• No results found

Perception of visual motion coherence by rats and mice

N/A
N/A
Protected

Academic year: 2021

Share "Perception of visual motion coherence by rats and mice"

Copied!
6
0
0

Loading.... (view fulltext now)

Full text

(1)

www.elsevier.com/locate/visres

0042-6989/$ - see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.visres.2006.02.025

Perception of visual motion coherence by rats and mice

R.M. Douglas

a,b,¤

, A. Neve

b

, J.P. Quittenbaum

b

, N.M. Alam

a

, G.T. Prusky

b

a Department of Ophthalmology and Visual Sciences, University of British Columbia, 2550 Willow Street, Vancouver, BC, Canada V5Z 3N9

b Canadian Centre for Behavioural Neuroscience, The University of Lethbridge, 4401 University Drive, Lethbridge, Alta., Canada T1K 3M4 Received 13 November 2004; received in revised form 17 January 2006

Abstract

The coherence thresholds to discriminate the direction of motion in random-dot kinematograms were measured in rats and mice.

Per-formance was best in the rats when dot displacement from frame-to-frame was about 2 degrees, and frame duration was less than 100 ms.

Mice had coherence thresholds similar to those of rats when tested at the same step size and frame duration. Although the lowest

thresh-olds in the rats and mice occasionally reached human levels, average rodent values (»25%) were 2–3 times higher than those of humans.

These data indicate that the rodent and primate visual systems are similar in that both have local motion detectors and a system for

extracting global motion from a noisy signal.

© 2006 Elsevier Ltd. All rights reserved.

Keywords: Extrastriate cortex; MT; Random dot; Kinematograms; Visual water task; Psychophysics

1. Introduction

The detection and e

Vective use of visual motion is

widespread in the animal kingdom. Experimental

investi-gations of the neural mechanisms of motion perception

have been conducted in species as diverse as

Xies, pigeons,

and monkeys (

Cli

Vord & Ibbotson, 2002

), and simple

neural circuits capable of detecting local motion of single

elements have been studied in the

Xy (

Reichardt, 1961

)

and rabbit retinas (

Barlow & Levick, 1965; Fried, Munch,

& Werblin, 2002

). In recent years, it has become common

to use dynamically moving random-dot patterns (or

kinematograms) to assess visual motion, as they enable

the study of motion perception in the absence of

posi-tional or form cues (

Nakayama & Tyler, 1981

).

Percep-tual mechanisms that can detect common motion of

many elements have been identi

Wed in humans and other

primates (

Braddick, 1974; Morgan & Ward, 1980;

Wil-liams & Sekuler, 1984

) and localized to extrastriate

cor-tex (

Baker, Hess, & Zihl, 1991; Newsome & Paré, 1988

). It

is, however, an open question whether primates have

unique cortical capabilities, or whether such a functional

organization is a fundamental property of mammalian

visual systems. If extrastriate analysis of global visual

motion is common to mammals, this would have

implica-tions for its evolution and ecological utility. In addition,

studying motion perception in laboratory rodents would

facilitate experiments into the cellular and molecular

substrates of visual motion processing. To this point,

however, no systematic study of motion perception has

been conducted in rodents, largely because it has been too

di

Ycult to test their visual motion thresholds

psycho-physically. Over the past several years, we have used the

Visual Water Task (

Prusky, West, & Douglas, 2000

) to

quantify acuity and contrast sensitivity thresholds in rats

and mice (

McGill, Douglas, Lund, & Prusky, 2004;

Pru-sky et al., 2000; PruPru-sky & Douglas, 2004

), and we report

here that this task can be adapted to measure their visual

motion thresholds.

In order to determine whether rodents have visual

motion capabilities and anatomical substrates comparable

to primates, we compared the dot motion coherence

thresh-olds of rats with those of humans.

* Corresponding author. Fax: +1 604 876 6553.

(2)

2. Methods

2.1. Subjects

Six adult female Long–Evans rats and six adult C57BL/6 mice were used in the experiments. The animals were between 90 and 150 days of age at the start of training, and testing continued over the subsequent 4–6 months. All procedures were authorized by the University of Lethbridge Animal Care Committee, which approves procedures that are conducted in accordance with the standards of the Canadian Council on Animal Care.

To facilitate comparison with human motion vision, three human observers viewed the same stimuli from the same distance. They used pushbuttons to start each trial and to record their decisions about which side had the positive stimulus. Human testing was conducted in accor-dance with the guidelines of the University of Lethbridge Human Subjects Research Committee.

