Continuous measurement of visible persistence.

Journal of Experimental Psychology:
Human Perception and Performance
1985, Vol. I I , No. 6, 711-725
Copyright 1985 by the American Psychological Association, Inc.
0096-1523/BJ/S00.75
Continuous Measurement of Visible Persistence
Erich Weichselgartner and George Sperling
New York University
In the synchrony judgment paradigm, observers judge whether a click precedes or
follows the onset of a light flash and, on other trials, whether or not a click precedes
light termination. The interclick interval defines the duration of visible persistence.
An elaboration of this method consists of two phases: In Phase 1, the luminance
of a reference stimulus is psychophysically matched to the peak brightness of the
test flash. Five luminance values between . 1 and 1.0 of the reference stimulus are
used subsequently. In Phase 2, a random one of the five reference stimuli, a test
flash, and a click are presented; the observer judges whether the click occurred
before or after the brightness of test flash reached the reference value (on onset
trials) or decayed below it (on termination trials). This method was validated on 3
subjects with test stimuli whose luminance rises and decays slowly in time, and then
was used to trace out the precise subjective rise and decay (temporal brightness
response function) of brief flashes.
Visible Persistence
A brief visual stimulus presented to a subject
is not perceived to end abruptly but to fade
out gradually. The time difference between the
physical termination of the stimulus and its
perceptual termination has been investigated
in a variety of paradigms. These paradigms
fall into two main classes: those that infer vi-
sual storage from the accuracy of subject's re-
ports and those that depend on subjective re-
ports (e.g., Brindley's [1960] Class A and Class
B procedures).
Typical accuracy procedures are the partial
report paradigm and various picture comple-
tion paradigms. In the partial report paradigm
(Sperling, 1960), the observer views a brief
flash of a matrix of letters. Afterwards, a tonal
cue that can be precisely located in time is
used to request report of a randomly selected
row of the stimulus. The decline of response
accuracy with cue delay indicates the duration
of short-term visual storage (iconic memory—
Neisser, 1967). Picture completion paradigms
require the observer to integrate information
from two successive flashes in order to identify
This research was supported by the United States Air
Force, Life Sciences Directorate, Grant AFOSR-80-0279.
The authors wish to thank Misha Pavel for helpful sug-
gestions.
Requests for reprints should be sent to George Sperling,
Psychology Department, New York University, 6 Wash-
ington Place, Room 980, New York, New York 10003.
a target letter (e.g., Eriksen & Collins, 1967)
or to detect a missing dot in a regular dot ma-
trix (Hogben & DiLollo, 1974).
Subjective procedures require the observer
to make a subjective judgment of perceived
presence or absence of a stimulus. An example
is the synchrony judgment paradigm (Sperling,
1967), in which the subject adjusts an auditory
event (e.g., click) to a visual event (e.g., light
termination). In synchrony judgment para-
digms there are no correct or incorrect re-
ports—the report is simply taken as its face
value as an indication of the subject's percep-
tual state. In a technical sense, the synchrony
judgment paradigm is a semantic exploration
of how the words visible or see are applied in
tachistoscopic viewing conditions.
On the whole, estimates of the duration of
visible persistence obtained in the two classes
of paradigms are similar, although there are
some apparent discrepancies. Typical persis-
tence times are found to be in the range of
100-300 ms (see Coltheart, 1980, and Long,
1980, for reviews). According to the unitary
persistence hypothesis, the discrepancies in the
duration of visible persistence between the two
classes of paradigms are caused by different
task requirements, which depend on different
segments of one-and-the-same decay function.
For example, Long (1980) states that a driving
force for the popularity of the visible persis-
tence paradigm is the assumption that the de-
cay functions for iconic memory (accuracy

712 ERICH WEICHSELGARTNER AND GEORGE SPERLING
procedures) and visible persistence are the
same. However, no one has come up yet with
a measure for those decay functions.
