Perceptual and Motor Skills, 1975,41, 55-59. @ Perceptual and Motor Skills 1975 LATENCY DIFFERENCES T O MONOCHROMATIC STIMULI MEASURED BY DISJUNCTIVE REACTION TIME PAUL D. JONES AND HILDE WILKINSON University of Louisville
Sirmmary.-Two studies were done in order to assess the effects of wavelength on visual perceptual latency as measured in a disjunctive RT paradigm. The results of the first study, though not statistically significant, suggested a trend toward shorcer RT to longer wavelength stimuli. In the second study, using well-practiced subjects, significant differences were found between disjunctive RT to red and green stimuli. The results suggest that latency differences as a function of wavelength are demonstrable in an experimental situation i n which the subject must react to chromatic information, as differentiated from brightness information. The question of whether differences exist in the latency of perception of various wavelengths is of considerable theoretical importance for theories of color vision. Physiological evidence seems to indicate that neural conduction time does vary as a function of wavelength. Lennox (1956), working with cats, compared latencies of response to colored lights at the lateral geniculate nucleus and at the striate cortex. Using four wavelengths, blue (445 n m ) , green (515 n m ) , yellow (578 nm) and red (620 n m ) , she showed that response latencies to all four wavelengths, when ERG amplitude was equated, differed at the cortex with red having the shortest latency, followed by yellow, blue, and green. Until very recently efforts to demonstrate a psychophysical relationship between wavelength and visual latency have failed. Guth ( 1964) used a dynamic displacement technique (Charpentier, 1893) to assess the effects of wavelength on perceptual latency at three luminance levels. The results of his experiment closely approximated the photopic luminosicy function and yielded the conclusion that response latency is independent of wavelength. Pollack (1968) utilized a simple reaction time paradigm in order to assess the effects of 6 different wavelengths at five levels of luminance. Her results showed no difference in reaction time as a function of wavelength at the four highest luminance levels. The marked difference in R T as a function of wavelength at the lowest luminance level in Pollack's study is attributable to rod intrusion. Finn and Lit (1971) also used simple R T to measure the effect of wavelength at low photopic and scotopic levels of illumination. Data were reported for two subjects for white, red, yellow, green and blue stimuli and for an additional subject for only white, red and blue. Analysis of the data for two of the subjects showed no systematic effect of wavelength. However, the third subject showed a tendency toward shorcer latencies to the red stimulus at rnesopic levels.
56
P. D.JONES 8r H. WILKINSON
Each of the above psychophysical studies has in common a methodology in which both wavelength and luminance are varied concurrently. PiCron (1931) reported that simple R T to red stimuli was shorter than to green or blue. The details of Piiron's heterochromatic brightness matching procedure are not provided nor is the level of illumination reported. It is clear, however, that his study used the method of stimulus substitution in which the level of illumination of the target remains unchanged while the hue is changed. Weingarten (1972) also reports positive results using the method of stimulus substitution. In his experiment the response measure was apparent movement (phi movement) in response to simultaneous or delayed onset of two photometrically matched chromatic stimuli. The discrepant results obtained with the various methods of assessing the effect of wavelength on perceptual latency seem to depend upon whether or not the subject is able to utilize brightness information in making his response. The use of the stimulus-substitution paradigm eliminates brightness cues; however, it is difficult to achieve in practice since transients often accompany the change of wavelength. A paradigm which is simpler to use, but in which the subject must make use of chromatic information in making his response, is disjunctive RT. The present study is an attempt to determine the effect of wavelength on perceptual latency using that paradigm.
