I l {I || 145 052 HTHS A METHQDCJLQGKAL ENVESTEGAHQN Q? ngcawga :1me RATE. Thesis £0? Hm Degree 0? M. A. WCHSJ’EN STATE UNIVERSITY Richard R; Knighf 1966 THESIS Michigan State University ABSTRACT A METHODOIOGICAL INVESTIGATION OF PERCEIVED FLICKER RATE by Richard R. Knight Under conditions of photic intermittency, the subjective flicker rate may, or may not, exactly correspond to the objective pulse rate. Two investigations concerning the measurement of the subjective rate are in distinct disagreement. It was the purpose of this study to attempt to resolve the previously conflicting results, and to apply a different method to the measurement of subjective flicker rate. Three observers matched visual and auditory rates 'equal’ or 'unequal,’ by the method of limits and the method of just noticeable differences, resPectively. The results obtained by the method of limits failed to conclusively replicate those pre~ viously found, i.e., for two observers the subjective flicker rate was lower than the objective pulse rate, but the third observer indicated that the flicker rate was higher than the pulse rate. Those results obtained by the second method showed the 'driving' phenomenon, as previously, but comparison of the two sets of results seemed to indicate that the 'driving' phenomenon was an artifact of the method employed. In addition, one observer attempted to apply two constant _n.- 0"" thugs; .- 1 '_0.. .- Richard R. Knight stimulus methods to the problem. It was found, however, that the observer was inadvertantly doing the easiest thing possible while attempting a difficult discrimination, and was responding only to the auditory stimulation, not judging its equality or inequality with the visual rate. Some suggestions for future research were made, and some possible explanations examined for the failure to replicate results. Approved: yJw/Ssafll—i Date: Mr ”n, will A METHODOIOGICAL INVESTIGATION OF PERCEIVED FLICKER RATE By Richard R. Knight A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1966 TABLE OF CONTENTS Page ACKNOWLEDGMENTS......... ....... . ............ .. ....... . ..... .. ii INTRODUCTION ...................... . .......................... 1 METHOD ....................................................... 9 RESULTS ...................................................... 15 DISCUSSION ....................................... ..... ....... 30 SUMMARY........ ...... ... ..... ...... .......................... 37 REFERENCES ................................................... 38 iii Table 10 LIST OF TABLES Page Length of interval (in cps) between upper and lower limits, as established by the methods of limits and just noticeable differences, for three observers....... ............................. 21 Difference and absolute limens, and standard deviations (SD), obtained by the methods of constant stimulus differences and constant stimuli (photic source 'on' and 'off'), as calculated by: A) Spearman arithmetic mean; B) Sum— mation method; C) Normal probability graph; D) Least squares, without weights ...................... 29 iv Figure 1. Pulse generator producing acoustic stimulation ...... 2. Waveform of acoustic stimulus ....................... 3. Points of subjective equality (PSE) for the three LIST OF FIGURES photic standard rates and three observers, as de- termined by the method of limits .................... Intervals of ’equal' judgments obtained by the method of just noticeable differences, for the three photic standard rates and three observers ..... Normal probability graph of percent of judg— ments 'auditory rate Faster then visual rate' (left) and 'auditory rate Fast' (right), obtained by the constant stimulus methods, for the 7 cps Standard stimulus ................................... Normal probability graph of percent of judg— ments 'auditory rate Faster than visual rate' (left) and 'auditory rate Fast' (right), ob— tained by the constant stimulus methods, for the 11 cps Standard stimulus ........................ Normal probability graph of percent of judg— ments 'auditory rate Faster than visual rate' (left) and 'auditory rate Fast’ (right), obtained by the constant stimulus methods, for the 15 cps Standard stimulus ................................... Page ... ll ... ll ... 18 ... 20 ... 23 ... 25 ... 