The studies concerning visible persistence
fall into two groups: (a) Studies concerned with
the time difference between stimulus termi-
nation and perceptual termination (Adelson,
1978; Appelman, 1980; Bertelson & Tisseyre,
1969; Bowen, Pola, & Matin, 1974; Sakitt,
1976a, 1976b) and (b) studies concerned with
the total phenomenal duration of the stimulus
that require reliable judgments of both onset
and termination, the estimated stimulus du-
ration being the time difference between judged
onset and termination (Efron, 1970a, 1970b,
1970c; Haber & Standing, 1969, 1970; Sper-
ling, 1967). There are no statements concern-
ing the actual form of the rise and decay curves
or of the complete representation of the stim-
ulus as a function of time in the subject's visual
system.
Clearly, a method to measure the entire
moment-to-moment time course of visible
persistence is needed. Because the term visible
persistence already has several meanings (e.g.,
it is used to describe the time difference be-
tween physical stimulus termination and per-
ceptual stimulus termination), the entire
function will be called temporal brightness re-
sponse (TBR). The TBR describes how the
perceived brightness of a brief visual stimulus
changes as a function of time. The purpose of
this article is to prove the feasibility of mea-
suring the TBR and to measure TBRs to brief
flashes for 3 observers.
Elaborated Synchrony Judgment Paradigm
In an elaboration of Sperling's (1967)
method, the following intermodal synchrony
judgment paradigm was developed. An ob-
server views two adjacent stimuli: a reference
stimulus, which is presented at the beginning
of the trial and remains on with constant lu-
minance during the trial, and a test stimulus
of varying luminance, which is presented and
terminated sometime during the middle of the
trial. Luminances are are adjusted so that from
the observer's point of view the test stimulus
initially appears dimmer than the steady-state
reference stimulus but increases in intensity
until eventually it becomes as bright or brighter
(Figure 1).
At some instant, the brightnesses of the test
and reference stimuli appear to match; that is
the match time. The aim of the method is to
discover precisely when the match time occurs.
In a time interval around match time, the sub-
ject is presented with a click.
If the subject perceives the click to have oc-
curred before match time, he or she presses a
left-hand response key; if the click occurs after
match time, a right-hand response is made.
Alternatively, the task of the subject can be
construed as pressing the right key if, at the
instant the click occurred, the test was brighter
than the reference stimulus, and otherwise to
press the left key. On termination judgment
trials, the keys are reversed.
Repeated judgments result in a psychomet-
ric function, the probability of a right response
as a function of time. The point of subjective
equality (pse) is the 50% point of the psycho-
metric function; it corresponds to the time at
which the brightness of the reference matches
that of the test. A psychometric function ob-
tained with a particular luminance of the ref-
erence stimulus determines only one match
time. The entire TBR function is obtained by
obtaining match times for a full range of lu-
minances of the reference stimulus, and by de-
termining match times near the onset and also
near the termination of the test stimulus.
Verification Procedure
Our primary interest is in the TBR function
of a very brief stimulus. However, it is impor-
tant to first demonstrate that a subject can
make reasonable match time judgments with
a slowly varying stimulus for which the TBR
can be assumed to approximately track the
temporal luminance function of the physical
stimulus. An example of such a physical stim-
ulus is the gradual fading out of a light bulb
turned off with a dimmer. Therefore, we first
determine TBRs for a control condition in
which test field luminance increases linearly
during a 600-ms period, stays at its maximum
for 300 ms, and then turns off linearly during
another 600-ms period. The data from this
control condition with real "physical persis-
tence" can be used to evaluate the method.
For example, when the rising part of the test
stimulus is under investigation (onset trials),
we expect the match times to increase with

VISIBLE PERSISTENCE 713
increasing luminance of the reference stimu-
lus, and when the decaying part of the test
stimulus is under investigation (termination
trials), we expect the match times to increase
with decreasing luminance of the reference
stimulus. After Experiment 1 (with ramped
onsets and terminations) we proceed to Ex-
periment 2, which determines TBRs for brief
stimuli.
