057-070 a study of individual differences in motion acuity at scotopic levels of illumination

7/28/2019 057-070 a Study of Individual Differences in Motion Acuity at Scotopic Levels of Illumination.

1/14

A STUDY OF IN D IV ID U A L D IF F E RE N CE S IN MOTIONACUITY AT SCOTOPIC LEVELS OFILLUMINATION *

BY C. J. WARDEN, H. C. BROWN AND SHERMAN ROSSColumbia University

INTRODUCTION

Little is known regarding motion acuity at parafoveal and pe-ripheral retinal positions under low levels of illumination. The samemay be said of the relationship between form and motion acuity atboth photopic and scotopic levels of illumination. The purpose ofthe present investigation was to study certain aspects of these rela-tionships.This paper is a continuation of the work recently reported bearingon the relationship between form and motion acuity (2). The earlierreport was concerned mainly with the standardization of apparatusand method for use in the present investigation. The emphasisthroughout the present study is upon the factor of individual dif-ferences.From the standpoint of military psychology, the primary interestwas in the selection of night-fliers. The question had been raisedas to whether or .not the Snellen Test is adequate for this purpose.If not, the problem of developing a test of scotopic acu ity would needto be solved. Moreover, it seemed to the present authors thatmotion acuity rather than form acuity ought to be stressed in thefield of aviation. For motion acuity is obviously the most importantvisual function utilized in flying. In fact, the logic of the situationwould suggest the use of a scotopic motion acuity test as the mostappropriate screening device.The specific attack upon the problem was made along the follow-ing lines in the present study. First of all, a group of Ss was selectedon the basis of Snellen scores. These 5s were then dark-adaptedfor an appropriate period, and given the Motion Acuity Test at twolow levels of illum ination. Threshold determinations were madethrough a series of six retinal positions, ranging from 7 degrees to 55degrees of peripheral angle.

* The work described in this paper was done under a contract recommended by the Com-mittee on Medical Research, between the Office of Scientific Research and Development andColumbia U niversity.1 The senior author was Responsible Investigator, and the junior authors were ProfessionalScientific Assistan ts, on this project (OEM cmr-264). 57


2/14

58 C. J. WARDEN, H. C. BROWN AND SHERMAN ROSSThis experimental design enabled us to make such comparisonsas the following: (i) the range of Snellen Scores and of the motionacuity threshold values of the group, (2) the correlation between

Snellen Scores and motion acuity threshold values, (3) the com-parison of the motion acuity threshold values for the two levels ofillumination tested, and (4) the application of these findings tohypothetical combat conditions.METHOD AND PROCEDURE

In this section, a brief description of the Ss, the apparatus, and the test procedure is pre-sented.Snellen test.This tes t was administered under standard conditions of distance and illumina-tion to each S prior to acceptance. A group of 28 male students a t Columbia University wereselected by this means. They were between the ages of 17 and 26 years, and had a Snellen Indexof 20/20 or better in both eyes, without glasses, and reported no history of gross visual defect.Most of the men had already passed the visual requirements for the Army or Navy Officer'sReserve. They were retested, however, before being accepted as Ss . Th e actual Snellen Scoresof those accepted were recorded for use in computing correlations between photopic form acuityand scotopic motion acuity. The test of motion acuity required three sessions of about twohours each in the dark-room. The Ss were paid a nominal sum at the close of the experiment,and were well motivated throughout.Mo tion acuity apparatus.The apparatus has been described in detail in a previous paper (2).It consisted essentially of the following parts: (a) A motor-gear-pulley system for rotating thestimulus. This system could be manipulated by the S in securing a threshold speed (method ofadjustment), (b) A dark cubicle for the S, lighted from above by a projector set for a given levelof illumination . A relatively high (log. 3.4 ml.) and a relatively low (log. 4.0 ml.) scotopiclevel of illumination was used. The latte r is presumably abou t the lowest level involved inordinary night-flying, (c) A perim etry set-up for securing the desired peripheral angle. A seriesof six retinal positions was employed: 7, 10, 25, 35, 45, and 55 degrees from the fovea. (d) Aseries of stimuli, in various sizes, as described in detail in Table I . Th e stimulus figu re consistedTABLE I