27 INTRODUCTION Consideration of the relationship between subjective flicker rate and objective photic pulse rate, under conditions of inter— mittent photic stimulation, has in the past been of some interest. The assumption that the subjective and physical rates exactly correspond was investigated by Bartley (1938). Basing his investi— gation on previous physiological studies concerning the mechanism of the visual portion of the nervous system, Bartley observed: 1) that the last vestige of subjective flicker before critical flicker frequency (CFF) is reached is about the same rate regard— less of the pulse rate needed to reach CFF, and 2) that change in the stimulus intensity, with constant pulse rate, results in a change in the subjective flicker rate. V For the purpose of clarification, the photic pulse rate is defined as the rate produced by the electronic circuit driving the photic source, and is distinguished from the observer's per— ceived rate, the (subjective) flicker rate. In the same vein, the acoustic rate is distinguished from the auditory rate, the rate which the observer perceives. Reese (1943) attempted to construct an equal unit scale for visual rate. He used the fractionation procedure, asking subjects to set a comparison stimulus to % the rate of the standard stimulus. He found that as the standard stimulus rate increased, the "%" judgment became progressively less than % the standard. Further— more, he found a discontinuity in the judgments between 5 and 7 l 2 cps, for which he suggested that subjects were judging 'time between pulses' at the slower rates, and 'rate’ at rates above 7 cps. Reese did successfully construct an equal unit scale. Studies using limited trains of intermittent stimuli Taubman (1950, a, b) studied judgment of auditory and visual number. In the first investigation (1950, a), he used l—lO short clicks (moree dots) with varying interclick intervals. He found underestimation of the number of clicks, which underestimation increased with shorter intervals and higher number of clicks pre~ sented. The study in visual number (1950, b) obtained the same results. Taubman noted, as Reese had, that there was a change in the criterion of judgment at slower rates, in such a way that the subjects were judging time instead of rate. In a series of studies, Cheatham and White investigated temporal numerosity (Cheatham and White, 1952; White, Cheatham, and Armington, 1953; Cheatham and White, 1954; White and Cheatham, 1959). The first study (Cheatham and White, 1952) was concerned with perceived number as a function of pulse rate. Using trains of 5, 10, 15, and 20 pulses (corresponding to 10, 15, 22.5, and 30 cps), they found low estimates of number as none higher than 8 per second, though the subjects were presented with as many as 30 per second. They concluded that the number of flashes perceived is dependent on the number of pulses presented, and the total time. They suggested that the errors are not an underestimation in the judgment process, but due to a process in the visual system. 3 The second study (1953) was based on the above assumption. The authors questioned whether the above results were due to some action at the retina, or to some more central process. ERG recor— dings showed that the retina reacted to every pulse presented under the same conditions as in the previous study, and the authors con— cluded that the limitation of visual number was due to some higher process in the visual system. Cheatham and White (1954) used essentially the same procedure as in the previous studies, with 1—17 pulses of 1000 cps tone. They again found underestimation of perceived number, and found more variability in the judgments than in the visual tasks. The last study (1959) involved tactual stimulation. The same results obtained for judgment of tactual number as for auditory number, i.e., underestimation. The authors further found the pre— sence or absence of an illuminated surround affected judgments of visual number. An illuminated surround resulted in better judgments of number. Forsyth and Chapanis (1958) included retinal displacement as a variable in investigating perceived visual number. They found, essentially as Cheatham and White, that the underestimation of number increased with number presented (1-20 photic pulses), and also increased as the target was moved temporally on the retina. The above investigations give an indication of the disparity between input rate and perception of rate or number. However the generality of these results to conditions of extended trains of 4 intermittent stimulation is questionable. Nelson and Bartley (1965) and Nelson, Bartley, and Bochniak (1965) have indicated that conditions in the optic pathway differ markedly when fewer than 20 pulses are presented, and when extended trains are used. Specifically, a reorganization of the cortical response to the first 10-20 pulses takes place, when the rate is such that no single unit of the pathway can fire to each photic pulse, approximately 10—20 cps. Furthermore Nelson, Bartley, and Ranney (1960) have shown a sensory parallel to this cortical reorganization period. Nelson, Bartley, and Jordan (1963) attacked the