The temporal frequency response function of pattern ERG and VEP: changes in optic neuritis

ELSEVIER

Electroencephalography and clinical Neurophysiology 100 (1996) 428-435

o

The temporal frequency response function of pattern ERG and VEP: changes in optic neuritis Benedetto Falsini a'*, Vittorio Porciatti b aEye Clinic, Catholic University, Lgo F. Vito 1, 00168, Rome, Italy blnstitute of Neurophysiology, CNR, Pisa, Italy

Accepted for publication: 8 May 1996

Abstract Steady-state pattern electroretinograms (PERGs) and visual evoked potentials (VEPs) in response to sinusoidal gratings (2 c/deg), sinusoidally counterphased at closely spaced temporal frequencies (TFs) between 4 and 28 Hz, were recorded from 11 patients with unilateral optic neuritis (ON; 11 affected eyes and 10 healthy fellow eyes) and 7 age-matched normal subjects (7 eyes). Amplitude and phase of responses' second harmonics were measured. Responses' apparent latencies were estimated from the rate at which phase lagged with TF. When compared to control values, mean PERG and VEP amplitudes of ON eyes were reduced (by about 0.4 log units) at both low (5-10 Hz) and high (16-20 Hz) TFs. Mean PERG amplitudes of fellow eyes were selectively reduced at low TFs (by about 0.3 log units). Mean PERG apparent latencies of both ON and fellow eyes were delayed (by 15 and 9 ms, respectively). Mean VEP apparent latency of ON eyes was delayed at both low and high TFs (by 24 and 30 ms, respectively), while that of fellow eyes was selectively delayed at high TFs (by 28 ms). The results in ON eyes indicate non-selective abnormalities of PERG and VEP generators responding at both low and high TFs. VEP TF losses may be in part accounted for by corresponding PERG losses. In the fellow eyes of patients, more selective PERG and VEP TF abnormalities may suggest differential impairment of retinal and postretinal subsystems responding better to low and high TFs (i.e. parvo-and magnocellular streams). Keywords: Pattern electroretinogram; Visual evoked potentials; Optic neuritis; Temporal frequency; Steady-state analysis

1. Introduction A number of psychophysical and electrophysiological studies have demonstrated that optic neuritis (ON) affects the temporal properties of visual system (Plant and Hess, 1985; Foster, 1986; Hess and Plant, 1986). Delayed visual perception, determined either psychophysically by a twoflashlight paradigm (Regan et al., 1976), or electrophysiologically by evoked potentials (Halliday et al., 1972), has been reported following an attack of ON (see Hess and Plant, 1986, for a review). Psychophysical temporal sensitivity losses have also been described in ON patients (Hess and Plant, 1986). In some patients, these losses have been reported to involve specific bands of temporal frequencies (TF), the responses to low TFs ( 1 - 1 0 Hz) being predomi* Corresponding author. Tel.: +39 6 30155201; fax: +39 6 3051274; e-mail: [email protected]

nantly affected (Hess and Plant, 1986; Wall, 1990), while in others non-selective abnormalities in both the low and high TF ranges have been found (Honan et al., 1990). Selective TF losses suggest that some visual neural pathways may be more vulnerable than others, or may deteriorate at a different rate during the course of disease. Temporal losses can occur at any stage between the retina and the visual cortex. Retinal and cortical evoked potentials (EPs) may provide an objective estimate of the temporal properties of the visual system. The electroretinogram obtained in response to contrast reversing gratings (pattern electroretinogram, PERG) is known to be correlated with retinal ganglion cell activity (Maffei and Fiorentini, 1982; Maffei and Fiorentini, 1986), and may reflect the temporal responsiveness of both outer and inner retina. Pattern evoked cortical potentials (pattern VEPs) represent an index of both retinal and postretinal functions, and may give an indication of how

0168-5597/96/$15.00 © 1996 Elsevier Science Ireland Ltd. All rights reserved Pll S0921-884X(96)95695-7