2.2. Apparatus

The Visual Water Task (Prusky et al., 2000) uses a trapezoidal tank (120 cm L £ 80 cm W £ 26 cm W £ 55 cm H) with two ViewSonic EF70, 17⬙ computer monitors facing into one end (Fig. 1). A midline divider (46 cm) extends into the pool from between the monitors, each of which subtended about 30° degrees when viewed from the end of the divider. The tank was Wlled with water, and an escape platform was placed just below the surface directly under one of the monitors. The platform was always located below the positive (+; reinforced) stimulus. Visual stimuli were generated by, and the experiments were controlled with custom software.

2.3. Stimuli

Kinematograms were computed separately for each monitor before every trial, and were then played continuously until the end of the trial. A random-dot kinematogram consisted of 24 frames or images. The diame-ter of the round dots was 1.9° for the rats, and 2.9° for the mice as studies from our laboratory (e.g., Prusky et al., 2000) have shown that the grating acuities of Long–Evans rats and C57BL/6 mice are 1.0 and 0.55 cycles per degree (c/d), respectively. The total area of the dots was about 20% of the screen area, and this corresponded to there being 62 (rats) or 28 (mice) dots on each image frame. The dots had a luminance of 96 cd/m2 and the background was 1.1 cd/m2.

Each dot could move from one frame to the next, or disappear. The step size, or amplitude of the displacements from fame to frame, was the same for all dots, with only the direction of movement varied between frames. Motion noise was introduced by having a percentage of the dots move in random directions (Fig. 2) with the additional constraint that there was no net movement. The remainder of the dots all moved in one, coherent direction. Dot displacements of 0.24°, 0.47°, 0.94°, 1.88°, 3.77°, 5.65°, and 7.54° were used in diVerent blocks of trials.

After moving for 2 or 12 frames (lifetime) each dot was randomly repo-sitioned, thus ensuring a constant number of dots on screen. A constant percent of dots died and were reborn elsewhere on each frame, and motion wrapped from the 24th to the 1st frame so that the kinematogram could be played continuously. Longer lifetimes were not possible with a looping requirement, and thus dot lifetimes of 2 and 12 frames limited the maxi-mum eVective coherence to 50 and 92%, respectively (Bischof, Reid, Wylie, & Spetch, 1999).

Frame duration set the stimulus onset asynchrony (SOA) for the dots to appear on two successive frames, and frame duration was a multiple of the video screen refresh rate. In the experiments with rats, 1, 2, 4, 8, 13, 19, and 26 screen refreshes were employed. As the monitors had an 85 Hz ver-tical refresh rate, these values translate to frame durations of 12, 24, 48, 94, 153, 224, and 306 ms.

2.4. Behavioral testing

The animals were trained to associate swimming to the positive stimu-lus, regardless of its left/right location, with escape from water. The initial training consisted of training the rat or mouse to distinguish between 100 and 0% coherence. Rats learned to grasp the end of the divider and inspect both screens before making a choice; mice tended to slow down at the bar-rier and swim back and forth before making a choice. After an animal had Wnished a trial, it was returned to a holding box to dry and groom while the other 5 animals completed the same trial. Sessions consisted of 15–20 trials per animal. The training kinematograms had a step size of 0.471°, a frame duration of 12 ms, and a lifetime of 10 frames.

Once the rats reached at least 80% accuracy over a block of 10 trials, we changed the task to identifying the direction of motion. This had the advantage that the kinematograms on the two monitors were identical in coherence, step size, frame duration, and lifetime, and only diVered in

Fig. 1. Schematic top-view of the Visual Water Task. The task uses a trape-zoidal pool with two computer monitors at one end displaying either a pos-itive (+, reinforced) or negative (¡, non-reinforced) stimulus. An escape platform is placed below the monitor showing the positive stimulus and the left/right position of the pair is randomized. The platform surface is just below the water level and is not visible to the animal. Animals are released at one end and they swim to the end of a divider where they inspect the two monitors before choosing one arm of the maze to swim into.

Submerged Platform Choice Plane Video Monitors Start Point

+

Water

Fig. 2. Schematic diagram of stimulus attributes. Motion coherence dis-plays consisted of multiple dots that could move in any direction across frames. After moving on 2 or 12 successive frames, the dot disappeared and a new dot was created at another random location on the screen (dashed circles). At 0% coherence, all dots moved in random directions with the average being zero. At maximal coherence all moving dots moved coherently in one direction.