problem of perceived rate in a different way. They investigated discrimination of the duration of intervals filled with photic stimuli. They found that conditions increasing amplitude and synchrony of optic pathway activity resulted in the longest duration discriminations. The authors mentioned that it was necessary to deter subjects from pro— ducing their own time standards, such as tapping or counting. Studies using_extended trains of stimuli Investigations involving extended trains of intermittent stimulation can be categorized as those involving photic or acoustic stimulation alone (primarily for establishing difference limens), and those involving both photic and acoustic stimuli. Gebhard, Mowbray, and Byham (1955) found difference limens (DLs) for photic stimulation, using the method of adjustment for standard rates 0 to 50 cps. Of their results those of interest are that they found DLs of .12 cps, .18 cps, and .25 cps for stan- dard rates 7, ll, 15 cps. When their results were compared with those obtained by Miller and Taylor (1948), who used a dif4 5 ferent method and acoustic stimulation, the relative difference limens (Af/f) were smaller for visual than auditory discrimination. This was contradictory to the common finding that the ear is a much better temporal discriminator than the eye. To resolve this conflict, Gebhard, Mowbray, and Byham (1956) found DLs for rate of acoustic stimulation, using the same method as used for photic stimulation, the method of adjustment. With the methods thus equated, they found DLs for acoustic rate to be smaller than those for photic rate, e.g., at 10 cps, Af=.080 cps, and at 20 cps, Af=.159 cps. The following are studies involving photic and acoustic inter— mittent stimulation in extended trains, and have bearing on the problem to be presented here. Pieron (1952, p. 308) relates a study by J. Segal (1939): "J. Segal attempted to determine with precision the evaluations of apparent frequency by comparison with a sound cadence (motor signalling being unable to follow beyond about 8 per second). He verified Bart— ley’s 1938 observations on the relation between brightness and perceived rhythm: when the critical fusion frequency is lowered (a brightness decrease facilitating fusion), the apparent rhythm is in— creased. On varying the brightness from 1 to 10, to 100, and to 1000 a stimulation at the rate of 20 per second is judged to be equal to the rhythm of about 40, 30, 20, 15 per second.... But the equilization of two disparate stimulations does not represent in the strict sense an evaluation, for it is hard to estimate-with any precision—~rhythms higher than 10 per second... Difference limens were matched for visual flicker matched to auditory flutter, and flutter matched to flicker, by Gebhard and Mowbray (1959). Both stimulus inputs were bilateral, and the method 6 of adjustment was used. The authors found that DLs for cross— sensory matches were considerably larger than those from intra— sensory matches, and that matching flutter to flicker was less accurate than flicker to flutter. They made the following obser— vations: 1) that "when the sound was under the control of 0, it appeared to ‘drive’ the light, i.e., a change in the physical rate of the flutter was immediately followed by a change in the pheno— menal rate of the flicker;" 2) that the flicker would not 'drive' the flutter, and 3) that the more variable and less accurate cross— sensory matching must be due to the difficulty of the task, since neither training no manipulation of variables reduced the errors. Another example of an interaction between auditory and visual rates was given by Ogilvie (1956). He found that CFF was raised when the acoustic stimulus was in phase with the photic stimulus; though the change in CFF was significant it was also small, aver— aging only .2 cps. The Controversy Nelson, Bartley, and Bochniak (1965), attempting to quantify flash frequency under varied brightness conditions, including the brightness enhancement phenomenon, used an acoustic tick with which to match the perceived flicker frequency. The photic target was a 10 5' disk, and the acoustic source was the tick emmitted by a General Radio Strobotac. The input was largely contralateral, right eye, left ear. The authors used a criterion such that "the sound and flash trains had to appear to accompany one another or one had to appear to 'cause' the other." They mentioned that 7 satisfying the criterion was difficulty, and that at higher rates, "ticks in the region of the flash rates seemed to depress the latter's strikingness." The method used was the method of average error. They found that flash frequency was consistently lower than photic rate, and that higher PCFs produced higher