EEP 95695

429

B. Falsini. V. Porciatti / Electroencephalography and clinical Neuraphysiology 100 (1996) 428-435

the temporal information is encoded at cortical level. PERGs and VEPs in response to sinusoidally-modulated stimuli (steady-state responses) (Regan, 1966; Strasburger, 1987), when recorded as a function of TF, can therefore be used to determine the suprathreshold response functions at retinal and postretinal levels. Changes in responses' phases as a function of TF can also provide a measure of the system's apparent delay (Spekreijse et al., 1977; Porciatti et al., 1992). The above-mentioned steady-state EPs technique has not been of widespread clinical use in the past. This is at least in part related to the technical limitations of most of the commercial visual stimulators, which do not allow adequate sampling at medium-to-high temporal rates due to the slow refresh rate (Simon, 1992). In addition, these stimulators are not provided with a suitable linearization, that is necessary to avoid contamination by luminance flicker. The available studies (e.g. Regan, 1977; Hess and Baker, 1984; Plant et al., 1986; Porciatti et al., 1992; Simon, 1992) indicate that, in normal subjects, the steadystate PERG and VEPs are temporally band-pass tuned in amplitude, with a maximum at low temporal frequencies ( 6 - 1 0 Hz), a secondary maximum at high temporal frequencies ( 1 6 - 2 0 Hz) and a high frequency cut-off at 3 0 40 Hz. The phases of both PERG and VEPs lag linearly with temporal frequency with slopes corresponding to the apparent latencies of responses (PERG, about 60 ms; VEPs, about 100 ms). Both PERG and VEPs therefore display in normals complex frequency-response functions that probably reflect the contribution of different temporal channels with their own response functions. The present study was designed to evaluate the PERG and VEP frequency-response functions in patients with unilateral ON. PERGs and VEPs showed in the affected eyes non-selective amplitude losses at both low and high TFs. Increased phase lags as a function of TF were also observed for both responses. The unaffected fellow eyes of patients showed more selective abnormalities, involving

PERG and VEP amplitudes at low TFs, and VEP phase at high frequencies.

2. Methods Eleven patients (6 males and 5 females), aged between 21 and 59 years, with a clinical history of one or more attacks of ON in one eye, were included in the study. Standard ophthalmologic examination including anterior segment biomicroscopy, visual acuity testing, Goldmann and Humphrey perimetry, applanation tonometry and ophthalmoscopy (direct and indirect) was performed in all cases. Transient VEPs to high-contrast reversing checkerboards (Celesia et al., 1986) were also recorded in all cases. Patients were selected on the basis of having normal clinical results in their fellow eyes. This allowed us to compare the PERG and VEP frequency functions of either the ON-affected or the fellow asymptomatic eyes to those of the normal control eyes. No concomitant ocular diseases were clinically evident. Refractive errors, when present, did not exceed +3 sph, _+1 cyl and were fully corrected for electrophysiological recordings. Clinical data of patients are summarized in Table 1. All cases were unilateral. In two patients (2 and 10) the diagnosis of ON may appear to be questionable since their clinical findings could not be very common for ON. However, in these patients the diagnosis was well supported by both clinical history and laboratory test results. Patient 2 was diagnosed as having a papillitis in her left eye with 3 relapses over the last 3 years. In the acute phase of disease, a pathological VEP delay and a dense paracentral scotoma were documented in the affected eye. With resolution of the acute episodes, VEP latency recovered to normal values, although the amplitude remained lower than that of the fellow eye, and perimetric sensitivity showed an improvement. Optic disk had a slight oedematous appearance throughout the course of disease. MRI and C A T scan ruled out compressive lesions, and no evidence of vascular systemic disor-

Table l Clinical findings of patients Patient Optic neuritis n/age (years)/ sex RE LE

Duration Visual acuity of symptoms (years) RE LE

Transient VEPs latency (ms)

Visual field

Optic disk

RE

LE

RE

LE

RE

LE

1/22/F 2/52/F 3/43/M 4/55/M 5/31IF 6/21/F 7/22/F 8/24/M 9/40/M 10/59/M 11/30/M

3 1 3 2 3 2 1 2 13 1 1

104 95 129 112 116 134 nt 98 108 I04 114

126 99 92 111 96 110 128 92 120 104 126

n n c ps rcs n nt ps n n n

bse ps n n n n rcs n rcs ps, rcs rcs

n n p p n n nt n n n n

n oed n n n n n n n p n

No No Yes Yes Yes Yes No Yes No No No

Yes Yes No No No No Yes No Yes Yes Yes

1.0 1.0 0.5 0.9 1.0 1.0 nt 0.9 1.0 1.0 1.0

1.0 1.0 1.0 1.0 1.0 1.0 0.8 1.0 0.5 0.5 0.9

nt, not tested; bse, blind spot enlargement; ps, paracentral scotoma; rcs, relative central scotoma; n, normal; p, temporal pallor; oed, optic disk edema.