General Method
Overview
For both experiments, the verification experiment and
the main experiment, the same apparatus and a similar
method were used. We begin with a detailed description
of Experiment 1 and add the differing details of Experiment
2 later.
Apparatus and Stimuli
All experiments were completely computer controlled.
The stimuli were presented on a Hewlett Packard 1310A
cathode-ray tube (CRT) with a fast P4 phosphor. The CRT
was driven by a Digital Equipment Corporation PDP-11/
34 computer via an especially designed display interface
(Kropfl, 1975) and software for real-time vision experi-
ments (Melchner & Sperling, 1980). The equipment made
it possible to control display parameters in the millisecond
range. The click was presented with Sennheiser HD 414
headphones.
Spatial arrangement. The stimuli consisted of two, ad-
jacent, square-wave gratings (10 cycles/degree), separated
by a gap of 0.25° visual angle, and each subtending a visual
angle of 1.6° X 1.6° (Figure 2). The two gratings were
viewed binocularly from a distance of 90 cm, with a fixation
dot in between. Each stimulus was composed of a 32 X
32 dot matrix, with every second horizontal line suppressed.
The individual dots can be seen in Figure 2, Panel a. This
grating was used because it is a spatial frequency to which
the visual system is highly sensitive and because the visible
presence of the grating might facilitate the observer's task
in judging the persisting presence of the stimulus. It is pos-
sible, indeed probable, that visible persistence depends on
the precise geometric form of the test stimulus. However,
in their judgments, observers were instructed to match the
overall brightness of the stimuli, and the grating structure
appears to have been unimportant.
Stimulus intensity. Over trials, reference stimuli with
five different luminances were used (see first column of
Table 1). The intensity of the test stimulus followed a
ramped square-wave function in Experiment 1 (control
condition) and a brief pulse function in Experiment 2.
The display was refreshed at a rate of 67 Hz, well above
the critical flicker-fusion frequency of about 45 Hz for these
stimuli. The high refresh rate guaranteed that the variations
in intensity were perceived to be continuous. The intensity
of the test stimulus changed with every refresh, while the
test
3 —
2 —
I
3
BKGD
Figure 1. The elaborated synchrony judgment paradigm with the stimuli of Experiment 1. (The ordinate
indicates the luminance of the stimuli; the abscissa indicates time. All times are given relative to the onset
of the test stimulus, which consists of two 600-ms ramps connected by a 300-ms plateau. The events defected
are the subject's response [R] on the previous trial; a random time interval of 200-2,000 ms [a] after which
the reference stimulus appears; a random time interval of 450-900 ms [b] before test stimulus begins; a click
at a predetermined time [c]. The subject judges whether the click occurred before or after the brightness of
the test equaled the brightness of the particular reference stimulus [0.5 of maximum reference luminance is
shown). A uniform background light [BKGD] of 0.35 cd/m2 is present continuously.)

714 ERICH WEICHSELGARTNER AND GEORGE SPERLING
intensity of the reference stimulus remained constant (Fig-
ure 2, Panels b and c). The luminance levels were externally
checked with a United Detector Technology 40x Opto Me-
ter. Two incandescent lamps were mounted to the left and
to the right side of the display, providing a background
illumination of 0.35 cd/m2 throughout.
The click was generated with a 0.5-ms wide pulse to the
headphones. A relative measure of click intensity was ob-
tained by having subjects match the loudness of a 1000
Hz tone to the loudness of the click. The click and the
tone were perceived to be equally loud when the intensity
of the 1000 Hz tone was 80 dB above its own threshold
(average of 3 subjects).
Procedure
Individual trials. At the beginning of a sub-block of 100
trials, the subject was shown a message indicating whether
onset or termination judgments were required. After a
random time interval of 200-2,000 ms (Interval a in Figure
1), the reference stimulus was turned on. After another
1cm
Time i Time in ms
figure 2. Panel a: photograph of the reference and test
stimuli on the cathode-ray tube (CRT). (Each stimulus
subtends 1.6" X 1.6" [degrees of visual angle], and they
are separated by a 0.25" gap. The spatial frequency is 10
cycles/degree. In this example, both stimuli have the same
intensity.) Panels b and c: stimulus generation (schematic).