flash frequencies. Finally, they found that the flash rate was not a submultiple of the photic input rate, and that these results——underestimation of flash frequency——agree with those earlier obtained for discrimination of duration. Further investigation of the auditory 'driving' of the visual stimulus was undertaken by Shipley (1964). Shipley's procedure was as follows: the subject viewed a 250 visual target (right eye) and heard a tick from a General Radio Strobotac (right ear); the experimenter set the photic and acoustic rates equal to each other at a rate between 2 and 10 cps, and the O adjusted the acoustic rate up or down slowly until the apparent synchrony disappeared. He found that, with both stimuli set at 10 cps. the acoustic rate could be increased to as much as 22 cps for one 0, before the syn- chrony between the visual and auditory rates ceased. He described the experience thus: "the light and sound were initially pulsating strongly together; as the flutter frequency changed, the apparent flicker frequency followed right along with it until some point was reached where an apparent asynchrony intruded." Shipley concluded that auditory rate cannot be used to measure visual rate, because the 'driving' is so compelling. 8 The latter two investigations are at odds.- It should be noted that the methods employed were almost directly Opposed: Nelson, et. al., used a method of average error, in which each trial consisted of an ascending run and a descending run; the Variable (acoustic) stimulus is set unequal to the Standard (photic) stimulus rate, then increased or decreased until 0 re— ports that the two frequencies are just equal. Shipley used a method (method of just noticeable differences) in which both stimulus rates are set at objective equality, and 0 varies the acoustic rate until flutter no longer appears equal to flicker. Each method yields an interval on the acoustic frequency scale over which the two types of sensory response are judged equal. PROBLEM It is the purpose of this study, therefore, to attempt to resolve the conflicting results obtained by the above two methods, with the hope of showing that the results obtained are dependent on the method used; to determine that an acoustic source can in— deed be used to measure flicker rate, and to devise a new method of measuring the subjective rate. METHOD Apparatus The source of intermittent photic stimulation was a Sylvania R1131C Glow Modulator Tube, driven by a Lafayette Model l202D flicker—fusion apparatus. The flux from the Glow Modulator tube was passed through a planoconvex lens to enlarge the image, a Wratten neutral density (1.00) filter, and a 3.18 mm. aperture. The observer's head was held steady with a chin rest, and the target was viewed from 18.5 cm., resulting in a target subtense of 10. The PCP employed was .01, and the intensity was 233 c/ft2. All observations were carried out in a well darkened room. The acoustic stimulation was provided by a 2% x 2 inch oval speaker, driven by the circuit shown in Figure l. The circuit and Speaker were enclosed in a fiberboard box, 6m x 5" x 4", with a 1 inch diameter hole in front of the speaker. This apparatus produced ‘ticks' of approximately 20 msec. duration, in 2 over— lapping frequency ranges, from 3.57 to 20.41 cps. The frequency was varied by a decade potentiometer, with exponential characteris— tics. The acoustic apparatus was calibrated by leading the pulses driving the speaker into a Tektronix 502A dual beam oscilloscope. In this way the action of the speaker at any reading on the poten— tiometer scale was presented visually, and the duration of one cycle read in milliseconds. This msec. reading was then converted to cycles per second. 10 The fiberboard enclosure was located 30 cm. from Os left ear and the photic stimulus was viewed with the left eye. The ticks produced are shown in Figure 2. They were recorded with a Lexington dynamic microphone at the position occupied by OS ear, and the microphone output led into the oscilloscope and photo— graphed with a Polaroid Land camera. The intensity of the acoustic stimulus could not be measured by any intensity—measuring device, due to the transient nature of the pulse and the inability of the apparatus to generate a fused sound. The intensity was estimated as 80—90 db. Observers The 3 observers were all well—experienced in the task of judging the equality of 2 frequency inputs, though RRK and SSK more so than RWW. The observers were familiar with the rates used and the phenomena to be observed. Procedure A. Though Nelson, Bartley, and Bochniak used the method of average error, it was found to be too difficult for O to control the acoustic frequency with the apparatus used here, so the method of