430

B. Falsini, V. Porciatti / Electroencephalography and clinical Neurophysiology 100 (1996) 428-435

ders (i.e. systemic hypertension, carotid occlusive disease) was found at clinical examination. Patient 10 had recurrent episodes of acute ON in his left eye (3 episodes over 2 years). An increased VEP latency with reduced amplitude was documented in the acute phase. VEP latency recovered to near normal values 6 months after the last ON attack, while amplitude remained lower than that of the fellow eye. Visual field examination showed a relative central scotoma (i.e. 5 - 7 dB reduction in the foveal and parafoveal sensitivities) and a paracentral scotoma. This patient was a non-smoker subject with no evidence of vascular systemic diseases. Compressive lesions where ruled out by CAT scan and MRI. In patient 7, the unaffected fellow eye was amblyopic, and therefore was not included in the study. In all, there were 11 ON-affected eyes and 10 unaffected fellow eyes. None of the patients had clinically definite multiple sclerosis according to the criteria proposed by Poser et al. (1983). The criteria for a diagnosis of ON included a history of sudden visual loss associated with pain in the affected eye, followed by a period of visual recovery. All cases fulfilled the criteria of Fleishman et al. (1987) for ON, and had resolved ON at the time of examination. The time elapsed from the onset of the most recent attack of ON to the testing of the study ranged 12-156 months. All patients had received systemic cortisone treatment during the acute phase of disease. At the time of electrophysiological testing, ocular motility was normal in all patients. Mean decimal visual acuity of affected eyes was 0.8 (range 0.5-1.0). The latency of transient VEPs to 1 Hz reversing checkerboards, or the intereye latency difference, was abnormal in 8 out of 11 patients. Most ON eyes (including those with a normal latency) also had subnormal amplitudes of transient VEPs. None had evidence of inflammatory or vascular retinal involvement. Seven eyes of 7 normal subjects, whose age distribution was comparable with that of the patients (mean 36 years, SD 10 years), were also tested. Informed consent was obtained from all subjects or patients after the nature of the technique and the aim of the research were fully explained. The visual stimulus consisted of vertical sinusoidal gratings of fixed contrast (90%) and spatial frequency (2 c/ deg), sinusoidally reversed in contrast at several closely spaced TFs between 4 and 28 Hz. The gratings were generated on a Joyce (Cambridge) monitor operated at 233 frames/s. The monitor was masked to a circular field of dia 21 cm, surrounded by a large, white dome (91 cm diameter) of matched mean luminance (200 cd/m2). Subjects fixated monocularly at the center of the stimulus from a viewing distance of 70 cm. All the eyes included in the study were tested at all the reported temporal frequencies. Pupils were natural, and their size did not differ between patients (both affected and unaffected eyes) and controls after adaptation to the stimulus mean luminance. PERGs were recorded with an Ag-AgC! electrode taped

on the skin over the lower eyelid. A similar electrode, placed over the eyelid of the contralateral patched eye, was used as reference (interocular PERG; Fiorentini et al., 1981). As the recording protocol was extensive, the use of skin electrodes with an interocular recording represented a good compromise between signal-to-noise ratio and signal stability. Discussion on the interocular PERG by skin electrodes and its relationship with the PERG by corneal electrodes can be found elsewhere (Porciatti, 1987; Hawlina and Konec, 1992; Porciatti and Falsini, 1993). VEPs were recorded with Ag/AgC1 electrodes placed 2 cm above the inion (active) and at the vertex. The common ground was on the forehead. PERG and VEP signals were amplified (PERG 100000-fold, VEP 50000-fold), band-pass filtered between 1 and 100 Hz (6 dB/oct) and averaged (12 bit resolution, 256 Hz sampling rate, 200-400 repetitions in 5 - 1 0 blocks each). The averaging time (i.e. the sweep duration) was varied according to the stimulus period. Single sweeps exceeding a threshold voltage (4 V) were rejected, to minimize noise coming from blinks or eye movements. A discrete Fourier analysis was performed off-line in order to isolate the PERG and VEP second harmonic components, whose half-to-peak amplitudes (in /xV) and phases (in 7r radians) were estimated. Examples of steady state PERG and VEPs to sinusoidal gratings modulated in counterphase at different TFs have been shown previously (Porciatti et al., 1992; Padovano et al., 1995). For the whole range of TF tested, the PERG and VEPs had a sinusoidal-like waveform, with a period corresponding to the second harmonic of the stimulus frequency. Thus, the second harmonic response component well represents the main response power. Averaging and Fourier analysis were also performed on signals sampled asynchronously at 1.1 times the TF of the stimulus, to give an estimate of the background noise. Under the present experimental conditions, the PERG and VEP responses recorded individually from both control subjects and patients were well above the noise level (see Section 3) and sufficiently reliable (i.e. the coefficient of variation in amplitude was typically 25%). In order to analyze the PERG and VEP phases, the raw data were first plotted as a function of TF on a linear scale. The phase lagged smoothly with TF. However, as the range of linear phase lag is bound between +l~r radians, the raw phase data had obvious discontinuities at the transition between +lTr radians, and could not be directly fitted by a linear regression line. Multiple integers of 27r radians were then subtracted to subsets of raw data when necessary to sort them over the whole TF range (as described in Strasburger, 1987; Porciatti et al., 1992). Following the linear system analysis approach first applied to evoked potentials by Regan (1966), the temporal phase of steady-state responses to a number of closely spaced TFs can provide an estimate of the system's apparent latency. Indeed, if the relationship between TF and response phase is linear, the system can be treated as a linear filter incor-

