Influence of the spatial periodicity of
moving gratings on motion response
Max]. Keck, Frederick W. Montague, Jr., and Thomas P. Burke
The present experiments examine the suprathreshold response of the motion or directionselective portion of the human visual system by means of the motion aftereffect (MAE). The
MAE was measured as a function of the contrast and spatial frequency of moving sinusoidal
gratings. For spatial frequencies less than 1 cy/deg, the MAE speed was found to increase
linearly with log contrast up to 80%. For spatial frequencies greater than 1 cy/deg, the rate of
increase of the MAE speed with log contrast was not found to be linear over the entire range of
contrast. The nonlinearity was greatest for the 8 and 10 cy/deg gratings, which showed very
little increase in MAE speed with contrast above 25%. We conclude that the direction-specific
mechanisms in human vision show a more limited contrast response to the high spatial frequencies than does the visual system as a whole.
Key words: moving gratings, motion aftereffect, spatial frequency,
contrast sensitivity, suprathreshold
T
he visual process of detecting a moving
grating appears to involve a motion-sensitive
system and a pattern-sensitive system. The
minimum contrast required to detect the
movement of a grating is not the same as the
minimum contrast required to detect its spatial structure. Thus different contrast sensitivity curves are obtained for different detection criteria.1'2 In the present research we
are examining the response of the motion or
direction-specific system.
Direction-specific adaptation. Evidence for
the existence of direction-specific mechanisms comes from experiments by Sekuler
and Ganz3 and Pantle and Sekuler,4 which
demonstrate that moving gratings produce
direction-selective adaptation. In other words,
adapting to a grating moving in a particular
From the Department of Physics, John Carroll University, Cleveland, Ohio.
This research was supported by a Cottrell College Science Grant from Research Corporation.
Submitted for publication July 5, 1979.
Reprint requests: Max J. Keck, Ph.D., Department of
Physics, John Carroll University, Cleveland, Ohio
44118.
1364
direction elevates the threshold more for detecting that grating when it is moving in
the same direction as the adapting grating.
This selective adaptation procedure has been
used to obtain information about the suprathreshold contrast response of directionspecific mechanisms.4' 5 These results show
that direction-specific adaptation for 3.4
cy/deg gratings increases linearly with the
logarithm of the adapting contrast for lowcontrast values. At intermediate contrasts the
amount of adaptation begins to saturate, resulting in very little further increase at higher
contrasts.
Motion aftereffect. Another approach to
measuring the response of direction-specific
mechanisms is to make use of the motion aftereffect (MAE). Barlow and Hill6 found cells
in the retina of the rabbit which respond only
to a stimulus moving in a particular direction.
When a stimulus was moved through the receptive field of the cell in the preferred direction, the response of the cell immediately
increased above its maintained level of discharge. As the stimulus motion continued,
the response decreased but remained above
0146-0404/80/111364+07$00.70/0 © 1980 Assoc. for Res. in Vis. and Ophthal., Inc.
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
Volume 19
Numlier 11
Spatial periodicity of moving gratings 1365
100
1
1
1 1
1 1
1 1
PH
80 -
O,
o
o
O,/
60 —
-
O s"
1
/
/
o
/
4
/
1/
/l
,A-A
—
^-^^
/
D
A _ .
A
/y / y
—
8
/
/
1
A
o
A
B
/
)
—
20 —
0.2
o
A
40 —
-
1
—
/
| |
1 1 1
10
1.0
100
ADAPT CONTRAST
Fig. 1. Initial MAE speed as a function of the contrast of the moving grating for 1 cy/deg (0), 4
cy/deg (A), and 8 cy/deg (El) for Subject P. H. The data for the 1 and the 8 cy/deg gratings were
normalized by the factors given in the caption of Fig. 2.
its normal baseline response (adaptation).
When the stimulus motion was stopped, the
response dropped abruptly to zero and then
slowly recovered to its maintained level of
discharge. The MAE is believed to result
from an imbalance in discharge among cells
sensitive to various directions of motion during the recovery phase.6"8 A stronger MAE
would be due to a greater imbalance, which
in turn implies a greater response of the
direction-specific system during adaptation.
We therefore assume that the magnitude of
the MAE can be used as a measure of the
response of the direction-specific motion
system.