DESCRIPTION OF STANDARD STIMULUS SETPeripheral Angle atWhich Stimulus was Used

7 , 1 025354555

Circumference inMillimeters160.3188.622O.O285.I349-0

Arc of Gap inMillimeters8.910.513.2I5 .819.4

Arc of Gap in Degreesof Visual AngleI.041.22I.43I.8s2.25

The manner of determining the size of this series of stimuli has been discussed in an earlierpaper (2). They were large enough for the Ss with poorer motion acuity, but larger than neces-sary for the better Ss.of a Landolt ring modified for rotary motion. Instead of a single gap , there were nine parallel-sided breaks or gaps (white areas), equidistan t from each other. A white bristol board back-ground, ben t in the proper arc for perimetry, served as the surround. The stimulus was exposedby means of a white bristol board shutter operated manually.As will be noted, the stimulus involved a circular figure of black wedges against a whitesurround and general background. A t one stage in the experiment, these brightness relation-ships were exactly reversed to see whether or not suchja change would make any difference inmotion perception. Since this stimulus-background change was not effective in raising or lower-


3/14

INDIVIDUAL DIFFER ENC ES IN MOTION ACUITY 59ing the motion threshold, no further mention of this variation from the standard procedure willbe made.Procedure.Motion threshold determinations were made on the right eye only, the leftbeing covered with an Adjusto eye-shield. This device perm itted the eye to remain open, toblink and wink, but prevented binocular accommodation on the fixation point.Each S usually required three sessions of about three hours each, given in the following order :(a) training session, (i) experimental session at the log. 4.0 ml. level, and (c) experimentalsession at the log. 3.4 ml. level. Each session was preceded by a dark-adapta tion period of 25min. (Navy red dark adaptor goggles), followed by a 5-min. period within the cubicle, in orderto reach the precise testing level.The Ss were all untrained in this type of observation, hence the first session was devotedlargely to training and dem onstrations. Each S was given information concerning the anatomyof the eye, the specific functions of the rods and cones, the meaning of the curve of dark adapta-tion, etc. He was then instructed how to fixate, how to operate the rheostat controlling thespeed of the stimulus, how to relax between judgm ents, etc. He was also trained in the use ofthe criterion of threshold motion, and was required to make sample judgments. During thissession, he became adjusted to a three-second exposure time. The Ss were usually proficientin making threshold judgments by the end of the first session.In the two test sessions, the method of adjustment was applied as follows in determining thethreshold rate of movement: (1) the E set the general speed level far above the threshold speed,(2) the S, using the rheostat control, then decreased the speed in large steps to a rate below thethreshold, (3) the S then increased the speed in small steps up to the threshold level, in terms ofour criterion.A high criterion was used here in order to avoid doubtful judgments and inconsistencies.At the threshold level, the Ss were required to see definite motion with absolute certa inty. If themotion was so slow as to appear doubtful, or to involve inference or suggestion, it was not re-garded as criterial. In such a case the S was required to establish the threshold value a t a some-what higher speed. The standard requirement was tha t the S be able to make a judgment ofgood motion in two out of three exposures at th e same speed. Three determinations were madefor each retinal position a t both levels of illumination. The threshold values determined by theuse of this criterion were probably somewhat above the absolute threshold.Th e exposure time was controlled by the shu tter. When the S indicated that he w as ' ready,'th e E counted aloud ' 1-2-3 ' a n d raised the shutter. At the end of three seconds, the shutter wasclosed. A three-second exposure time had been found to be adequate in preliminary experi-ments (2).The threshold speed of rotation of the stimulus was read from the chronoscope, in terms ofseconds, for a given angle on the timing disc. These values were then transmuted into degreesof visual angle per second, as the most convenient index of motion acuity.