B. Falsini, V. Porciatti / Electroencephalography and clinical Neurophysiology 100 (1996) 428-435

porating a pure delay. This delay can then be estimated from the slope of the best-fitting straight line. Responses' apparent latencies were calculated from the slopes (determined by least-squares fit of the data; Porciatti et al., 1992) of the phase lag as a function of TF, according to the formula: apparent latency (ms) = phase slope (Tr radians/ Hz) x 1000/4 (period of the second harmonic: 2 x 27r radians). PERG and VEPs amplitude data from control and ONaffected eyes were statistically compared by analysis of variance (ANOVA), after logarithmic transformation of PERG and VEP amplitudes to better approximate a normal distribution. Repeated measures ANOVAs with one 'within' subjects and one 'between' subjects (grouping) factor were performed with amplitudes as dependent variables. Two sets of TF data were separately evaluated: one including the peak at 8 Hz ('low' TF range 4 - 1 2 Hz), and the other including the remaining frequencies up to 28 Hz ('high' TF range 13-28 Hz). TF was treated as the 'within' subjects factor and group (normal versus affected eyes) as

PERG

-0.2-

o ,_1 V

LU

the 'between' subjects factor. The Greenhouse-Geisser correction (Greenhouse and Geisser, 1959) was employed to adjust the probabilities for the repeated measures analyses. PERG and VEP apparent latencies from control and ON eyes were statistically compared by independent t tests. The results from unaffected fellow eyes of patients were separately analyzed by the same procedures. Since it was assumed that these eyes could not have been fully normal (Fleishman et al., 1987), their results were compared with those obtained from control eyes. 3. R e s u l t s

Fig. 1A,B shows the mean amplitudes of PERG and VEPs as a function of TF for control eyes, ON eyes and fellow eyes. In the same plots, the dotted lines represent the asynchronous noise levels averaged across subjects. As previously described (Porciatti et al., 1992), the PERG frequency-response of control eyes displayed a maximum at 8 Hz, a secondary peak at 16 Hz with a dip in between,

(A)

-0.6-

£3

-o.8-

_.~

-1.0-

"'''"""'../:"'> ........ ~'~'~-~

-o.8

.... .......... ..........."",, ..,.

n <

- - ' - - CONTROLS - - v - - FELLOW EYES - - o - - OPTIC NEURITIS EYE5

VEP

0,4 =

=

-0.41

431

-1.2 -

-1.4

0 0-

-1.2,

"'"'.% ........... "........... ...-."

;

1'0

1~

2'0

2~

3b

(c)

-2-

-1.6 0-

""--.. ....................

o

g

1'0

is

2b

2;

s'0

-2-

t'-

-4-

-4"0

-6-. -8-

iii < "I"

-lO-

-10-

t,'t

-12 -

-12-

-14 -

-u-

-16

0

;

lb

l's

2b

2'5

3'o

-16

~....~

o

~

1'0

1;

2'0

2's

3'o

T E M P O R A L F R E Q U E N C Y (Hz) Fig. 1. (A,B) Mean ( + SEM) PERG (A) and VEP (B) amplitude values plotted as a function of temporal frequency for controls, fellow eyes and optic neuritis eyes. The dotted lines represent the asynchronous noise levels averaged across subjects. (C,D) Mean ( + SEM) PERG (C) and VEP (D) phase values plotted as a function of temporal frequency for controls, fellow eyes and optic neuritis eyes. In (C) and (D) phases of fellow eyes and optic neuritis eyes have been shifted on the y-axis by 1 and 27r radians, respectively, for clarity of presentation.

B. Falsini, V. Porciatti / Electroencephalography and clinical Neurophysiology 100 (1996) 428-435

432

and a cut-off beyond 30 Hz. The VEP frequency-response showed a maximum between 5 and 8 Hz, a shoulder at 16 Hz, and a cut-off beyond 30 Hz. The results of the statisTable 2 Results of statistical analysis (ANOVA) on PERG and VEP amplitudes recorded from controls and patients Between factor

Within factor

Interaction

1,16

8,128

8,128

1.99 0.65 0.05 5,80

0.64 0.65 0.74 5,80

4.1 0.72 0.002 8,120

1.15 0.72 0.33 8,120

2.35 0.61 0.02 5,75

0.66 0.61 0.72 5,75

0.4 0.74 0.84

PERG amplitude Normals vs ON eyes (4-12 Hz)

Normals vs ON eyes ( 13-28 Hz)

Normals vs fellow eyes (4-12 Hz)

Normals vs fellow eyes (13-28 Hz)

df

F G-G P df

F G-G P df

15.8 0.01 1,16

4.84 0.05 1,15

F G-G P df

5.25

F G-G P

1.13 0.30

8 0.74 0.00

df

1,16

8,128

8,128

F G-G P df

4.9

1.78 0.33 0.08 5,80

0.73 0.33 0.66 5,80

F G-G P df

5.8

26 0.58 0.00 8,120

0.75 0.58 0.59 8,120

F G-G P df

0.99

0.9 0.53 0.52 5,75

1.35 0.53 0.22 5,75

F G-G P

0.036

31.1 0.57 0.00

0.2 0.57 0.96

0.04 1,15

VEP ampli-

tude Normals vs ON eyes (4-12 Hz)