Sekuler and Pantle8 used the MAE to determine the recovery rate and the dependence on stimulus velocity of directionspecific cells. Pantle9 measured the MAE as a
function of the velocity (temporal frequency)
for 3 and 6 cy/deg gratings. He concluded
that the magnitude of the MAE has a broad
maximum for a temporal frequency near 5 Hz
for both 3 and 6 cy/deg gratings. To achieve
this optimal temporal frequency of 5 Hz, the
grating velocity must be changed accordingly
for each spatial frequency.
Keck et al.10 examined the dependence of
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
the MAE on the contrast of 4 cy/deg moving
gratings. Their results showed that for low
contrasts, the MAE magnitude increased
rapidly with grating contrast. However, for
contrasts greater than five to six times
threshold, further increases in the MAE
magnitude with contrast occurred at a greatly
reduced rate. This is in agreement with the
results of Pantle and Sekuler4 and Pantle et
al.5 obtained from direction-selective adaptation experiments.
In the present experiments we measured
the MAE as a function of the contrast and
spatial frequency of moving gratings. At each
spatial frequency the velocity of the adapting
grating was adjusted to result in a temporal
frequency of 5 Hz. The magnitude of the
MAE was determined by two psychophysical
methods: magnitude estimation and tracking.
The gratings subtended a visual angle of 7.6°
X 5.9°.
Experimental method
Grating display. The gratings used for this
study were generated electronically on the face of
a Tektronix 5103 (P31 phosphor) oscilloscope. The
manual intensity control of the oscilloscope was
positioned so that the luminance of the screen was
rs(. Ophthalmol. Vis. Sci.
November 1980
1366 Keck, Montague, and Burke
100
0.2
1.0
10
ADAPT CONTRAST
100
(%)
Fig. 2. Initial MAE speed as a function of the contrast of the moving gratings for Subject P. H.
for the spatial frequencies indicated next to each curve, ranging from 0.34 to 10 cy/deg. The
data for each spatial frequency were normalized by (from low to high) 1.2, 1.3, 1.0, 1.0, 0.84,
0.93, and 0.80. The mean value of the stmdard errors was 4.12.
directly proportional to the DC voltage applied to
the external z-axis input. An amplifier was used to
sum the AC and DC voltage applied to the z-a\is.
Because the luminance of the screen was a linear function of the applied z-;wis voltage, the contrast was the peak amplitude of the AC voltage
divided by the DC voltage (up to 80% contrast).
This was confirmed by optical measurements of
the contrast.
The gratings were made to drift across the
screen by continuously phase-delaying the signal
applied to the trigger relative to the z-axis signal.
This was accomplished by rotating a synchroresolver with a motor. The grating speed was determined by the angular speed of the synchrore solver.
The experiments took place in a room that was
dark except for the light emitted by the oscilloscope screen. For all measurements, the subject
viewed the display binocularly from a chair situated so that the viewing distance was 1 m. A headrest was mounted on the chair to maintain a constant viewing distance. At 1 m the screen subtended a visual angle of 7.6° horizontally and 5.9°
vertically. A small black dot situated in the center
of the display served as a fixation point for the
subjects. The average luminance of the screen was
0.7 millilamberts.
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
Subjects. The subjects were three John Carroll
University students. P. H. and F. M. were experienced psychophysical observers. T. B. was
familiarized with the procedures used and was
given some practice in observing prior to being
used as a subject. All three subjects were myopic
and wore corrective lenses.
Procedure. Before each run, the subject darkadapted for 10 min. He was then presented with a
set of trials. Each trial consisted of the following: a
30 sec presentation of a moving grating at the desired contrast, a 3 sec uniform field, and then a
stationary test grating (same spatial frequency as
the moving grating) at four times threshold contrast (the MAE varies with the test contrast10' ").
For both the adapt and test phase, the subject was
instructed to maintain fixation on the small black
dot located at the center of the display. The MAE
generated with the test grating was measured
either by magnitude estimation or tracking. During a particular run, only one method was used.
When using the method of magnitude estimation, the subject was asked to estimate the backward speed of the MAE within 2 sec after the
stationary grating was turned on. The first two
trials were with a 4 cy/deg grating (14% contrast)
which was referred to as the standard. The subject
was told that the MAE arising from the standard
Volume 19
Number 11
Spatial periodicity of moving gratings 1367
100
0.2
1.0
10
ADAPT CONTRAST
100
(%)
Fig. 3. Initial MAE speed as a function of the contrast of the moving grating for Subject F. M.
for the spatial frequencies indicated next to each curve, ranging from 0.26 to 10 cy/deg. The
data for each spatial frequency were normalized by (from low to high) 0.60, 0.73, 0.85, 1.1, 1.0,
1.0, 0.56, and 0.68. The mean value of the standard errors was 7.27.
had a speed of 50 and that all other MAEs were to
be estimated relative to this standard speed.