limits was utilized, with the experimenter varying the frequency. The observer was seated and dark adapted for 10 minutes. The Standard stimulus was the photic stimulus rate, and the Variable stimulus was the acoustic rate. Standard stimulus rates of 7 cps, ll cps, and 15 cps were used. These rates were randomly ordered such that 10 ascending and 10 descending matches were made at each rate. ll .JL 7/ _L\_L TT *GV Figure 1 Pulse generator producing acoustic stimulation A L L All A A v Trrr rrrv «- vvvvvvv —H»+-» FF} I» ~++H Figure 2 Waveform of acoustic stimulus ZMV IO MSEC 12 The Variable stimulus, acoustic rate, was controlled by the experimenter. He first found acoustic rates well above and well below each of the 3 Standard rates, at which 0 reported that the visual and auditory rates were definitely unequal, and at which 0 could determine which rate was the faster or slower. These acoustic rates were 4.43 and 10.58 cps for the 7 cps Standard; 6.55 and 16.06 cps for 11 cps; 9.17 and 18.38 cps for 15 cps, and were used as points around which to start ascending and descending trials. For any Standard rate, the ascending trial was run first, then the descending. The photic source was on continuously, and the experimenter gave a ready signal when the trail was to begin. He then turned on the acoustic stimulus with a silent switch, and either increased or decreased the frequency until 0 said "stop" at the first point at which the visual and auditory rates appeared equal. Such a session of 30 ascending trials and 30 descending trials lasted 1% hours. Each of the 3 Os followed the above procedure. B. This procedure was that used by Shipley, and is essentially Weber's method of just noticeable differences (Guilford, 1954). The observer was dark adapted as before. The Standard stimulus was again the photic rate, at 7, 11, and 15 cps, and the Variable stimulus was the acoustic frequency. Again, the photic source re— mained on continuously. The experimenter preset the acoustic frequency equal to the photic frequency, gave a ready signal, and turned on the acoustic source. He then turned the acoustic rate up or down until 0 re— l3 ported that the 2 perceived rates were no longer equal or synch— ronous. As in Procedure A, 10 inequality matches were made when the acoustic rate was increased, and 10 when the rate was decreased, at each of the 3 Standard rates. Each of the observers completed such a session, which lasted 1% hours, with a 15 minute break midway through the session. C. This procedure involved the use of the methods of constant stimulus differences, and constant stimuli. Only one 0, RRK, participated. After preliminary observation, a series of 7 acoustic fre— quencies was selected for each Standard stimulus (again, photic rate). These acoustic frequencies were as follows: for the 7 cps Standard, 6.85, 7.00, 7.15, 7.30, 7.45, 7.60, 7.75 cps; for the 11 cps Standard, 10.8, 11.0, 11.2, 11.4, 11.6, 11.8, 12.0 cps; for the 15 cps Standard, 14.4, 14.6, 14.8, 15.0, 15.2, 15.4, 15.6 cps. Each of the above acoustic rates was paired with its respective Standard rate a total of 60 items; for each Standard rate, a total of 420 judgments were involved. As before, the photic stimulus remained on continuously, but did not change rate, i.e., all 420 trials were completed at 7 cps before doing those for 11 cps, and so for 15 cps. The observer was presented with one of the 7 acoustic stimulus rates paired with the Standard stimulus, for a 5 second period as controlled by a Hunter interval timer. The acoustic frequencies 14 within each of the 3 sets were randomly ordered such that no fre— quency could follow itself more than once, i.e., the frequencies were randomized in blocks of 7. The observer was required to re— port whether the tick rate was "Faster" or "Slower" than the flick— er rate, before the 5 seconds had elapsed. The preliminary obser— vations had shown that there was some tendency to judge the equality of the 2 rates by comparing the flicker rate before, during, and after the acoustic rate was introduced. This was thought to com- plicate the judgment, and was eliminated by the forced response. The observations were made 60 trials at a time, occupying 45 minutes. A total of 21 such sessions were completed. In order to determine that the observer was responding to both sets of stimuli, the above procedure was repeated, with the fol— lowing change. The photic source was turned off, all other con— ditions remained constant, and O was presented with the tick fre- quencies alone. He was required to scale the 7 frequencies into "Fast" and "Slow." As before, 60 judgments were made of each fre— quency. The 420 judgments were made in 1% hours. In each of procedures A, B, and C, above, 0 