EXPERIMENTI Method Subjects.-Four subjects, three males and one female (the second author), with normal color vision as determined by the Hardy, Rand, and Rittler plates (American Optical Co.) were tested. The subjects ranged in age from 20 to 56 yr. Apparatus.-The target stimulus was a 1'45' disc surrounded by a lo0 achromatic annulus which subject viewed in free view. Monochromatic stimuli were provided by a Bausch and Lomb grating monochromator with a 3-mm exit slit. Accurate selection of the desired wavelength was assured by adding a detente ring around the wavelength selection knob. The achromatic test stimulus and the surround were provided by 300-watt zirconium lamp (with correlated color temperature of 3200" K ) which was controlled by a regulated power supply. The chromatic and achromatic beams were brought to a focus in the plane of a stepping-motor-driven episcotister disc which was positioned so that one beam was interrupted while the other passed uninterrupted. Each beam transilluminated a flashed opal glass screen. The brightness of the surround field was controlled by Kodak Wratten neutral density filters, while that of the test beams was controlled by Kodak Inconel neutral density wedges which could be adjusted by subject for hetero-
LATENCY TO MONOCHROMATIC STIMULI
57
chromatic brightness matching. Viewing was through an achromatizing lens and a 3-mm artificial pupil. Head position was maintained with a chin-headrest. The stepping motors were controlled by multivibrators and one-shots. A white noise masking stimulus was produced by a GR noise generator. A l-sec. interruption of the masking noise served as a ready signal 2.5 sec. prior to the onset of the stimulus. A sub-threshold current was passed through subject in order to activate a Digibit drinkometer control when subject's finger was in contact with a metal disc. An Atec digital counter was started at the onset of each trial and stopped when the subject lifted his finger from the response plate. Latencies were automatically printed on a Beckman printer. The subject was dark-adapted 10 min. prior to each experimental session. Heterochromatic flicker photomecry was used to equate the luminance of the chromatic stimuli with the achromatic standard which was set at 1.0 ft-L. The achromatic surround was switched off during flicker photometry. During R T testing the subject was instructed to lift his finger when a chromatic stimulus appeared and to withhold his response when the stimulus was achromatic. Three monochromatic stimuli, with wavelengths of 440 nm, 510 nm, and 645 nm, were used as the stimuli. For each of these stimuli subject was given three practice trials with the positive stimulus followed by 12 positive and 12 negative stimuli in random order. The duration of a session was approximately 30 min. The order of presentation of stimuli was randomly varied and 5-min. rest periods were interspersed between wavelengths. Each of the three wavelengths was presented in three different experimental sessions. Results
Because RT tends to be skewed, median RTs were calculated for each of the three wavelengths on each of the three test days. The mean of these medians for each of the four subjects is presented in Table 1. Although the means are suggestive of the general trend reported by Lennox ( 1956), namely, that response to red is shorter than to blue or green, the trend is not uniform across all subjects and the mean differences are not statistically significant. The maximum number of responses to the negative stimulus in any single session was three. The results of this study do not support the hypothesis that perceptual latency varies as a function of wavelength, but the trend of the means suggests that such differences do exist. The decision process introduced by the disjunctive R T paradigm may have resulted in an increase in the response variability which overshadowed the small differences which might be expected in response latency.
P. D.JONES
58
~ 4 :H.WILKINSON
TABLE 1 MEANOF MEDIANRTs ( I N MSEC.) TO EACHOF THREEWAVELENGTHS Subject
Wavelength
440
510
645
The two subjects in the above experiment for whom the responses were least variable were pracciced observers in psychophysical experiments. The possibility exists that response variability in the disjunctive R T paradigm might be reduced through extended practice. Accordingly, a second experiment was designed, using two of the wavelengths of Exp. I, in which considerable practice was provided to each subject prior to testing.
EXPERIMENTI1 Method
Subjectr.-Two male subjects, including the first author, neither of whom had served in the first experiment, were tested in Exp. 11. Both subjects had normal color vision as determined by che Farnsworch D-15 test. Appa~atzls.-The apparatus was similar in all essential aspects to that used in Exp. I. Timing was provided by an Iconix programmable timing system and latencies were displayed on an Iconix digital counter and recorded manually. The achromatic stimulus, as well as the surround, was provided by a ribbon filament bulb operating at 6-v dc at 18 amps. A Kodak Wracten 79 filter was used to raise the color temperature of the achromatic source to 4800° K. Procedzire.-Both subjects were given five training sessions with each of the two monochromatic stimuli, 510 nm and 645 nm, used in the experiment. Training sessions consisted of 24 exposures of each positive stimulus and 24 exposures of a negative (achromatic) stimulus presented in random order. Prior to each experimental session the subject was dark-adapted for 10 min. Heterochromatic flicker photometry was used to equate each of the two test stimuli to the achromatic standard at 1.0 ft-L. Each subject was then given three practice trials with [he monochromatic stimulus prior to data being collected. An experimental session consisted of 24 positive stimuli and 24 negative stimuli in random order. After a 5-min. rest period the procedure was repeated for the remaining chromatic stimulus. All presentations were preceded by 2.5 sec. by an auditory warning signal. The entire session lasted approximately 45 min. Each subject was run for 12 days for a total of 288 determinations of disjunctive RT to each of the two monochromatic stimuli.
LATENCY T O MONOCHROMATIC STIMULI
Median R T for each of the two stimuli was calculated for each of the 12 sessions. Means of these medians are presented in Table 2. An analysis of the difference berween mean RTs to red and green stimuli for each of the two sub.05) jects showed that the differences are statistically significant ( t = 2.45, p for Subject PJ and ( r = 2.74, p < .O5) for Subject FS. N o false positive responses were observed for either subject during the experimental sessions.