When using the tracking method to measure
the MAE, the subject moved a lever at the same
speed as the test grating appeared to be moving.
The lever was the wiper of a linear potentiometer.
Its position was recorded on a strip chart recorder
with a speed of 15 cm/min. The slope of the resulting curve was proportional to the lever speed.
Tangents to the first 2 sec of the recorded curves
were drawn to obtain the initial MAE speed.
The contrast of each moving (adapting) grating
was the independent variable tested for each spatial frequency. The contrasts tested ranged from
0.6% to 80%. The range of spatial frequencies
tested was 0.26 to 10 cy/deg, with only one spatial
frequency utilized per run. The speed of the moving grating of each spatial frequency was such that
it produced a 5 Hz temporal frequency at any
point on the oscilloscope face.
A run consisted of the following: two standards
(defined to have an MAE speed of 50 for magnitude estimation); two familiarization trials which
served as practice for the subject; a set of trials,
one at each contrast tested, presented in random
order; a standard, the speed of which was estimated (or tracked) by the subject; a second complete set of trials, also presented in random order;
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
and a final standard trial. There was a 90 sec delay
between trials during which the subject looked
away from the display. Each run took approximately 1% hr, with the actual time depending on
the number of contrasts presented.
Results
Fig. 1 shows representative results for
Subject P. H. The initial MAE speed resulting from the adaptation to moving gratings is
shown as a function of the contrast of the
moving grating for three different spatial frequencies. Each point is the geometric mean
of four speed estimates obtained during two
separate runs by the method of magnitude
estimation.
The straight lines are obtained by leastsquares fits to the data. The data is divided
into a high-contrast and a low-contrast segment, each with its own straight-line fit. A
choice was made by inspection as to which
data point should be used for the break point
between the high-contrast segment and the
low-contrast segment. If the resulting lines
indicated that a poor choice had been made,
the procedure was repeated until a good fit
Invest. Ophthtttmol. Vis. Sci.
November 1980
1368 Keck, Montague, and Burke
10
1.0
100
ADAPT CONTRAST
Fig. 4. Initial MAE speed as a function of the contrast of the moving grating for Subject T. B.,
who is using the method of tracking. The spatial frequency is indicated in cycles per degree
next to each cui-ve. These data were not normalized. The mean value of the standard errors for
Subject T. B. was 0.50.
was obtained. The two straight-line segments
are not assumed to be the actual functional
form of the initial MAE speed but are used
merely as an aid in analyzing and describing
the data.
In order to make meaningful comparisons
across spatial frequencies, the magnitude estimation data must be normalized to compensate for session-to-session variability in the
estimates. A 4 cy/deg grating was included in
each sequence to serve as a standard for
comparison. The mean values of the 4 cy/deg
standards presented during each run were
used to normalize all spatial frequency data to
the 4 cy/deg data. (The normalizing factor for
each spatial frequency is given in each figure
legend.)
Fig. 2 shows the normalized MAE speed at
all the measured spatial frequencies for Subject P. H. Similar curves are shown for Subject F. M. in Fig. 3. And the curves in Fig. 4
are fits to data obtained from Subject T. B. by
the tracking method rather than the magnitude estimation method. Note that dashed
curves are used for spatial frequencies less
than 1 cy/deg and that the labels next to each
curve refer to the spatial frequency.
The families of curves presented in Figs. 2
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
to 4 for the three subjects showed the following similar trends. (1) The threshold contrast
for obtaining an MAE was at a minimum at an
intermediate spatial frequency (about 1 cy/
deg); it increased for both higher and lower
spatial frequencies. (The threshold contrast is
the intercept of each curve at zero MAE
speed.) (2) For spatial frequencies less than 1
cy/deg the MAE speed increased linearly
with the logarithm of the grating contrast
above threshold. (3) For higher spatial frequencies the MAE speed did not increase
linearly with log contrast over the entire contrast range. For contrasts greater than about
six times threshold, the rate of increase of the
MAE speed was reduced. This was most pronounced at the highest spatial frequencies
where we began to see a saturation of the
MAE speed.
Discussion
The MAE results from the fatiguing of that
portion of the visual system which detects the
motion of a stimulus. The present experiments examine the response of this direction-specific system as a function of the
stimulus contrast and spatial frequency.