100% Coherence 0% Coherence

(3)

direction of the coherent motion. For the rats, the discrimination was between leftward (negative) and rightward (positive) coherent motion. For the mice, the task required a discrimination between horizontal (positive) and vertical (negative) motion.

Coherence thresholds were then established by reducing the percent of dots moving coherently for blocks of ten trials. A preliminary threshold was established when performance fell below seven correct in a block of 10 trials. The threshold was assessed several times or until a reliable pattern of performance was generated. Generally, three or four estimates were so obtained.

3. Results

3.1. Behavior and training

All the rats and mice were eventually able to

discrimi-nate the direction of motion in random-dot

kinemato-grams, and did so even when coherence was 25% or less.

The directional discrimination, however, was a di

Ycult test

for the animals to learn. Initially we tried testing the rats

with a dot lifetime of 2 frames, but none of the rats could

learn to make the discrimination. This may have been

because 50% was the maximum coherence possible with

single frame motion. When dot lifetime was increased to 12

frames we could start testing at 92%, and the animals had

fewer problems learning, taking 86 trials on average to

learn the new discrimination (range 30–200 trials). Once the

rats learned the task, performance was close to perfect at

high coherencies, and then declined as the percentage was

decreased (

Fig. 3

). Testing was discontinued once

perfor-mance dropped below 70% because animals rapidly become

confused when they made many errors. Although most of

the experiments used dot lifetimes of 12 frames, we tried

reducing dot lifetime later, after the rats had had extensive

experience with task. On the second attempt, all of the rats

were able to discriminate the direction of motion with

two-frame lifetimes, but there was no signi

Wcant diVerence in

coherence thresholds for 2-frame (31.0 § 2.85) and 12-frame

motion (25.21 § 4.59) kinematograms (t D 1.0, df D 5,

p D 0.36).

We attempted to train the mice in the same way as the

rats, but they could not learn the initial discrimination of

maximal coherent motion versus one that had no coherence

(i.e., 0%). After this initial failure, we tried training the mice

on an alternative task in which the requirement to

discrimi-nate between horizontal motion (+) and vertical motion

(¡). This they did learn in an average of 360 trials (range

262–520). Like the rats, once the mice had learned the task,

they could perform at a high level.

Fig. 4

A shows one

mouse making only occasional errors over 544 trials as

coherence was lowered progressively six times.

Perfor-mance improved over the experiment: Initially the mouse

started making errors at dot coherencies of 60%, but on

subsequent runs the mouse made fewer errors and the

threshold stabilized at about 25%. On average 449 (§4.1)

Fig. 3. Performance of the rats for kinematograms with a dot

displace-ments of 3.8° per 47 ms frame. Testing started at 92% coherence, and that was reduced until performance dropped below 70%. This was repeated several times, and an average of 239 trials was needed to establish the threshold for each animal. Each point is the mean percent correct for the 6 animals, and the error bars are for one standard error.

0 10 20 30 40 50 60 70 80 90 100 40 50 60 70 80 90 100

% Coherence

% Correct

Fig. 4. (A) Example of coherence threshold testing in one mouse. Initially the mouse made errors almost as soon as the coherence was reduced, but with further testing, performance stayed near 100% until the coherence had dropped to below 30%. (B) The average performance of the 6 mice are plotted as a function of coherence. The error bars show one standard error above and below the means.

0 100 200 300 400 500 0 10 20 30 40 50 60 70 80 90 100 40 50 60 70 80 90 100

% Correct

0 10 20 30 40 50 60 70 80 90 100 40 50 60 70 80 90 100

% Coherence ( )

% Correct

% Coherence

Trial

A

B

(4)

trials were used to measure one coherence threshold. The

corresponding threshold in rats was obtained more rapidly

(239 § 9.9 trials) but with a shallower frequency-of-seeing

curve (

Fig. 3

) than that for the mice (

Fig. 4

B).