RESULTSThe results secured on the 28 Ss for both levels of illuminationare given in Tables II and III, and the data are plotted in Figs. 1and 2. The threshold values for motion acuity are given in term sof degree of visual angle per second. These measures represen t thethreshold rates of rotation of the stimulus at the various retinalpositions, under each level of illumination.The general form of the function is essentially the same for bothlevels of illum ination. However, the curves for th e higher level(log. 3.4 ml.) lie below the corresponding curves for the lowerlevel (log. 4.0 m l.). Th is relationship was found to be consist-ently true for all Ss, as might have been expected.In general the curves show a rise in trend as the peripheral angleis increased. Th is trend is very definite from th e 10 degree retina l


4/14

ouUlV)o rua

u_ioz3-

oVIklU

Q:O

F l C -PERIPHERAL ANGtC

' Log- - 3 - 4 L level of i llumination.


5/14

INDIVIDUAL DIFFERENCES IN MOTION ACUITY 61

7 10 24 35PERIPHERAL. ANCLE

FIG . 2. Log . 4.0 m l. level of illum ination.


6/14

62 C. J. WARDEN, H. C. BROWN AND SHERMAN ROSSposition outward. The results for the 7 degree position are somewhatirregular. This was due to the fact that the rather large stimulusused here was very close to the fixation point, and overlapped thenear-foveal region. The data at this point may be ignored, however,since the 10 degree position is commonly regarded as the mostsensitive (1).When the threshold values are plotted on semi-log paper (Figs.3 and 4), a straigh t line relationship ob tains. This is especially clear

TABLE IIR E S U L T S FOR THE H I G H E R S C OT O P IC L E V E L O F I L L U M I N A T I O N (LOG

D E G R E E S O F V I S U A L A N G L E

Subject

I234567891 0

1 11 2131415161718192 02 12 2232 4252 62 728

7.608351293234.257386.269328.281374351.328.199257.222.164.176.152.087.140.328.049O34.071.026.316445398

10 .679.562.222234.316.386-257304.316398304.328.199234.199.176.187.187.071.158304.049043.062.020.491.491.421

Retinal

2 5 I.030585.269.328386.410433.410374.410398.386234.281.187.257.211.187075.176.246.090.063.062.047597573573

PER SECOND

Position

35 I.205.714374.421.480515550.480.491503.491.468293.363.222293.222234.089.269.281.117.076.074.061.714.690655

- 3 . 4 ML.)

45I.427

.889.456.620632597.679.585.667.6325855733Si.410.269339.246.328.103319339.152.095.089.082.866.819.842

IN T E R M S OF

55 1-7551.041.620.807749.854.866.702.819.725.679.725.421.46835i386.328.410.129.391374.2113.112.105

1.0651.0061.030

from the 25 degree retinal position outward. The curves, in general,take the form of an exponential function, such as y = kax.The range of individual differences in motion acuity thresholdvalues is extremely m arked . These values for the 10 degree retinalposition at the lower brightness level (log. 4.0 ml.) are plotted inFig. 5. As will be seen, the histogram takes the general form of anormal distribution with some tendency toward positive skewness.It seems likely that with a sufficiently large number of cases a fairly


7/14

INDIVIDUAL DIFFERENCES IN MOTION ACUITY 63typical normal curve would be obtained . This would mean th a t agroup, highly selected on a cone level form test (Snellen), would bepractically unselected so far as scotopic motion acuity is concerned.In brief, the fact of normal (20/20 or better) photopic form visionoffers no basis for predicting high motion acuity at low levels ofillumination.This conclusion is further substantiated by a correlation betweenSnellen Index and motion acuity threshold va lue . The biserial cor-