Normals vs ON eyes (13-28 Hz)

Normals vs fellow eyes (4-12 Hz)

Normals vs fellow eyes (13-28 Hz)

0.049 1,16

0.02 1,15

0.33 1,15

0.85

G-G, epsilon factor value of the Greenhouse-Geisser correction (see also Section 2).

tical comparisons between control and patients' eyes, performed by ANOVA, are reported in Table 2. In ON eyes, as compared to controls, both PERG and VEP mean amplitudes were reduced at both low and high TFs, with greatest losses (about 0.4 log units) around the peaks of the normal TF functions. ANOVA showed a significant effect of the group at both low and high TF, but no significant interaction of group by TF. In the fellow eyes of patients, mean PERG and VEP amplitudes were larger than those of ON eyes, but smaller than those of controls. Fellow eyes' PERG and VEP amplitudes tended to be, on average, selectively reduced at low TFs by about 0.3 log units. A N O V A showed a significant effect of the group for PERG, but not VEP amplitude at low TFs. The interaction of group by TF was not statistically significant. Fig. 1C,D shows how the average phase of PERG and VEPs changes as a function of TF for control, fellow and ON eyes. The mean phases of fellow and ON eyes have been shifted along the y-axis by 1 and 2 7r radians, respectively, for clarity of presentation. The PERG phase decreased steadily with TF in both controls and patients and was well fitted by a single linear regression. In both controls and patients, the VEP phase showed a discontinuity at 12 Hz and was best fitted by separate linear regressions for the low (4-11 Hz) and high (12-30 Hz) frequency branches. In both ON and fellow eyes, PERG and VEP phase slopes tended to be steeper than those of controls. Apparent latencies were evaluated from the slopes of individual phase plots and displayed in Fig. 2A,B as histograms. The mean PERG apparent latency was delayed, as compared to that of controls, by 15 ms in ON eyes (t = 3.5; P < 0.01) and 9 ms in the fellow eyes (t -- 2.6; P < 0.05). The mean VEP apparent latency was delayed in ON eyes, as compared to that of controls, by 24 ms at low TFs (t = 2.2, P < 0.05) and 30 ms at high TFs (t = 3.9; P < 0.01). In the fellow eyes, mean VEP apparent latency was selectively delayed at high TFs, as compared to that of controls, by 28 ms (t = 9.9; P < 0.001). 4. Discussion In normal eyes, the VEP amplitude versus TF function showed a bimodal shape, with a maximum at around 6 - 8 Hz and a shoulder between 16 and 20 Hz. The frequency response function of the PERG was also bimodal with two amplitude maxima, either in the low (8 Hz) or high (16 Hz) frequency region. Both PERG and VEP phases lagged progressively with increasing TF. We found that while the PERG phase data were well fitted by a single regression line, those of VEP phase were best fitted by two regression lines, with a change in slope at 12 Hz. This yielded two different estimates of VEP response latency, in the low and high TF regions, respectively. The present findings are in general agreement with those previously reported on normal PERG and VEP TF functions by different studies (Regan, 1966; Hess and Baker, 1984; Plant

B. Falsini, V. Porciatti / Electroencephalography and clinical Neurophysiology 100 (1996) 4 2 8 - 4 3 5

(B)

433

VEP

160,

TFs: 4 - 11 Hz

(A)

PERG

12o

16o.

E 80,

v

E r.,3

>0

120.

zw

<~

I..z w n, ,~ 13_ <

Z

40.

IZW

16o-

W 7-

80.

7-

"-r-

CONTROLS

n, < 13_

40

FELLOW EYES

OPTICNEURITISEYES

I TFs:12-28Hz I

13. 12o

CONTROLS

FELLOW EYES

IPTIC NEURITISEYES

<

80 40-

,

,2 ,, \ ' - ,

,\

,\

, Q,

i<

CONTROLS

FELLOW EYES

OPTICNEURITISEYES

Fig. 2. Histograms showing the mean ( _+ SEM) PERG (A) and VEP (B) latencies, estimated from the corresponding phase responses obtained in controls, fellow eyes and optic neuritis eyes. PERG latency was estimated form the slope of the phase response of each observer, over the entire temporal frequency range (4-28 Hz). For the VEP, the latency was calculated separately for the lower and upper limbs of the temporal frequency-response curves (above and below 12 Hz, respectively) of each observer.