3.2. Dot motion

In the

Wrst quantitative experiment with the rats, dot

step size was varied and coherence thresholds were lowest

between 1° and 4° (

Fig. 5

A). At 26.9%, the rat thresholds

were 3.9 times that of the average of 6.8% for the humans

over the same range. The coherence thresholds were higher

for both rats and humans at larger step sizes, and at 7.54°

two rats were unable to discriminate at the maximum

avail-able coherence of 92.5%. Smaller step sizes were also

chal-lenging for the rats, whereas humans had little trouble with

the same sizes, or even at 0.047° at which their threshold

was 10.4% (not shown).

The average rat thresholds were higher than the human

thresholds for all frame durations (

Fig. 5

B). The lowest,

22.9%, was 2.2 times higher than the corresponding 24 ms

threshold for the three humans. Rat thresholds became

pro-gressively higher for longer durations, whereas the human

thresholds varied little. Four of the six rats were unable to

discriminate when the frame duration was 306 ms.

Changing either the size of the jumps a dot makes, or the

time between the jumps can vary the speed of dot motion.

The data from

Figs. 5

A and B are replotted in

Fig. 5

C as a

function of velocity. The two speeds from the two

experi-ments cover the same velocity range, and the coherence

thresholds are similar and low for the 20° s

¡1

to 80° s

¡1

range, but diverge at each extreme. At higher velocities, a

lower threshold was present when the increase in speed was

due to shorter frame durations than it was with larger step

sizes. Conversely at low velocities, small step sizes were

more discernible than long frame durations.

4. Discussion

All the rats and mice learned to discriminate the

direc-tion of random dot kinematograms. The lowest coherence

thresholds were about 20–25%, which was higher than

that of humans, but comparable to those in cats (

Rudolph

& Pasternak, 1996

). The experience in our laboratory is

that mice are harder to train and test, and do not perform

as well as rats in visual tasks. They have a lower acuity

and contrast sensitivities, and we expected that they

would not do as well with motion coherence. The fact they

did as well as the rats may have been due to the use of an

“easier” task. Discriminating vertical from horizontal

motion may be an easier task than leftward versus

right-ward motion. This could be a cognitive problem with the

directional signal having a much lower saliency than the

orientation signal. Alternatively, the local motion

detec-tors may be oriented, but not directional selective, and

this would make the leftward versus rightward

discrimina-tion impossible.

Fig. 5. EVects of varying dot step size and delay on coherence thresholds. (A) Rat and human motion coherence thresholds for displays with dots moving diVerent distances per frame. Each Wlled circle indicates the aver-age value for the six rats, and open squares depict the averaver-age for 3 human observers. Two rats were unable to discriminate above 70% accuracy at the largest step size. Frame duration was 48 ms. (B) Motion coherence thresholds for displays with diVerent frame durations (and thus diVerent stimulus onset asynchronies). Four rats were unable to discriminate at better than 70% accuracy for the largest SOA. Dot displacement was 3.78°. (C) Changing either dot displacement or frame duration alters the speed of the dots. The data from (A and B) are plotted here as a function of velocity. The two manipulations were not equivalent as diVerent thresh-olds were obtained for the same speed. The error bars show one standard error around each mean.

0.2 0.4 0.6 1.0 2.0 4.0 6.0 8.0 0 10 20 30 40 50 60 70 80 90 100 Humans Rats Coherence Threshold (%) Frame Duration (ms) Coherence Threshold (%) Varying Frame Duration Varying Displacement Size 4 6 8 10 20 40 60 100 200 400 0 10 20 30 40 50 60 70 80 90 100

Velocity (degrees/second)

Coherence Threshold (%) 10 20 40 60 100 200 400 0 10 20 30 40 50 60 70 80 90 100 Rats Humans

Displacement per Frame (degrees)

A

B

(5)

4.1. Global motion

The present results clearly demonstrate that rodents

have a global motion system. The tasks can not be solved

by tracking a single dot. Rather, information about many

dots must be integrated over a larger part of the visual

Weld.

In primates, the biological basis of the perception of motion

coherence has been thought to rely on the area around MT,

as lesions there impair the extraction of a motion signal

from a noisy display (

Baker et al., 1991; Newsome & Paré,

1988

). Similarly, area PMLS in the cat appears to be

spe-cialized for motion processing, and lesions of PMLS impair

motion coherence perception (

Huxlin & Pasternak, 2004;

Rudolph & Pasternak, 1996

). At present, there is no

electro-physiological evidence for a homologous extrastriate area

in rat, but induction of c-fos immunoreactivity in the area

after viewing moving stimuli has suggested that rat

antero-lateral (AL) visual area is the equivalent of primate MT

and cat PMLS (

Montero & Jian, 1995

).