TABLE IIIR E S U L T S F O R T H E L O W E R S C O TO P IC L E V E L O F I L L U M I N A T I O N ( L O G . 4.0 M L . ) I N T E R M S O F

D E G R E E S O F V I S U A L A N G L E P E R S E C O N D

Subject

I234567891 0

1 11 213H15161718192 02 12 22 32 4252 62 72 8

7983374515.468.46857335i.421.49136337435i304.491.211.164304.199.117.199339059.046.140.047538433433

10.971655.48035i363538.316.398.328.410.328339304351.211234257.222.187.176339.077.082.117.023.49153.456

Retinal Position2 5

I.2I7.819515538.620.608542.456445445433.421.386363.269304.281304.222234.269 152.129. 1 1 1.056644.661.610

351.4041.018.632655772.761.644.56255o538550503445445363.363.3283O4

.281.281293.199.146.129.070.784.778714

45 I.7201.229.866.842936.924.819.714.714.679.679.644573538445445.421398

35 '.328363.246.181.164.094.983959.866

552.2231.5211.322119 31.1581.13S1.030.913.878.878.866.807 7 H.667.562538-SIS515

433433.410.316.228.222.117I-I931.1701.077

relation technique was used. Th is involved arranging the individualsinto two Snellen categories (20/20, 20/15 or better), and ranking themean motion acuity threshold values for all six retinal positions at agiven level of illumination. Th e biserial correlation for the higherlevel of illumination (log. 3.4 ml.) was found to be only .07 .10.This means that an individual with a high Snellen Index (20/15 o rbetter) is not likely to have any better scotopic motion acuity thanan individual with an average Snellen Index (20/20).


8/14

64 C I. W A R D E N ) H . C. BROWN AND S H E R M A N R 0 S S

.01 7 fO 2 5 5 55 45PERIPHERAL ANGLEFIG. 3. Semi-log curves (log. 3.4 ml.).

APPLICATION TO NIGHT-FLYING CONDITIONSIn the present section an attempt is made to apply our results tothe night-flying situation. The question of major interest is as towhether or not the individual differences in motion acuity herereported are large enough to influence efficiency in night-flying.


9/14

INDIVIDUAL DIFFERENCES IN MOTION ACUITY 65

3 r -

Ozvi

u.OuUiccouo6o,

. 0 5

.Oi l 1 I 1 1 I7 10 2 5 3 5 4 5P E R I P H E R A L A N G L E

FIG. 4. Semi-log curves (log. 4.0 m l.).

55


10/14

66 C. J. WARDEN, H. C. BROWN AND SHERMAN ROSSAre some of our Ss so insensitive to moving objects at low illumina-tions that they would not make good night-fliers ?We found it impossible to obtain precise information on theactu al flying situa tion . It was necessary, therefore, for us to baseour com putations upon a set of hypothe tical conditions, which wouldseem to agree fairly well with the actual flying situa tion . In thecom putations reported, the specific conditions assumed are as follows:(i) size of plane: 90 feet in the longest dimension (wingspread), (2)distance of this object from the S: 1.5 or 2.0 miles, and (3) level ofillumination: scotopic (log. 4.0 ml.).

10N

0 2 3 " 8 213 . 3 0 6 4 0 3 4 9 8 593 .688 .763 .878 .973D E G R E E S O f V I S U A L A N G L E P E R S E C C A T E G O R I E S

FIG. 5- Frequency distribution of motion acuity threshold values at the 10 degree retinalposition (log. 4.0 ml. level of illumination).There is reason to believe that the distances indicated aboverepresent about the limit of form visibility for a plane of this sizeunder th is low level of illumination . These values (1.5 and 2.0 miles)correspond closely with the projection of the form threshold value,determined on four Ss, at this level of illumination. These formthresholds represent a reduction in visibility to about 1/20 of the