(The height of each bar represents luminous energy [Sper-
ling, 1971] on the CRT. The distance between bar onsets
represents the time between refreshes [ 15 ms]. [b] Reference
stimulus: 1,500 ms of the 3,000-ms long reference stimulus
are shown; [c] Test stimulus, Experiment 1. Because of
limitations in print quality, only 1/3 of the actual flashes
are shown.)
random time interval of 450-900 ms (Interval b in Figure
1), the test stimulus began. On the first trial, the onset time
of the click was randomly chosen within an interval of ±
200 ms around starting points (initial values), which were
determined in preliminary experiments (see below). The
reference stimulus was physically present for 3,000 ms; the
test stimulus was on for 1,500 ms in Experiment 1 (Figure
1), and for 31 ms (3 refreshes) in Experiment 2; the back-
ground illumination was constant throughout.
The subject's task was to decide whether the click had
occurred before or after the instant of perceived brightness
match between the two stimuli, and, accordingly, to press
one of two response keys. This is a forced-choice task with
two-alternatives, "click before" and "click after." The sub-
ject had unlimited time to arrive at a decision but typically
responded within 2 s. The response automatically initiated
the next trial.
Blocks of trials. The subjects were run in two blocks
of 400 trials, each block being subdivided in four 100-trial
sequences of onset or termination judgments, respectively.
The points of subjective equality (i.e., the match times
when the brightnesses of test stimulus and reference stim-
ulus appear to match) were determined by a transformed
up-down procedure (see below). There was a total of 20
conditions per experiment: reference stimuli of five lu-
minances, test stimulus to the left or to the right side of
fixation, and onset or termination judgment (5 X 2 X 2).
Blocks of 400 trials lasted about 45 min, separated by a
break.
Staircase procedure. To find the match times most ef-
ficiently, a transformed up-down procedure described by
Levitt (1971) was used. Two interleaved staircases to es-
timate the X29 and X/,, points on the psychometric func-
tion were run; from these, both the X
5 point (pse) and the
variance of the underlying distribution are computable. In
addition to the well known advantages of staircase methods,
the interleaved estimation of two (or more) points elimi-
nates sequential stimulus dependencies and therefore en-
sures that the subject cannot anticipate stimuli and adjust
responses accordingly.
The first click in any one of the 20 conditions was ran-
domly placed around an initial value found in preliminary
experiments. The following clicks were determined ac-
cording to Levitt's transformed up-down procedure. In
Staircase I, which estimates X.M, the click time is decreased
by a fixed time interval (step size), when the subject re-
sponds "Click time occurred after match time," and in-
creased by step size, when the subject responds two con-
secutive times "Click time occurred before match time."
In Staircase 2, which estimates X.7,, the converse applied.
As determined by preliminary experiments, a step size
of 100 ms was chosen for the first five trials, and a step
size of 20 ms afterwards. This step size coincides with re-
ports on the resolving power for temporal order judgments,
which was found to be 15-44 ms. (Exner, 1875; Hirsh &
Sherrick, 1961; Rutschmann, 1966). Per 400 trials, 20
staircases were run simultaneously (and interleaved), one
for each combination of reference stimulus (5), visual field
(2), and point on the psychometric function (2). The Sub-
jects BW and SW did not know about this procedure.