In the present study, human thresholds were consistently

lower than the rat thresholds for the same stimuli, usually

by a factor of 2–4, and sometimes by much more. This

dis-crepancy is probably a di

Verence in their visual capabilities

and not due to the animals giving up when the task became

more di

Ycult. Performance was often close to 100% until

the coherence approached the eventual thresholds, and the

Wnal threshold estimates were consistent across animals.

Interestingly,

Bischof et al. (1999)

report that pigeons also

have motion coherence thresholds much higher than

humans. Coherence thresholds of 10–20% in guppy

Wsh

have been reported (

Anstis, Hutahajan, & Cavanagh, 1998

)

so it may be that all vertebrate visual systems have some

ability to extract motion in the presence of noise. This

func-tion may be partly localized in rat to AL and in cat to LS,

but the homology to primate MT is not exact, as these areas

do not appear to be as pro

Wcient as MT is in the primate.

The higher coherence thresholds in the rodents

com-pared to primates could be due to poorer temporal

integra-tion. Each dot appeared on 12 successive frames, and thus

there were multiple displacement and stimulus onset

asyn-chronacy (SOA) combinations that could, in principle, have

been used to compute a motion signal. Human observers

integrate information over multiple frames (

Burr &

Sant-oro, 2001; Watamaniuk & Sekuler, 1992

) but the absence of

a signiWcant diVerence between the coherence thresholds

when using 2- and 12-frame kinematograms suggests that

temporal integration was poor in the rat. Reduced temporal

integration, relative to that for humans, has also been

reported in pigeons (

Bischof et al., 1999

).

4.2. Local motion

The variations in coherence thresholds at di

Verent dot

displacements and SOAs re

Xect the properties of the

mech-anisms that detect the motion of the individual dots. For

both humans and rats, the thresholds increased for

dis-placements larger than 4°, suggesting a maximum

displace-ment (D

max

) of about 8° (

Fig. 5

A). This is a much larger

D

max

than the 0.25°

Wrst found by

Braddick (1974)

, and is

most likely due to the much large dot sizes (

Morgan, 1992

),

long lifetimes (

Todd & Norman, 1995

), and large display

areas (

Eagle & Rogers, 1996

) which extended into

periph-ery (

Baker & Braddick, 1985

). Another way to look at the

data is to consider that, for the rats, the large dots are

roughly equivalent to the small dots used in many human

studies. If one scales a typical human D

max

of 0.25° by the

30-fold di

Verence in acuity, one gets an predicted D

max

of

7.5° for rats. These considerations suggest that the rats and

humans have local motion systems that have very similar

properties when working at the same spatial scale.

Frame duration or stimulus onset asynchrony (SOA) is

also critical for obtaining apparent motion, and the rats

had increasing diYculty as SOA increased and were quite

poor beyond 150 ms. Similar limits have been seen in

human studies of random-dot motion using small dots

(

Braddick, 1974

), but the human performance reported

here is quite diVerent as it was good across all SOAs tested.

This may be because the optimal (and presumably

maxi-mal) delay grows with larger dot displacements (

Eagle &

Rogers, 1997

).

5. Conclusions

The limited spatial and temporal ranges for dot

move-ment are consistent with simple local motion detectors

being used in the rodent to detect individual dot motion.

The low thresholds for motion coherence are evidence for a

more global mechanism that combines the output of these

detectors. Although the visual capabilities of rodent are

more limited than in primates, there are advantages of

using rats and mice as a experimental models in uncovering

the fundamental mechanisms underlying the perception of

visual motion. Genetic tools are more readily available in

rodents than in cats or primates, and the Xat cortex in

rodents has advantages for optical imaging of cortical

activity (

Cang, Kalatsky, Lowel, & Stryker, 2005; Ohki,

Chung, Ch’ng, Kara, & Reid, 2005

).

References

Anstis, S., Hutahajan, P., & Cavanagh, P. (1998). Optomotor test for wave-length sensitivity in guppyWsh (Poecilia reticulata). Vision Research, 38, 45–53.

Baker, C. L., Jr., & Braddick, O. J. (1985). Eccentricity-dependent scaling of the limits for short-range apparent motion perception. Vision Research, 25, 803–812.