First, consider the minimum contrast re-
Volume 19
Number 11
Spatial periodicity of moving gratings 1369
1000
100 —
0.2
SPATIAL FREQ. (c/deg)
Fig. 5. Contrast sensitivity as a function of spatial frequency for moving gratings obtained from
threshold for detecting motion (0) and extrapolation to zero MAE (A).
quired to obtain a response of this system at a
particular spatial frequency. An approximate
value of this minimum contrast could be obtained from our data by extrapolating the
MAE speed curves to zero; the intercepts
would represent the threshold contrasts required for obtaining an MAE. The reciprocal
of the threshold contrast values for T. B.
(from Fig. 4) are plotted as a function of the
spatial frequency in Fig. 5 (triangles). These
values ought to be compared with the more
conventional measurement of a threshold
contrast for the detection of a moving grating,
with motion direction used as the detection
criterion. We have measured these threshold
contrasts for Subject T. B., and their reciprocals are also plotted in Fig. 5 (circles). Note
that the sensitivity is greater for the latter
measurements. This is a reasonable result,
since the MAE is a fatiguing effect, and one
might expect that the threshold for obtaining
it would be somewhat higher than that required for detecting the grating motion.
However, the striking similarity between the
two curves supports the notion that the cell
population which mediates the detection of a
moving grating is also responsible for the
MAE.
Next, consider the response above threshold. For low spatial frequencies, less than 1
cy/deg, the MAE speed was found to increase linearly with log contrast up to 80%
contrast. However, for spatial frequencies
greater than 1 cy/deg the increase of the
MAE speed with log contrast was not found
to be linear over the entire range of contrast.
The nonlinearity is especially pronounced for
the highest spatial frequencies, 8 and 10
cy/deg. For these spatial frequencies there
is very little increase in MAE speed
above 25% contrast. Thus, for high spatial
frequencies, the MAE speed shows a limited
or compressed contrast response. These results are in agreement with published data
available for some spatial frequencies.4' 5> 10
The MAE measurements which we have
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
1370 Keck, Montague, and Burke
reported in this paper must be attributed to
the direction-specific mechanisms in human
vision. We have found that this subset of cells
has a more limited contrast response to high
spatial frequencies than does the visual system as a whole. Although the transient cell
populations have been shown12 not to respond well to high spatial frequencies, there
is no convincing evidence that all directionspecific cells fall solely within the transient
system. It remains to be shown whether
those direction-specific cells which respond
to high spatial frequencies can be classed as
either transient or sustained.
We thank Paul Hrich for assistance in the laboratory
and for being a subject on numerous occasions. We also
thank Allan Pantle for many helpful discussions.
REFERENCES
1. van Nes FL, Koenderink JJ, Nas H, and Bouman
MA: Spatiotemporal modulation transfer in the
human eye. J Opt Soc Am 57:1082, 1967.
2. Watanabe A, Mori T, Nagata S, and Hiwatashi K:
Spatial sine-wave responses of the human visual system. Vision Res 8:1245, 1968.
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
Incest. Ophlhalmol. Vis. Sci.
November 1980
3. Sekuler R and Ganz L: Aftereffect of seen motion
with a stabilized retinal image. Science 139:149,
1963.
4. Pantle A and Sekuler R: Contrast response of human
visual mechanisms sensitive to orientation and direction of motion. Vision Res 9:397, 1969.
5. Pantle A, Lehmkuhle S, and Caudill M: On the capacity of directionally selective mechanisms to encode different dimensions of moving stimuli. Perception 7:261, 1978.
6. Barlow HB and Hill RM: Evidence for a physiological explanation of the waterfall phenomenon and
figural aftereffects. Nature 200:1345, 1963.
7. Sutherland NS: Figural aftereffects and apparent
size. Q J Exp Psycho] 13:222, 1961.
8. Sekuler R and Pantle A: A model for aftereffects of
seen movement. Vision Res 7:427, 1967.
9. Pantle A: Motion aftereffect magnitude as a measure
of the spatio-temporal response properties of direction-sensitive analyzers. Vision Res 14:1229, 1974.
10. Keck MJ, Palella TD, and Pantle A: Motion aftereffect as a function of the contrast of sinusoidal gratings. Vision Res 16:187, 1976.
11. Keck MJ and Pentz B: Recoveiy from adaptation to
moving gratings. Perception 6:719, 1977.
12. Harwerth RS and Levi DM: Reaction time as a measure of suprathreshold grating detection. Vision Res
18:1579, 1978.