).

2826

NEAL

S. P E A C H E Y

et al. 0.0

0

4-

A -10~3 Q) E~ E o

--

"0 EL

1

E <

e

6.2°pd

::::::::::::::::::::::::::::::: -10 -5 0 5 10

15

20

z'x

O

~Z~__Zk__~__~.•....../~1.~/~ u

~

90

0 0 cEL

0

0

0

-90

-5

-10

A--,i

5

10

1,5

°

20

25

.

• 5.0

ZX--ZX 6.20 cpd -30-

,*--e.

5 u~3

3.10 cpd 12.4 cpd

C) D

-7.5 -40-

-50-

A..~. A -

lO.O 3 c~ o

Time

0 v

-20-

:::::::::::::::::::::::::::::::::::: -60 -lO -5 0 5

180

B

0 0) o3 o cO_

V - - ~ " 0.77 cpd 1.55 cpd

~7--V

10

15

20

(seconds)

FIGURE 7. Change in phase values during the course of trials using a uniform field during the first four segments. Points represent the change in phase from the first value measured following the presentation of the test grating• Each symbol represents a different test grating spatial frequency•Each point representsthe averagechange for the three subjects; error bars are omitted for clarity•

Time (seconds) FIGURE 5. VEP amplitude (A) and phase (B) obtained during the course of stimulus trials using a 6.2 c/deg test grating. Conventions used are the same as in Fig. 2. Note that amplitude first increased and then decreased during the trial while phase decreased.

In Figs 2~5, VEP phase decreased from the initial level following the presentation o f the test grating. However, the magnitude of this decrease was fairly small. To summarize the phase data, Fig. 7 expresses each value as the difference between the first measure obtained following presentation of the test grating and subsequent measures (values obtained during the presentation of the uniform field have been omitted from this analysis). Each data point in Fig. 7 represents these phase differences averaged across the three subjects. In all cases,

4

0 z]

A

T t

•

?J. \ N! ,\

2

•

-°---o-. o

CL

E <

i -5

--10 180

0 {,9 0 cG_

124cpd 5

10

15

20

A

B 03 © q~

s

90 ~0~0~0~0~--0--~

0

0

~0 ~---i~ant~A~o~9/1 \o~. i

0 O 0

-90

Z~

--180 :::::::::::::::::::::::::::::: -I0 -5 5 10

15

20

Time (seconds) FIGURE 6. VEP amplitude (A) and phase (B) obtained during the course of stimulus trials using a 12.4 c/deg test grating. Conventions used are the same as in Fig. 2. Note that amplitude first increased and then decreased

d u r i n g t h e trial w h i l e p h a s e d e c r e a s e d .

phase declined from the initial value. The largest phase decline occurred for the 3.1 c/deg test grating; phase changes were smaller for lower and higher spatial frequencies• At the second harmonic of the alternation rate (i.e. at 13.3 Hz), 1 msec is equivalent to approx. 4.8 deg, a conversion used to generate the righthand scale axis. The largest phase change observed corresponds to a peak latency delay of about 10.5 msec. The analysis applied above provides information about the VEP second harmonic. While this allows the characteristics of VEP harmonics to be rapidly assessed, it ignores the actual VEP waveform. Therefore, averaged VEPs were constructed off-line from the EEG signals obtained during a total of 10 trials. Figure 8 presents representative segments of this average for one subject for two test spatial frequencies: 3.1 c/deg [Fig. 8(A)] and 12.4 c/deg [Fig. 8(B)]. In each panel, the lower tracing represents the rate of stimulus contrast reversal. The upper tracing represents the VEP obtained at the onset of the test grating. This record contains two types of response, a transient response to the onset of the pattern stimulus (arrows) and a steady state response that gradually becomes established. In both panels, the amplitude of the steady state response increases from the first to the second tracing. At 3.1 c/deg, the amplitude of the response noted after 10.5 or 18 sec of stimulation is larger than that seen earlier. In comparison, VEP amplitude decreases to a low level at 12.4 c/deg. These peakto-peak amplitude changes are consistent with the changes in the second harmonic noted for these spatial frequencies [Figs 4(A) and 6(A)]. It is also possible to see the changes in VEP timing in these waveforms. The phase delay is easiest to see at 3.1 c/deg [Fig. 8(A)] where VEP amplitude remains at a higher level throughout the trial. Overall, there is good agreement between this qualitative waveform analysis and the analysis provided above for these two spatial frequencies (Figs 4 and 6). The above results indicate that large changes in VEP amplitude may be observed during the presentation of a test grating. Responses to stimuli with spatial frequencies of 0.77 and 1.55 c/deg did not change during the