et al., 1986; Porciatti et al., 1992; Simon, 1992), and suggest multiple underlying generators for both retinal and postretinal functions. In particular, at least two generators with different temporal properties could contribute with a different weight to the responses, depending on the temporal frequency: a generator with more sustained/tonic properties dominates the response at low TFs, while another generator with more transient/phasic properties dominates the responses at higher TFs. Neurons with these differential characteristics are present in the visual pathway of primates, the so-called parvocellular (P) and magnocellular (M) streams (Kaplan and Shapley, 1986; Kremers et al., 1992). Parvo- and magno-pathways remain separate at least to the level of primary visual cortex (Shapley, 1990). P cells respond tonically and more slowly than M cells, whereas M cells respond better than P cells to fast stimuli of low contrast (De Monasterio, 1978). The notch at 10-12 Hz, observed in both PERG and VEP amplitude functions, and the change in slope at 12 Hz, found for the VEP phase function, would suggest a kind

of interaction between the two classes of generators. This multi-channel modelling of retinal and postretinal TF functions is also in keeping with the psychophysical results on temporal modulation sensitivity (see Hess and Plant, 1986, for a review), which indicate that different channels, although broadly tuned and with overlapping sensitivity profiles, may contribute to the visual TF function. In the ON-affected eyes, amplitude and phase abnormalities involved both PERG and VEP functions. PERG and VEP amplitude losses with respect to control values were found either in the low or high TF region and were largest at the peak of the normal TF curves. Therefore, affected eyes had shallower frequency functions as compared to controls. PERG and VEP apparent latencies were delayed in ON as compared to normal eyes. VEP latency delays involved both low and high TF regions. Although the PERG and VEP have been already evaluated in optic nerve demyelination (e.g. Plant et al., 1986; Holder, 1991; Bradshaw, 1992), their TF response function has not yet been investigated. The present method, as corn-

434

B. Falsini, V. Porciatti / Electroencephalography and clinical Neurophysiology 100 (1996) 428-435

pared to previous, more routine, electrophysiotogical techniques employed in ON, could offer the advantage of evaluating in greater detail the temporal properties of different functional channels at both retinal and postretinal level. This may help to reveal abnormalities, involving either the amplitude TF function or the apparent latency, which are not readily detectable by other methods. For instance, the increase in the PERG apparent latency found in ON eyes represents a relatively new finding. It is worth noting that, in a recent study evaluating PERG and VEP TF responses to chromatic contrast stimuli, Porciatti and Sartucci (1996) reported an abnormal PERG apparent latency in patients with multiple sclerosis and unilateral ON. We may presume that, like amplitude attenuation, a pathological PERG delay in ON may result from either retrograde ganglion cell dysfunction or primary retinal pathology (i.e. retinal vascular changes; Lightman et al., 1987). The pattern VEPs' TF function was evaluated in 15 ON patients by Abe et al. (1993). They found, as we have, non-selective amplitude losses in the range 4 - 2 4 Hz and suggested a potential diagnostic use of the VEP TF function in ON. Unfortunately, no VEP phase data were reported in their study. The present study provides evidence that both amplitude losses and latency delays of VEPs in ON may be accounted for, at least in part, by corresponding PERG abnormalities, and that both PERG and VEP abnormalities in ON tend to be non-selective as a function of TF. It may be interesting to discuss present electrophysiological results in the light of psychophysical studies evaluating threshold TF sensitivity and suprathreshold TF discrimination in ON. A number of studies have shown that the temporal acuity (i.e. critical fusion frequency) for flicker or pattern stimuli can be reduced in ON (see Hess and Plant, 1986, for a review), indicating an impairment of responses in the high TF spectrum. Low TF losses in sensitivity have also been reported for either flickering (Harding and Wright, 1986; Honan et al., 1990) or patterned stimuli (Hess and Plant, 1983; Hess and Plant, 1986). With sinusoidal gratings, threshold sensitivity losses appeared to be selective for low TFs when stimuli of low spatial frequency (0.5 c/deg) were employed (Hess and Plant, 1983). Travis and Thompson (1989) reported variable spatio-temporal losses in patients with multiple sclerosis who presumably had had previous ON: in some patients sensitivity losses were more marked at low TFs, while in others such losses were revealed more reliably at high TFs. Evaluating suprathreshold TF discrimination of 2 c/deg sinusoidal gratings, Plant and Hess (1985) found that some ON eyes exhibited psychophysical losses which were suggestive of a greater impairment of the lower, more sustained TF channel as compared to the higher, transient one. The general picture that emerges from the results of the psychophysical studies is that different patterns of TF sensitivity losses can be associated with ON. Many patients can indeed exhibit either selective losses in the low and high TF regions, or non-selective abnormalities.