11/14

INDIVIDUAL DIFFERENCES IN MOTION ACUITY 67normal distance perception in daylight, as might be expected. Asa matter of fact, the stimuli used in determining the motion acuitythresholds were somewhat larger than the stimuli used to secure theform acuity thresholds for projection. Itseems reasonable toassume, therefore, th a t a plane of this size would be definitely visibleat 1.5 or 2.0 miles.Individual thresholds in motion acuity were computed to agreewith the specific setting indicated above. The threshold values atthe 10 degree retinal position (log. 4.0 ml.) were taken as the basis

TABLE IVI N D I V I D U A L D I F F E R E N C E S IN M I N I M U M S P E E D N E C E S S A R Y FOR THE O BSE RV A T I O N OF A

P L A N E (90 F E E T W I N G S P R E A D ) AS M O V I N G

SubjectsRankOrder

I234567891 0

I I1 2131415161718192O2 12 2232425262728

ThfPOhmnInTaftPUkQ\tX IIItroQOlU Q J-st?4iCC9 UlVisual Angle/Sec.

.023077.082.117.176.187.211.222234257.304.316.328.328339339 3 5 1351363398.410456.480.491503538.655.971

Speed in Miles per Hour at Which

Plane 1.5 Mi.2.157.217.68

IO.9616.4917-5219.7620.7921.9224.0728.4829.603O.7230.7231-7531-7532.8832.8834.OO3 7 7 338.4042.7144.9645-9947.1250-396i-3S90.95

Motion is First SeenDistant Plane 2 Mi. Distant

2.869-5910.2114.5821.9323.3026.2827.6529.1532.OI37-8839-3740.8640.8642.2342.234373437345.2249.7851-0756.8059.8061.1762.6767.0281.60

120.96Note. Ss are ranked in order of threshold motion perception at 10 peripheral angle at levellog. 4.0 ml. The speed indicated in the last two columns refers to the difference in speed(resultant) between the observer 's plane and the observed plane.

of computation, since this retinal position is generally regarded asthe most sensitive at low illumination. The results of this analysisare presented in Table IV. The threshold values obtained in ourexperiment, arranged in rank order, are given in column 2 of the


12/14

6 8 C. J. WAR DEN, H. C. BROWN AND SHERMAN ROSStable . The values in the last two columns of the table represent thethreshold rate in miles per hour for the perception of motion of a90-foot plane at th e distances indicated. These speed values referto the resultant speed (difference) between the plane of the observerand the plane to be observed.T he meaning of the speed values in the table can best be indicatedby a few examples. The bes t S (No. 1) should be able to detect aplane as moving when the resultant speed is only 2.15 miles per hourat a distance of 1.5 miles. The rate for a distance of two miles wouldneed to be 2.86 miles per hour. On the other hand, the worst S(No. 28) could not detect the plane as moving, unless the resultantspeed were 90.95 miles per hour at 1.5 miles, or 120.96 miles per hourat a distance of 2.0 miles. The individual differences in the rate ofmovement of the object, for the threshold perception of motionunder these conditions, are thus seen to be extremely large.In order to determine whether or not our poorer Ss are tooinsensitive to moving objects to make good night-fliers, the criticalresulta nt speeds for actua l com bat should be known. We have beenunab le to secure definite information on this poin t. Th e demandsfor the perception of movement at different rates would vary fromone type of plane to another, and from one combat situation toanother.However, it is possible to demonstrate how the information inTable IV might be applied to the selection of night-fliers. Th is canbe done by assuming arbitrarily that certain resultant speeds arecritical. For example, if the resulta nt speed of 50 miles per hour iscritical, Ss 26, 27, and 28 would be unab le to see the plane as moving,at a distance of 1.5 miles. A t a distance of two miles, Ss N o. 21through 28 would not be able to observe it as moving. If we assumea resultant speed of 75 miles per hour to be critical, then S 28 wouldbe rejected at the 1.5 mile distance, and Ss 27 and 28 at the 2.0 miledistance. Even a t this high critical speed, 3.6 percent of th e Sswould be eliminated at the shorter distance, and 7.2 percent at thelonger distance.