RESPONSE CHARACTERISTICS OF VEP

course of a trial, whereas response amplitudes to the higher spatial frequencies changed considerably. Ho and Berkley (1988) suggested that the decrease in VEP amplitude may reflect a form of contrast adaptation. If so, then it should be possible to alter the time course of these changes by presenting adapting stimuli prior to the presentation of a test grating. Figures 9-11 present sessions designed to test this possibility. Specifically, test gratings were presented following one of three adaptation procedures. In addition to a baseline condition in which a uniform field was presented, two other adapting trials were run. In one, an adapting grating was presented during all four adapting segments. In the other, the adapting grating was presented during only

4-

2827

A

~

3.1

cpd

.f {

O) "U -1

{÷{T

E

0

<

-10

I~

0

0

-5

5

10

15

5

10

15

20

11) -t3

03

t(2-

-90. . . . .

A

-10

3.1 cpd

-5

0

20

Time (seconds)

0 2.5 "ID

10.5

e~

E

18 i

i

i

~

i

i

Time (200 msec/div)

B

12.4 cpd

~.

E

.¢

2.5

~

18

i

i

i

i

i

i

i

Time (200 msee/div) F I G U R E 8. Averaged VEP waveforms obtained from one subject within trials in which the test grating was either 3.1 c/deg (A) or 12.4c/deg (B). Each waveform represents the average of 10 trials identical to those used to obtain the data shown in Figs 4 and 6. The four response epochs illustrated were obtained at different time points within the trial. The delay (sec) between the initial presentation of the test grating and the starting point of each segment is indicated by the numbers at the right. The lower trace represents the time-course of contrast reversal.

F I G U R E 9. Effect of prior adaptation on the manner in which VEP amplitude (A) and phase (B) values change during the course of stimulus trials using a 3.1 c/deg grating. In Figs 9-11, open circles represent trials in which a uniform field was presented during the first four segments. Diamonds represent trials in which a low contrast 3.1 c/deg adapting grating was presented during each of the first four segments (solid diamonds) or only during the third and fourth adapting segments, with a uniform field presented during the first and second segments (open diamonds). Each point represents the average of the results obtained from three subjects; error bars indicate 4- 1 SEM.

the third and fourth adapting segment (a uniform field being presented during the first and second segments). Figure 9 presents average results obtained across the three subjects when a 3.1 c/deg test grating was used. Here, the different symbols represent the three adaptation conditions. The baseline condition (open circles) used no adapting grating and replicates the pattern of results seen previously (Fig. 4). During the presentation of the adapting grating (diamonds), a relatively small VEP was recorded as noted in the left panel in each figure. When an adapting grating was used, VEP amplitude did not increase when the test grating appeared. In addition, VEP amplitudes obtained later in the trial did not appear to differ across the three adaptation conditions. A two-way repeated measures analysis of variance (ANOVA) was used to compare the amplitude measures obtained during the presentation of the test grating for the three adaptation conditions. There was no significant difference in VEP amplitude across the three adaptation conditions (F2.52=0.60; P >0.05). However, VEP amplitude changed significantly during the trial (Fs.52 = 2.72; P < 0.05). A significant interaction between the adaptation condition and time shows that the manner in which VEP amplitude changed during the trial differed across the three adaptation conditions (F,6.52 = 2.64; P < 0.01). VEP phase was also affected by adaptation. The phase advance noted for the baseline condition was increased by prior exposure to an adapting grating. An ANOVA applied to the phase data revealed that VEP phases were significantly different