The present electrophysiological data point towards a nonselective nature of loss of the suprathreshold TF sensitivity in ON. Another finding of the present study was that the fellow eyes of our patients, as compared to control eyes, showed changes in the mean PERG and VEP frequency responses, although less marked than those observed in the affected eyes: mean PERG amplitudes tended to be selectively depressed at low TFs, and mean VEP phases were delayed only at high TFs. Abnormalities in the apparently healthy fellow eyes of patients with unilateral ON have been previously reported in numerous studies by using a variety of psychophysical or electrophysiological techniques (Meienberg et al., 1982; Honan et al., 1990; Holder, 1991; Bradshaw, 1992). These abnormalities could reflect subclinical optic nerve demyelination involving both axonal damage and conduction delay. The present data suggest, in addition, that sustained/tonic and transient/phasic mechanisms subserving PERG and VEPs frequency responses may be differently altered in subclinical demyelination. In particular, selective PERG amplitude losses to low TF could reflect a peripheral damage involving mainly P retinal ganglion cells. Indeed, selective losses at low TFs have been found in monkey's contrast sensitivity after selective, acrylamide-induced degeneration of P retinal ganglion cells (Merigan and Eskin, 1986). A selective delay of VEP latency at high TFs could result from an abnormal electric conduction affecting prevalently the magnocellular pathway. In this study, for obvious constraints in collecting data from patients, the PERG and VEP TF functions were recorded at the peak spatial frequency of the stimulus (2 c/deg). It would have been more informative, in order to evaluate the involvement of magno- and parvo-subsystems, to have at least two slices in the temporal domain, at high and low spatial frequency. Clearly, a better separation among generators with parvoand magno-properties can be achieved by selecting stimulus characteristics other than TF, i.e. equiluminant chromatic contrast (Morrone et al., 1994) which is processed primarily through the parvocellular pathway, and low-contrast, low-spatial frequency stimuli which are more specific for the magnocellular pathway. Studies employing such experimental paradigms of stimulation should be able to more fully address the issue of a differential impairment of visual pathways in subclinical optic nerve demyelination. References Abe, H., Hasegawa, S., Takagi, M., Yoshizawa, T. and Usui, T. Temporal modulation transfer function of vision by pattern evoked potentials in patients with optic neuritis. Ophthalmologica, 1993, 207: 9499. Bradshaw, K. Early onset of abnormality of the pattern-evoked ERG in patients with optic neuritis. Clin. Vis. Sci., 1992, 7: 313-326. Celesia, C.G., Kaufman, D. and Cone, S.B. Simultaneous recording of pattern electroretinography and visual evoked potentials in multiple sclerosis. Arch. Neurol., 1986, 43" 1247-1252.

B. Falsini, V. Porciatti / Electroencephalography and clinical Neurophysiology 100 (1996) 428-435 De Monasterio, F.M. Properties of concentrically organized X and Y ganglion cells of macaque retina. J. Neurophysiol., 41, 1978: 13941410. Fiorentini, A., Maffei, L., Pirchio, M., Spinelli, D. and Porciatti, V. The ERG in response to alternating gratings in patients with diseases of peripheral visual pathways. Invest. Ophthalmol. Vis. Sci., 1981, 21: 490-493. Fleishman, J.A., Beck, R.W., Linares, O.A. and Klein, G.W. Deficits in visual function after resolution of optic neuritis. Ophthalmology, 1987, 94: 1029-1035. Foster, D.H. Psychophysical loss in optic neuritis: luminance and colour aspects. In: R.F. Hess and G.T. Plant (Eds.), Optic Neuritis. Cambridge University Press, Cambridge, 1986, pp. 152-191. Greenhouse, S.W. and Geisser, S. On methods in the analysis of profile data. Psychometrika, 1959, 24:95-112. Halliday, A.M., McDonald, W.I. and Mushin, J. Delayed visual evoked response in optic neuritis. Lancet, 1972, i: 982-985. Harding, G.F.A. and Wright, C.E. Visual evoked potentials in acute optic neuritis. In: R.F. Hess and G.T. Plant (Eds.), Optic Neuritis. Cambridge University Press, Cambridge, 1986, pp. 230-254. Hawlina, M. and Konec, B. New noncorneal HK-loop electrode for clinical electroretinography. Doc. Ophthalmol., 1992, 81: 253-259. Hess, R.F. and Baker, J.R. Human pattern evoked electroretinogram. J. Neurophysiol., 1984, 51:939-951. Hess, R.F. and Plant, G.T. The effect of temporal frequency variation on threshold contrast sensitivity deficits in optic neuritis. J. Neurol. Neurosurg. Psychiatry, 1983, 46: 322-330. Hess, R.F. and Plant, G.T. The psychopbysical loss in optic neuritis: spatial and temporal aspects. In: R.F. Hess and G.T. Plant (Eds.), Optic Neuritis. Cambridge University Press, Cambridge, 1986, pp. 1/)9-151. Holder, G.E. The incidence of abnormal pattern electroretinography in optic nerve demyelination. Electroenceph. clin. Neurophysiol., 1991, 78: 18-26. Honan, W.P., Heron, J.R., Foster, D.H., Edgar, G.K., Scase, M.O. and Collins, M.F. Visual loss in multiple sclerosis and its relation to previous optic neuritis, disease duration and clinical classification. Brain, 1990, 113: 975-987. Kaplan, E. and Shapley, R.M. The primate retina contains two types of ganglion cells, with high and low contrast sensitivity. Proc. Natl. Acad. Sci. USA, 1986, 83: 2755-2757. Kremers, J., Lee, B.B. and Kaiser, P.K. Sensitivity of macaque ganglion cells and human observers to combined luminance and chromatic temporal modulation. J. Opt. Soc. Am., 1992, 9: 1477-1485. Lightman, S., McDonald, W.I., Bird, A.C., Francis, D.A., Hoskins, A., Batchelor, J.R. and Halliday, A.M. Retinal venous sheathing in optic neuritis: its significance for the pathogenesis of multiple sclerosis. Brain, 1987:110:405-411. Maffei, L. and Fiorentini, A. Electroretinographic responses before and after section of the optic nerve. Science, 1982, 211: 953-954. Maffei, L. and Fiorentini, A. Generator sources of the pattern ERG in man and animals, in: R.Q. Cracco and I. Bodis-Wollner (Eds.), Frontiers in Clinical Neuroscience, 3. Liss, New York, 1986, pp. 101-106.