Each staircase consisted of 20 trials (stopping rule). The
.50 point on the psychometric function (match time) was
estimated by computing, separately for each condition, the
average of the click times for each staircase (X
2», X T1)
from the first reversal on, and then by computing the arith-

VISIBLE PERSISTENCE 715
Table 1
Match Times and Other Data From Experiment 1
Reference
cd/m2
Onset
0.31
0.67
1.23
1.95
2.57
MT
186
240
328
463
703
BW
SE
28
27
33
31
60
DIP
225
196
269
289
258
MT
36
69
113
170
400
Subject
EW
SE
23
23
32
26
25
DIF
130
194
136
130
256
MT
70
136
235
342
452
SW
SE
15
18
11
30
43
DIF
75
73
100
170
182
Peak alone 703 10 563 686 13 317
Onset
2.57
1.95
1.23
0.67
0.31
712
1,014
1,242
1,389
1,467
56
40
26
22
23
297
360
242
370
170
705
1,108
1,336
1,403
1,485
18
21
33
36
31
395
304
145
102
145
940
1,212
1,331
1,420
1,466
39
18
31
21
17
349
168
161
154
86
Note. MT is mean match time, SE is standard error of the mean, and DIF is difference between 0.3 and 0.7 estimates
of the match-time psychometric function (equal to 1.09 X Standard Deviation of the Match-Time Density Function
assuming Normality). AH data are in millseconds. Each MT is based on 240 observations, except the MTs for the
absolute brightness peak of the test stimulus ("peak alone"), which were determined independently with 400 observations
per subject. DIF could not be computed for BW's and EWs peak alone judgments. The stimulus was a ramped square-
wave function.
metic mean of both averages: match time = 0.5 (X.29 +
X
71). Assuming that the psychometric functions are sym-
metric in their midrange (not necessarily normal), this
method is known to give bias-free estimation of X,0, the
match time. (The symmetry of the psychometric functions
was not tested formally because graphs of the psychometric
functions derived from the judgments did not show any
obvious asymmetries.)
Subjects
Three male graduate students (BW, EW, and SW) par-
ticipated in the two main experiments. All subjects had
normal orcorrected-to-normal vision. Two of the subjects
(BW and SW) were paid $4.00 per hour. Subject EW had
taken part in exploratory experiments; Subjects BW and
SW were initially naive subjects. There were only a few
practice trials for Subjects BW and SW.
Experiment 1
Experiment 1 consisted of two phases. In
Phase 1, stimulus parameters for Phase 2 were
empirically determined. In Phase 2, the TBR
for a ramped test stimulus was measured using
the parameters from Phase 1.
Method
Procedure
Phase 1. The experiments in Phase 1 determined (a)
the luminances of the reference stimulus that matches peak
test brightness and, hence, the choice of appropriate lu-
minances for the reference stimuli, and (b) the time of
occurrence of a click that co-occurs with the test peak and,
hence, the choice of appropriate initial click values. The
peak luminance of the test stimulus itself was arbitrarily
chosen at a low value (3.43 cd/m2) to aviod obvious retina]
afterimages while still being a moderately bright flash.
The elaborated synchrony judgment paradigm requires
determination of the luminance of a reference stimulus
whose brightness exactly matches the peak brightness of
the test stimulus. Luminance matching was carried out
with a symmetric up-down staircase of 30-50 presentations
on which the luminance of the reference stimulus was var-
ied, and the subject judged whether the reference was
brighter or dimmer than the peak brightness of the test
stimulus. The luminance of the brightness-matching ref-
erence was 0.75 of the peak luminance of the test stimulus.
The results were so nearly identical for the Subjects BW,
EW, and SW that for practical reasons the same reference
stimulus was used for all subjects. (Individual results are
given in the Results section below.)
In order to trace out the TBR, five reference stimuli
were chosen for Experiment 1, whose values were .125,
.25, .5, .75, and 1.0 of the luminance of the reference stim-
ulus that matched the brightness peak of the test stimulus.
In a separate sequence of trials, the subjects judged the
synchrony of the click and the peak of test brightness by
means of same kinds of staircases as used in the main
experiment. This judgment is made without regard to the
reference stimulus. Together, the isolated brightness and
temporal matches of Phase 1 pinpoint the brightness and
time of the perceived brightness peak of the test stimulus.
Phase 2. In Phase 2 of Experiment 1, the TBR was de-
termined for a ramped test stimulus by means of the psy-