2828

NEAL

across the three adaptation conditions (F2.52= 16.88; P <0.01), that VEP phase changed during the trial (F8,52 = 11.11; P < 0.01) and that the interaction between adaptation condition and time was significant (Fj6,52 = 6.35; P < 0.01). Figure 10 presents the results obtained when a 6.2 c/deg test grating was used. VEPs were again obtained to the adapting grating itself. This is shown by the large VEP amplitudes and relatively stable phase values plotted at the lefthand portion of the figure panels (diamonds). The amplitude level initially reached in the baseline condition (open circles; see also Fig. 5) was not achieved in trials where the test grating followed exposure to an adapting grating (open and solid triangles). An A N O V A revealed that VEP amplitudes were significantly different across the three adaptation conditions (F2,52 = 10.92; P < 0.01), that VEP amplitude changed significantly during the trial (F8,52= 8.58; P < 0.01) and that the interaction between adaptation condition and time was significant (F,6,52 = 5.72; P <0.01). For this stimulus condition, VEP phase did not appear to change across adaptation conditions nor during the trial. This impression was supported by the ANOVA, which showed that VEP phases were not significantly different across the three adaptation conditions (F2,52= 0.46; P > 0.05), that VEP phase did not change significantly during the trial (F8,52= 1.85; P >0.05) and that the interaction between adaptation condition and time was not significant (F16,52 = 1 . 5 4 , P > 0.05). Figure 11 presents the results obtained when a 12.4 c/deg test grating was used. During the baseline condition (open circles), there was an amplitude decrease comparable to that seen previously [Fig. 6(A)]. The amplitude levels observed in adapting trials (diamonds) never achieved the level of the baseline condition. An

4-

A

"

l

et al.

S. PEACHEY

4.

~-. > zzl.

A 3.

-0

2-

12.4 epd

O•/+XT --0

0-.. T

.

t T\* °%-oT ,

0

-10

, , , , I ,-,

-5

i

0

i

,

i

I

i

i

t

*

5

I

10

:

:

:

:

', :

:

15

:

:

,

20

180-

3 ~ © X3 Q) u) 0 cn

90-

• O.

-90

-

-180

-10

::

: : ; : : : : i : : :

-5

: : : : : : : : : : : ~ : : : : .

5

10

15

20

Time (seconds)

FIGURE 11. Effect of prior adaptation on the manner in which VEP amplitude (A) and phase (B) values change during the course of stimulus trials using a 12.4c/deg grating. Conventions used are the same as in Fig. 9. A N O V A revealed that VEP amplitudes were different across the three adaptation conditions (F2,52= 32.10; P < 0.01), that VEP amplitude changed during the trial (F8,52 = 10.61; P < 0.01) and that the interaction between adaptation condition and time was significant ( F 1 6 , 5 2 = 6.43; P < 0.01). Phase values appeared to advance to a stable value more quickly under both adaptation conditions. This observation was supported by an ANOVA, which showed that VEP phases were significantly different across the three adaptation conditions (F2,52 = 5.96; P < 0.01), that VEP phase changed during the trial (Fs,52 = 6.89; P < 0.01) and that the interaction between adaptation condition and time was significant ( F I 6 . 5 2 = 5.51; P < 0.01).

6.2 cpd

DISCUSSION 7

E <

o

@

o

o

0

-10

5

-5

180 B !

15

20

{+÷-H+{+{

i

C~ (1.) -0 {19 O3 0 CO-

10

O.

-90.

f

o -180

oT

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

-10

-5

0

5

10

15

20

Time (seconds)

FIGURE 10. Effect of prior adaptation on the manner in which VEP amplitude (A) and phase (B) values change during the course of stimulus trials using a 6.2 c/deg grating. Conventions used are the same as in Fig. 9.