Sirmmary.-Two studies were done in order to assess the effects of wavelength on visual perceptual latency as measured in a disjunctive RT paradigm. The results of the first study, though not statistically significant, suggested a trend toward shorcer RT to longer wavelength stimuli. In the second study, using well-practiced subjects, significant differences were found between disjunctive RT to red and green stimuli. The results suggest that latency differences as a function of wavelength are demonstrable in an experimental situation i n which the subject must react to chromatic information, as differentiated from brightness information. The question of whether differences exist in the latency of perception of various wavelengths is of considerable theoretical importance for theories of color vision. Physiological evidence seems to indicate that neural conduction time does vary as a function of wavelength. Lennox (1956), working with cats, compared latencies of response to colored lights at the lateral geniculate nucleus and at the striate cortex. Using four wavelengths, blue (445 n m ) , green (515 n m ) , yellow (578 nm) and red (620 n m ) , she showed that response latencies to all four wavelengths, when ERG amplitude was equated, differed at the cortex with red having the shortest latency, followed by yellow, blue, and green. Until very recently efforts to demonstrate a psychophysical relationship between wavelength and visual latency have failed. Guth ( 1964) used a dynamic displacement technique (Charpentier, 1893) to assess the effects of wavelength on perceptual latency at three luminance levels. The results of his experiment closely approximated the photopic luminosicy function and yielded the conclusion that response latency is independent of wavelength. Pollack (1968) utilized a simple reaction time paradigm in order to assess the effects of 6 different wavelengths at five levels of luminance. Her results showed no difference in reaction time as a function of wavelength at the four highest luminance levels. The marked difference in R T as a function of wavelength at the lowest luminance level in Pollack's study is attributable to rod intrusion. Finn and Lit (1971) also used simple R T to measure the effect of wavelength at low photopic and scotopic levels of illumination. Data were reported for two subjects for white, red, yellow, green and blue stimuli and for an additional subject for only white, red and blue. Analysis of the data for two of the subjects showed no systematic effect of wavelength. However, the third subject showed a tendency toward shorcer latencies to the red stimulus at rnesopic levels.
56
P. D.JONES 8r H. WILKINSON
Each of the above psychophysical studies has in common a methodology in which both wavelength and luminance are varied concurrently. PiCron (1931) reported that simple R T to red stimuli was shorter than to green or blue. The details of Piiron's heterochromatic brightness matching procedure are not provided nor is the level of illumination reported. It is clear, however, that his study used the method of stimulus substitution in which the level of illumination of the target remains unchanged while the hue is changed. Weingarten (1972) also reports positive results using the method of stimulus substitution. In his experiment the response measure was apparent movement (phi movement) in response to simultaneous or delayed onset of two photometrically matched chromatic stimuli. The discrepant results obtained with the various methods of assessing the effect of wavelength on perceptual latency seem to depend upon whether or not the subject is able to utilize brightness information in making his response. The use of the stimulus-substitution paradigm eliminates brightness cues; however, it is difficult to achieve in practice since transients often accompany the change of wavelength. A paradigm which is simpler to use, but in which the subject must make use of chromatic information in making his response, is disjunctive RT. The present study is an attempt to determine the effect of wavelength on perceptual latency using that paradigm.