Baker, C. L., Jr., Hess, R. F., & Zihl, J. (1991). Residual motion perception in a “motion-blind” patient, assessed with limited-lifetime random dot stimuli. Journal of Neuroscience, 11, 454–461.

Barlow, H. B., & Levick, W. R. (1965). The mechanism of directionally selective units in rabbit’s retina. Journal of Physiology, 178, 477–504. Bischof, W. F., Reid, S., Wylie, D. R., & Spetch, M. (1999). Perception of

coherent motion in random dot displays by pigeons and human. Per-ception & Psychophysics, 61, 1089–1101.

Braddick, O. J. (1974). A short range process in apparent motion. Vision Research, 14, 519–527.

(6)

Burr, D. C., & Santoro, L. (2001). Temporal integration of optic Xow, measured by contrast and coherence thresholds. Vision Research, 41, 1891–1899.

Cang, J., Kalatsky, V. A., Lowel, S., & Stryker, M. P. (2005). Optical imag-ing of the intrinsic signal as a measure of cortical plasticity in the mouse. Visual Neuroscience, 22, 685–691.

CliVord, C. W., & Ibbotson, M. R. (2002). Fundamental mechanisms of visual motion detection: models, cells and functions. Progress in Neuro-biology, 68, 409–437.

Eagle, R. A., & Rogers, B. J. (1996). Motion detection is limited by element density not spatial frequency. Vision Research, 36, 545–558.

Eagle, R. A., & Rogers, B. J. (1997). EVects of dot density, patch size and contrast on the upper spatial limit for direction discrimination in ran-dom-dot kinematograms. Vision Research, 37, 2091–2102.

Fried, S. I., Munch, T. A., & Werblin, F. S. (2002). Mechanisms and cir-cuitry underlying directional selectivity in the retina. Nature, 420(6914), 411–414.

Huxlin, K. R., & Pasternak, T. (2004). Training-induced recovery of visual motion perception after extrastriate cortical damage in the adult cat. Cerebral Cortex, 14, 81–90.

McGill, T. J., Douglas, R. M., Lund, R. D., & Prusky, G. T. (2004). Quanti-Wcation of spatial vision in the RCS rat. Investigative Ophthalmology and Vision Science, 45, 932–936.

Montero, V. M., & Jian, S. (1995). Induction of c-fos protein by patterned visual stimulation in central visual pathways of the rat. Brain Research, 690, 189–199.

Morgan, M. J. (1992). Spatial Wltering precedes motion detection. Nature, 355(6358), 344–346.

Morgan, M. J., & Ward, R. (1980). Conditions for motion Xow in dynamic visual noise. Vision Research, 20, 431–435.

Nakayama, K., & Tyler, C. W. (1981). Psychophysical isolation of move-ment sensitivity by removal of familiar position cues. Vision Research, 21, 427–433.

Newsome, W. T., & Paré, E. B. (1988). A selective impairment of motion perception following lesions of the middle temporal visual area (MT). Journal of Neuroscience, 8, 2201–2211.

Ohki, K., Chung, S., Ch’ng, Y. H., Kara, P., & Reid, R. C. (2005). Func-tional imaging with cellular resolution reveals precise micro-architec-ture in visual cortex. Namicro-architec-ture, 10, 597–603.

Prusky, G. T., West, P. W., & Douglas, R. M. (2000). Behavioral assess-ment of visual acuity in mice and rats. Vision Research, 40, 2201–2209. Prusky, G. T., & Douglas, R. M. (2004). Characterization of mouse spatial

vision. Vision Research, 44, 3411–3418.

Reichardt, W. (1961). Autocorrelation, a principle for the evaluation of sensory information by the central nervous system. In W. A. Rosenb-lith (Ed.), Sensory communication (pp. 303–317). Cambridge Massachu-setts: MIT Press.

Rudolph, K. K., & Pasternak, T. (1996). Lesions in cat lateral suprasylvian cor-tex aVect the perception of complex motion. Cerebral Corcor-tex, 6, 814–822. Todd, J. T., & Norman, J. F. (1995). The eVects of spatiotemporal

integra-tion on maximum displacement thresholds for the detecintegra-tion of coher-ent motion. Vision Research, 35, 2287–2302.

Watamaniuk, S. N., & Sekuler, R. (1992). Temporal and spatial integration in dynamic random-dot stimuli. Vision Research, 32, 2341–2347. Williams, D. W., & Sekuler, R. (1984). Coherent global motion percept

References

Related documents