The present study has examined the short-term changes in the response characteristics of the human VEP that occur during the presentation of a high contrast stimulus. Because the magnitude of these changes can be rather large, the results obtained challenge an underlying assumption of signal averaging-that an invariant stimulus will evoke an invariant response. Responses to stimuli of low spatial frequency changed little during the course of a trial. This result agrees with that reported by Xin et al. (1994), who examined VEP changes following a contrast increment for a 1.0 c/deg test grating. In comparison, reproducible changes in VEP amplitude were seen for the middle and high spatial frequency stimuli used. The results obtained for high spatial frequency stimuli are similar to those reported by Ho and Berkley (1988). However, H o and Berkley (1988) reported that a decline in VEP amplitude was always observed during the course of stimulus presentation;

RESPONSE CHARACTERISTICSOF VEP spatial frequency changed only the time-course of this decline. It is not clear what factor is responsible for the difference between the present results and those of Ho and Berkley (1988) at low and middle spatial frequencies. However, our results indicate that temporal frequency, field size and monocular versus binocular viewing are unlikely candidates. Several studies have been carried out at the single cell level using a paradigm similar to that used here (Maffei et al., 1973; Vautin & Berkley, 1977; Albrecht et al., 1984; Sclar et al., 1989). In the majority of cells, response amplitude increased initially before declining during the presentation of a test grating. This pattern of change resembles the present results obtained with 6.2 and 12.4c/deg test gratings (Figs 5 and 6). However, the responses of single cells may also increase during the course of presentation of the test grating, as seen when a 3.1 c/deg test grating was used (Fig. 4). In cat, Albrecht et al. (1984) noted that 28% of all cells studied showed a response increase during stimulus presentation. This suggests that the spatial frequency dependence of the present VEP results may reflect the response characteristics of different subpopulations of cortical cells. However, Albrecht et al. (1984) reported there was no correlation between the spatial frequency dependence of a cell and the manner in which the response of that cell changed during stimulus presentation. In monkey, Sclar et al. (1989) found that some complex cells showed a gradual increase in response amplitude during stimulation while simple cells did not. Although it is possible that the VEP changes noted here are related to a changing contribution of simple and complex cell subpopulations, our data cannot address this issue. Exposure to a high contrast grating reduces visual sensitivity to subsequently-presented stimuli of similar spatial frequency and orientation (e.g. Gilinsky, 1968; Blakemore & Campbell, 1969). Under similar conditions, a reduction in VEP amplitude is typically observed (e.g. Blakemore & Campbell, 1969; Campbell & Maffei, 1970; Mecacci & Spinelli, 1976; Nelson, Seiple, Kupersmith & Carr, 1984a; Odom & Norcia, 1984; Brigell, Peachey & Seiple, 1987; Suter, Armstrong, Suter & Powers, 1991), although an amplitude increase may be seen (Bach, Greenlee & Buhler, 1988; Suter et al., 1991). The time-course of this effect has been studied psychophysically (e.g. Blakemore & Campbell, 1969) and using VEP techniques (e.g. Campbell & Maffei, 1970; Nelson et al., 1984a; Ho & Berkley, 1988). Ho and Berkley (1988) noted that the time-course of the decrease in VEP amplitude was similar to that measured psychophysically (e.g. Blakemore & Campbell, 1969), and suggested that the decrease in VEP amplitude was a direct correlate of visual adaptation. The results shown in Figs 9-1 l support the idea that the changes observed represent a form of contrast adaptation. If this is the case, however, the present results indicate that an amplitude increase may result from contrast adaptation at medium spatial frequencies (cf. Bach et al., 1988). In addition, the amplitude stability noted at low spatial frequencies indicates another distinction between the present VEP