435

Meienberg, O., Flammer, J. and Ludin, H.P. Subclinical visual field defects in multiple sclerosis. J. Neurol., 1982, 227: 125-133. Merigan, W.H. and Eskin, T.A. Spatio-temporal vision of macaques with severe loss of P retinal ganglion cells. Vision Res., 1986, 26: t7511761. Morrone, C., Porciatti, V., Fiorentini, A. and Burr, D.C. Pattern-reversal electroretinogram in response to chromatic sfimulL I: humans. Vis. Neurosci., 1994, l l: 861-871. Padovano, S., Falsini, B., Ciavarella, P., Moretti, G. and Porciatti, V. Spatial-temporal interactions in the steady-state electroretinogram. Doc. Ophthalmol., 1995, 90: 169-176. Plant, G.T. and Hess, R.F. Temporal frequency discrimination in optic neuritis and multiple sclerosis. Brain, 1985, 108: 647-676. Plant, G.T., Hess, R.F. and Thomas, J.S. The pattern evoked electroretinogram in optic neuritis: a combined psychophysical and electrophysiological study. Brain, 1986, 109: 469-489. Porciatti, V. Non-linearities in the focal ERG evoked by pattern and uniform-field stimulation. Invest. Ophthalmol. Vis. Sci., 1987, 28: 1306-1313. Porciatti, V. and Falsini, B. Inner retina contribution to the flicker electroretinogram: a comparison with the pattern electroretinogram. Clin. Vision Sci., 1993, 8: 435-447. Porciatti, V. and Sartucci, F. Retinal and cortical evoked responses to chromatic contrast stimuli: specific losses in both eyes of patients with multiple sclerosis and unilateral optic neuritis. Brain, 1996, in press. Porciatti, V., Burr, D.C., Morrone, M.C. and Fiorentini, A. The effects of ageing on the pattern electroretinogram and visual evoked potentials in humans. Vision Res., 1992, 32:1199-1209. Poser, C.P., Paty, D.W., Scheinberg, L., McDonald, i., Davis, F.A., Ebers, G.C., Johnson, K.P., Sibley, W.A., Silberberg, D.H. and Tourtelotte, W.W. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann. Neurol., 1983, 13: 227-231. Regan, D. Some characteristics of averaged steady-state and transient responses evoked by modulated light. Electroenceph. clin. Neurophysiol., 1966, 25: 231-237. Regan, D., Milner, B. and Heron, J.R. Delayed visual perception and delayed visual evoked potentials in the spinal form of multiple sclerosis and retrobulbar neuritis. Brain, 1976, 99: 43-66. Simon, F. The phase of PVEP in maxwellian view: influence of contrast, spatial and temporal fiequency. Vision Res., 1992, 32: 591-599. Shapley, R. Visual sensitivity and parallel retinocortical channels. Ann. Rev. Psychol., 1990, 41: 653-658. Spekreijse, H., Estevez, O. and Reits, D, Visual evoked potentials and the physiological analysis of visual processes in man. In: J.E. Desmedt (Ed.), Visual Evoked Potentials in Man: New Developments. Clarendon Press, Oxford, 1977, pp. 16 89. Strasburger, H. The analysis of steady-state evoked potentials revisited. Clin. Vision Sci., 1987, 1: 245-256. Travis, D. and Thompson, P. Spatiotemporal contrast sensitivity and colour vision in multiple sclerosis. Brain, 1989, 112: 283-303. Wall, M. Loss of P retinal ganglion cell function in resolved optic neuritis. Neurology, 1990, 40: 649-653.