A SHORT TEST FO R NIGHT-FLIERSAlthough the present experiment involved the use of complicatedapparatus, a much-simplified test for practical use in the selectionof night-fliers could be readily developed. Such a standard testwould require only a short time per S, and would be easy to ad-minister. On the basis of the findings of the present experim ent, atest for selection could be simplified along the following lines:I. The apparatus for rotating the stimulus would be small, com-


13/14

INDIVIDUAL DIFFERENCES IN MOTION ACUITY 6 9pact and portable. By adapting the method used in the N avyAdaptometer (3) to motion acuity, only a single speed would berequired. Th is would be the median threshold speed, as determinedon a group of 100 or more Ss. Individual ra tings would be basedupon the percentage of correct responses to this median speed on 50or more trials.2. It would be restricted to a single rod level of illumination.This is possible since a high correlation (.94) was found betweenthreshold values at the two low brightness levels used in the presentexperiment.3. Binocular vision would be used, with a fixation point approxi-m ately 10 degrees above the stimulus (3). The determinations madeat a single retinal position would serve as a dependable index ofscotopic motion acu ity. This is tru e since the motion acu ity curvesobtained in the present study were regular out to 55 degrees.4. It is estimated that the time required to give a screeningtest with such an apparatus would be about 10 min. per S, in additionto the time required for dark adaptation.

SUMMARYA group of 28 male Ss, between 17 and 26 years of age, with aSnellen Index of 20/20 or better were tested for motion acuity.Threshold values were secured under two levels of illumination (log.

3.4 ml. and log. 4.0 ml.), at six retinal positions ranging from7 degrees to 55 degrees.The following facts and conclusions emerge from the experiment:1. The motion acuity function for each level of illumination wasfound to be regular, with a gradual rise as the peripheral angleincreased. As might be expected, the curves for the higher level(log. 3.4 ml.) were much lower than the curves for the lower level(log. 4.0 ml.). The rank order correlation between the averagemotion acu ity threshold values for th e two levels is .94 .18.2. The range of individual differences in motion acuity thresholdvalues is extremely marked, and shows a steady increase from the 7

degree retinal position ou tward . A t the 10 degree retinal position(log. 4.0 ml. level) the range in threshold values is from .023 to.971 degrees of visual angle per second (Table II I ) . Th e thresholdvalue of the worst S is thu s over 40 times as large as th a t of th e bestS. An at tem pt to apply these results to night-flying conditions isdiscussed in a separate section.3. No significant relationship was found between the SnellenIndex and the motion acu ity threshold value. A biserial correlationwas computed between two Snellen categories (20/20 and 20/15 or


14/14

70 C. J. WARDEN, H. C. BROWN AND SHERMAN ROSSbe t t e r ) and the motion acui ty threshold va lues at the log. 3.4 ml.level of i l lum inatio n. T his biserial correlation was only .07 .10.A Snellen Index of 20/20 is no gua r a n t e e t ha t an aviator possessesadequate mot ion acui ty at scotopic levels of i l luminat ion.4. A plan for a shor t and simplified test of scotopic motion acui ty,to be used in the selection of night-fliers, is outl ined in the final sectionof this report .

(Manuscript received May 9, 1944)BlBMOGRAPHY

1. LILJENCRANTZ, E., SWANSON, C. A., & CARSON, L. D. The use of the eyes at night. Proc.U. S. NOD. Insi., 1942, 68, 802-810.2. WARDEN, C. J., & BROWN, H. C. A preliminary investigation of form and motion acuity atlow levels of illumination. / . exp. Psychol., 1944, 34, 437-449.3. WEBSTER, A. P. The measurement of thresholds of perception and discrimination. 1943.Pp. 18. (Unpublished)

057-070 a study of individual differences in motion acuity at scotopic levels of illumination

Documents