2829

paradigm and contrast adaptation measured psychophysically, which does not show a similar spatial frequency dependence (cf. Graham, 1972; Jones & Tulunay-Keesey, 1975; De Valois, 1977). In sum, VEP amplitudes and psychophysically measured contrast thresholds may not behave identically in adaptation paradigms. Across the spatial frequency range tested, there were three patterns by which VEP amplitude changed during a trial. At low spatial frequencies, the VEP appears to represent the activity of a mechanism with a response that changes little in amplitude during stimulation. At high spatial frequencies, the VEP represents a different mechanism, the response of which decreases during stimulation. In between, VEP amplitude increased during a trial [Fig. 4(A)]. Strasburger, Murray and Remky (1993) recently provided evidence that the responses of low and high spatial frequency-sensitive mechanisms may cancel near 3 c/deg, presumably due to a near 180 deg phase difference. They interpret these two mechanisms to be analogous to the transient and sustained mechanisms identified psychophysically (e.g. Tolhurst, 1973; Kulikowski & Tolhurst, 1973). Strasburger et al. (1993) examined responses obtained only during the first three sec of stimulation. Therefore, the low amplitude VEP recorded to a 3.1 c/deg grating early during a trial [Fig. 4(A)] may reflect the same response cancellation proposed by Strasburger et al. (1993). If the magnitude of one of these mechanisms becomes smaller during the course of a trial, as is the case for high spatial frequencies [Figs 5(A) and 6(A)], an amplitude increase is predicted and is observed. Therefore, the amplitude increase noted at 3.1 c/deg may reflect a near-complete cancellation at the beginning of the trial, followed by a gradual revealing of the response of a stable low spatial frequency mechanism due to a decrease in amplitude of the canceling response of a high spatial frequency mechanism. In addition to the amplitude changes noted during stimulus presentation, VEP phase decreased for all spatial frequencies tested, which is equivalent to a prolongation of VEP latency. The magnitude of this decrease ranged up to 50 deg at 6.65 Hz, corresponding to a latency change of about 10.5 msec. Bearing in mind that this decrease occurred within a 20 sec trial, this is a surprisingly large magnitude change when compared to published values of normative VEP latency and phase (e.g. Tomoda, Celesia & Toleikis, 1991). As a result, a quantitative analysis of phase values obtained from sweep trials may be difficult to interpret. However, most sweep studies use phase only qualitatively, to indicate a reproducible amplitude signal (Norcia & Tyler, 1985) or to determine which channel of a lock-in amplifier to use (Nelson, Seiple, Kupersmith & Carr, 1984b). Nevertheless, sweep trials have been used to obtain phasespatial frequency functions for quantitative comparison to patient data (Celesia, Brigell, Gunnink & Dang, 1992). The possibility that VEP phase may change during continuous stimulation finds support in a study by

NEAL S. PEACHEY et al.

2830

Seiple and Holopigian (1989). The mechanism underlying the VEP phase changes observed, and the dependence of these changes upon spatial frequency, is not known. However, the VEP appears to be composed of several underlying components (Jeffreys & Axford 1972a, b; Zemon, Kaplan & Ratliff, 1980; Arakawa, Peachey, Celesia & Rubboli, 1993). This raises the possibility that the observed phase changes reflect alterations in the time courses of these components and, as a consequence, the manner in which these components interact to form the VEP. Further studies using animal models in which the components of the VEP have been differentially affected will allow an evaluation of this hypothesis. Finally, the present results bear on the interpretation of amplitude data obtained in sweep trials. For example, the relatively slow amplitude increase for 3.1 c/deg stimuli predicts the presence of an amplitude trough during a spatial frequency sweep. In fact, amplitude troughs in spatial frequency sweeps have often been reported near 3c/deg (Tyler, Apkarian & Nakayama, 1978; Tyler, Apkarian, Levi & Nakayama, 1979; Apkarian, Nakayama & Tyler, 1981; Strasburger & Rentschler, 1986; Strasburger, Scheidler & Rentschler, 1988; Strasburger et al., 1993). In addition, to obtain the highest estimate of acuity under the present conditions, an approx. 2 sec delay prior to data collection should be used for each spatial frequency. Although this delay will underestimate response amplitudes at intermediate spatial frequencies, this will allow optimal responses to be obtained at high spatial frequencies which contribute more to the acuity estimate. Indeed, a 1 sec delay has been incorporated by Strasburger et al. at each stimulus step in spatial frequency sweeps (Strasburger & Rentschler, 1986; Strasburger et al., 1988, 1993). A parametric study will be necessary to determine the extent to which different delays will affect swept acuity estimates.