EXPERIMENTI Method Subjects.-Four subjects, three males and one female (the second author), with normal color vision as determined by the Hardy, Rand, and Rittler plates (American Optical Co.) were tested. The subjects ranged in age from 20 to 56 yr. Apparatus.-The target stimulus was a 1'45' disc surrounded by a lo0 achromatic annulus which subject viewed in free view. Monochromatic stimuli were provided by a Bausch and Lomb grating monochromator with a 3-mm exit slit. Accurate selection of the desired wavelength was assured by adding a detente ring around the wavelength selection knob. The achromatic test stimulus and the surround were provided by 300-watt zirconium lamp (with correlated color temperature of 3200" K ) which was controlled by a regulated power supply. The chromatic and achromatic beams were brought to a focus in the plane of a stepping-motor-driven episcotister disc which was positioned so that one beam was interrupted while the other passed uninterrupted. Each beam transilluminated a flashed opal glass screen. The brightness of the surround field was controlled by Kodak Wratten neutral density filters, while that of the test beams was controlled by Kodak Inconel neutral density wedges which could be adjusted by subject for hetero-
LATENCY TO MONOCHROMATIC STIMULI
57
chromatic brightness matching. Viewing was through an achromatizing lens and a 3-mm artificial pupil. Head position was maintained with a chin-headrest. The stepping motors were controlled by multivibrators and one-shots. A white noise masking stimulus was produced by a GR noise generator. A l-sec. interruption of the masking noise served as a ready signal 2.5 sec. prior to the onset of the stimulus. A sub-threshold current was passed through subject in order to activate a Digibit drinkometer control when subject's finger was in contact with a metal disc. An Atec digital counter was started at the onset of each trial and stopped when the subject lifted his finger from the response plate. Latencies were automatically printed on a Beckman printer. The subject was dark-adapted 10 min. prior to each experimental session. Heterochromatic flicker photomecry was used to equate the luminance of the chromatic stimuli with the achromatic standard which was set at 1.0 ft-L. The achromatic surround was switched off during flicker photometry. During R T testing the subject was instructed to lift his finger when a chromatic stimulus appeared and to withhold his response when the stimulus was achromatic. Three monochromatic stimuli, with wavelengths of 440 nm, 510 nm, and 645 nm, were used as the stimuli. For each of these stimuli subject was given three practice trials with the positive stimulus followed by 12 positive and 12 negative stimuli in random order. The duration of a session was approximately 30 min. The order of presentation of stimuli was randomly varied and 5-min. rest periods were interspersed between wavelengths. Each of the three wavelengths was presented in three different experimental sessions. Results
Because RT tends to be skewed, median RTs were calculated for each of the three wavelengths on each of the three test days. The mean of these medians for each of the four subjects is presented in Table 1. Although the means are suggestive of the general trend reported by Lennox ( 1956), namely, that response to red is shorter than to blue or green, the trend is not uniform across all subjects and the mean differences are not statistically significant. The maximum number of responses to the negative stimulus in any single session was three. The results of this study do not support the hypothesis that perceptual latency varies as a function of wavelength, but the trend of the means suggests that such differences do exist. The decision process introduced by the disjunctive R T paradigm may have resulted in an increase in the response variability which overshadowed the small differences which might be expected in response latency.
P. D.JONES
58
~ 4 :H.WILKINSON
TABLE 1 MEANOF MEDIANRTs ( I N MSEC.) TO EACHOF THREEWAVELENGTHS Subject
Wavelength
440
510
645
The two subjects in the above experiment for whom the responses were least variable were pracciced observers in psychophysical experiments. The possibility exists that response variability in the disjunctive R T paradigm might be reduced through extended practice. Accordingly, a second experiment was designed, using two of the wavelengths of Exp. I, in which considerable practice was provided to each subject prior to testing.
EXPERIMENTI1 Method
Subjectr.-Two male subjects, including the first author, neither of whom had served in the first experiment, were tested in Exp. 11. Both subjects had normal color vision as determined by che Farnsworch D-15 test. Appa~atzls.-The apparatus was similar in all essential aspects to that used in Exp. I. Timing was provided by an Iconix programmable timing system and latencies were displayed on an Iconix digital counter and recorded manually. The achromatic stimulus, as well as the surround, was provided by a ribbon filament bulb operating at 6-v dc at 18 amps. A Kodak Wracten 79 filter was used to raise the color temperature of the achromatic source to 4800° K. Procedzire.-Both subjects were given five training sessions with each of the two monochromatic stimuli, 510 nm and 645 nm, used in the experiment. Training sessions consisted of 24 exposures of each positive stimulus and 24 exposures of a negative (achromatic) stimulus presented in random order. Prior to each experimental session the subject was dark-adapted for 10 min. Heterochromatic flicker photometry was used to equate each of the two test stimuli to the achromatic standard at 1.0 ft-L. Each subject was then given three practice trials with [he monochromatic stimulus prior to data being collected. An experimental session consisted of 24 positive stimuli and 24 negative stimuli in random order. After a 5-min. rest period the procedure was repeated for the remaining chromatic stimulus. All presentations were preceded by 2.5 sec. by an auditory warning signal. The entire session lasted approximately 45 min. Each subject was run for 12 days for a total of 288 determinations of disjunctive RT to each of the two monochromatic stimuli.
LATENCY T O MONOCHROMATIC STIMULI
Median R T for each of the two stimuli was calculated for each of the 12 sessions. Means of these medians are presented in Table 2. An analysis of the difference berween mean RTs to red and green stimuli for each of the two sub.05) jects showed that the differences are statistically significant ( t = 2.45, p for Subject PJ and ( r = 2.74, p < .O5) for Subject FS. N o false positive responses were observed for either subject during the experimental sessions.