was strongly dis— couraged from counting or tapping, or in any way self-producing a time standard. Furthermore, the conditions of photic stimulation (short PCF, moderate intensity) and acoustic stimulation (short pulse, moderate —— strong intensity) were set with the purpose of making it as easy as possible for O to discern flicker and auditory 'flutter', and so to increase the accuracy and ease of the judgments. RESULTS A. The results obtained using the method of limits are presented in Figure 3. The 450 line represents physical equality between the 2 input rates. Both the upper and lower limits, and the midpoint of the interval so defined (the point of subjective equality——PSE), are shown at each of the 3 Standard frequencies. The upper and lower limits as graphed do not represent a standard deviation or a confidence interval, but rather the average acoustic frequencies at which 0 indicated that the visual and auditory rates were just equal, established by descending and ascending trials respectively. Each limit is the mean of 10 observations. It can be seen that 2 Os, RRK and RWW, set the acoustic rate lower than the photic rate, though these differences are not large. The other 0, SSK, consistently matched the perceived rates at acoustic rates higher than the photic rate. Also, at the 15 Cps photic Standard, RRK made 6 of 10 descending judgments at acoustic rates lower than the matches on the ascending trials, thus probably accounting for the extremely small size of the equality interval. It should also be noted that none of the equality matches were made at a submultiple of the photic Standard frequency, e.g., a match at 10 cps for a Standard frequency of 15 cps. B. The results of the method of just noticeable differences, used by Shipley, are_presented in Figure 4. As in Figure 3, intervals are shown, bounded by the average limits, upper and lower. The 15 16 average limit is the acoustic frequency at which 0 indicated that the 2 perceived rates were no longer equal or synchronous. Each limit is the mean of 10 observations. In Figure 4, the amount of 'driving' is indicated by the length of the interval up and down from the physical equality point to each limit. The upward trials for RWW at 15 cps obtained no judgment of inequality at the upper limit of the acoustic gener— ator, so no definite point can be shown. Table 1 compares the length of the interval between upper and lower limits for the 2 methods and 3 Os. Only those intervals at 11 cps and 15 cps for RRK and RWW were larger under Shipley's con— ditions than when found by the method of limits. C. The results obtained using the 2 constant methods are shown in Figures 5, 6, and 7. The figures shown percent "tick Faster than light" judgments and "tick Fast" judgments (no photic stimulation), plotted on a probability scale, versus acoustic frequency. Plotted in this way, a normal distribution of percent scores would fall on a straight line. It can be seen that as the Standard rate increased, distri— butions were harder to fit to straight line. No manipulation of interval size or placement of frequencies around the Standard was able to improve this data. The straight lines are fitted by the averaged Z—scores method (Woodworth and Schlosberg, 1954, p. 205). 17 Figure 3 Points of subjective equality (PSE) for the three photic standard rates and three observers, as determined by the method of limits 18 .m. __ .h p mac £53 2.8:... .m. m magmas __ P K. .m— h >> 22.m v_m_~. v.mwmw rwmwnu SdO‘BlVH OLISOOOV 19 Figure 4 Intervals of 'equal' judgments obtained by the method of just noticeable differences, for the three photic standard rates and three observers 20 m. rt‘ 4 5.53.1 mac .53. 2.8:“. .... __ s D h 33¢ xmm xmm .mmo Is. ,1 — - I In ‘aivu ousnoov 5d!) 21 TABLE 1 Length of interval (in cps) between upper and lower limits, established by the methods of limits and just noticeable differences, for three observers standard Rate method of limits 7 cps method of just noticeable differences 11 cps 15 cps Observer SSK RRK RWW 4.37 2.70 2.95 1.62 2.05 2.03 6.38 2.83 3.19 3.16 3.99* 4.28* 5.41 0.03 1.81 3.66 4.98* 10.61* * indicates interval larger under conditions of 'driving' than under method of limits 22 Figure 5 Normal probability graph of percent of judgments ’auditory rate Faster than visual rate' (left) and 'auditory rate Fast' (right), obtained by the constant stimulus methods, for the 7 cps Standard stimulus was; mao .mEozuaouE 0.53004 .mbN Own. ch OnN 9N CON no.0 th OmN m¢N OMN m3. CON no.0 b b n P - P n b P b p - _IP :1. 4. _ . _ _ _ _ _ _ _ _ _ _ .. .1. .IIIIL. lllll T '|'|llll| 'I'lll'l|"l lllll|l|'|'l 23 IIIIIIIIIIIIIIII 5m ,ISVd mu, iNEIoaad __________1._________ I I I I I I I l I I l .1 lllll'll'll 'll'illllll om