REFERENCES

Albrecht, D. G., Farrar, S. B. & Hamilton, D. B. (1984). Spatial contrast adaptation characteristics of neurones recorded in the cat's visual cortex. Journal of Physiology, London, 347, 713 739. Apkarian, P. A., Nakayama, K. & Tyler, C. W. (1981). Binocularity in the human visual evoked potential: Facilitation, summation and suppression. Eleetroeneephalography and Clinical Neurophysiology, 51, 32-48. Arakawa, K., Peachey, N. S., Celesia, G. G. & Rubboli, G. (1993). Component-specific effects of physostigmine on the cat visual evoked potential. Experimental Brain Research, 95, 271-276. Armington, J. C., Corwin, T. R. & Marsetta, R. (1971). Simultaneously recorded retinal and cortical responses to patterned stimuli. Journal of the Optical Soeiety of America, 61, 1514-1521. Bach, M., Greenlee, M. W. & B/ihler, B. (1988). Contrast adaptation can increase visually evoked potential amplitude. Clinical Vision Sciences, 3, 185 194. Blakemore, C. & Campbell, F. W. (1969). On the existence of neurones in the human visual system selectively sensitive to the orientation and size of retinal images. Journal of Physiology, London, 203, 237 260.

Brigell, M. G., Peachey, N. S. & Seiple, W. H. (1987). Pattern electroretinogram threshold does not show contrast adaptation. Investigative Ophthalmology and Visual Science, 28, 1614-1616. Campbell, F. W. & Maffei, L. (1970). Electrophysiological evidence for the existence of orientation and size detectors in the human visual system. Journal of Physiology, London, 207, 635~52. Celesia, G. G., Brigell, M., Gunnink, R. & Dang, H. (1992). Spatial frequency evoked visuograms in multiple sclerosis. Neurology, 42, 1067 1070. De Valois, K. K. (1977). Spatial frequency adaptation can enhance contrast sensitivity. Vision Research, 17, 1057-1065. Gilinsky, A. S. (1968). Orientation-specific effects of pattern of adapting light on visual acuity. Journal of the Optical Soeiety of America, 58, 13-18. Graham, N. (1972). Spatial frequency channels in the human visual system: Effects of luminance and pattern drift rate. Vision Research, 12, 53~8. Harter, M. R., Seiple, W. H. & Salmon, L. (1973). Binocular summation of visually evoked responses to pattern stimuli in humans. Vision Research, 13, 1433 1446. Ho, W. A. & Berkley, M. A. (1988). Evoked potential estimates of the time course of adaptation and recovery to counterphase gratings. Vision Research, 28, 1287 1296. Jeffreys, D. A. & Axford, J. G. (1972a). Source locations of patternspecific components of human visual evoked potentials. I. Component of striate cortical origin. Experimental Brain Research, 16, 1 21. Jeffreys, D. A. & Axford, J. G. (1972b). Source locations of patternspecific components of human visual evoked potentials. II. Component of extrastriate cortical origin. Experimental Brain Research, 16, 22-4O. Jones, R. M. & Tulunay-Keesey, U. (1975). Local retinal adaptation and spatial frequency channels. Vision Research, 15, 1239- 1244. Kulikowski, J. J. & Tolhurst, D. J. (1973). Psychophysical evidence for sustained and transient detectors in human vision. Journal o/ Physiology, London, 232, 149-162. Maffei, L., Fiorentini, A. & Bisti, S. (1973). Neural correlate of perceptual adaptation to gratings. Science, 182, 1036~1038. Mecacci, L. & Spinelli, D. (1976). The effects of spatial frequency adaptation on human evoked potentials. Vision Research, 16, 477-479. Nelson, J. I., Seiple, W. H., Kupersmith, M. J. & Carr, R. E. (1984a). A rapid evoked potential index of cortical adaptation. Electroencephalograph)' and Clinical Neurophysiology, 59, 454-464. Nelson, J. I., Seiple, W. H., Kupersmith, M. J. & Carr, R. E. (1984b). Lock-in techniques for the swept stimulus evoked potential. Journal ¢~f Clinical Neurophysiology, 1, 409-436. Norcia, A. M. & Tyler, C. W. (1985). Spatial frequency sweep VEP: Visual acuity during the first year of life. Vision Research, 25, 1399 1408. Odom, J. V. & Norcia, A. M.(1984). Retinal and cortical potentials: spatial and temporal characteristics. Documenta Ophthalmologica Proceedings Series, 40, 29 38. Sclar, G., Lennie, P. & DePriest, D. D. (1989). Contrast adaptation in striate cortex of macaque. Vision Research, 29, 747-755. Seiple, W. H. & Holopigian K. (1989). An examination of VEP response phase. Eleetroencephalography and Clinical Neurophysiology, 73, 520-531. Strasburger, H. & Rentschler, I. (1986). A digital fast sweep technique for studying steady-state visual evoked potentials. Journal of Electrophysiological Techniques, 13, 265-278. Strasburger, H., Murray, I. J. & Remky, A. (1993). Sustained and transient mechanisms in the steady-state visual evoked potential: Onset presentation compared to pattern reversal. Clinical Vision Sviences, 8, 211-234. Strasburger, H., Scheidler, W. & Rentschler, I. (1988). Amplitude and phase characteristics of the steady-state visual evoked potential. Applied Optics, 27, 1069. 1088. Surer, S., Armstrong, C. A., Surer, P. S. & Powers, J. C. ( 1991 ). Spatial-frequency-tuned attenuation and enhancement of the

RESPONSE CHARACTERISTICS OF VEP steady-state VEP by grating adaptation. Vision Research, 31, 1167-1175.

Tolhurst, D. J. (1973). Separate channels for the analysis of the shape and the movement of a moving visual stimulus. Journal of Physiology, London, 231, 385-402. Tomoda, H., Celesia, G. G. & Toleikis, S. C. (1991). Effect of spatial frequency on simultaneous recorded steady-state pattern electroretinograms and visual evoked potentials. Electroencephalography and Clinical Neurophysiology, 80, 81-88. Tyler, C. W., Apkarian, P. & Nakayama, K. (1978). Multiple spatialfrequency tuning of electrical responses from human visual cortex. Experimental Brain Research, 33, 535 550. Tyler, C. W., Apkarian, P., Levi, D. M. & Nakayama, K. (1979). Rapid assessment of visual function: an electronic sweep technique for the pattern visual evoked potential. Investigative Ophthalmology and Visual Science, 18, 703 713. Vautin, R. G. & Berkley, M. A. (1977). Responses of single cells in cat visual cortex to prolonged stimulus movement: Neural

2831

correlates of visual aftereffects. Journal of Neurophysiology, 40, 1051-1065. Xin, D., Seiple, W., Holopigian, K. & Kupersmith, M. J. (1994). Visual evoked potentials following abrupt contrast changes. Vision Research, 34, 2813-2821. Zar, J. H. (1974). Cicular distributions. In Zar, J. H. (Ed.), Biostatistical analysis (pp. 310-328). Englewood Cliffs, N.J.: Prentice-Hall. Zemon, V., Kaplan, E. & Ratliff, F. (1980). Bicuculline enhances a negative component and diminishes a positive component of the visual evoked cortical potential in the cat. Proceedings of the National Academy of Sciences, U.S.A., 77, 747(~7478.

Acknowledgements--We are grateful to Drs Marc Bearse, Mitch Brigell, Gastone Celesia, Elmar Schmeisser, Bill Seiple and Neil Winchester for comments on the manuscript. Supported by grants from the U.S. Department of Veterans Affairs.