Brain Research, 100 (1975) 111-120 © ElsevierScientificPublishingCompany,Amsterdam- Printed in The Netherlands
111
REACTION TIME TO SQUARE-WAVE GRATINGS AS A FUNCTION OF SPATIAL FREQUENCY, COMPLEXITY AND CONTRAST
A N T O N I O TARTAGL1ONE, DAVID P. G O F F AND A R T H U R L. BENTON
Neurosensory Center and Department of Neurology, University of lowa, lowa City, Iowa 52242 (U.S.A.) and Clinica delle Malattie Nervose e Mentali dell'Universitgt, 16132 Genoa (Italy) (Accepted June 7th, 1975)
SUMMARY
Visuomotor reaction time (RT) of normal human subjects to briefly presented square-wave bar patterns were measured. The pattern varied with respect to magnitude and number of spatial frequencies (one cycle/degree of visual angle, 3 cycles/ degree, and both combined) with each spatial frequency stimulus presented at 3 levels of contrast. Space-average luminance was held constant. RT increased with the higher spatial frequency and was intermediate in value for the combination stimulus at each level of contrast. RT increased with decreasing contrast levels for each spatial frequency stimulus. There was no interaction between the two parameters. The findings indicate that a behavioral response measure can be employed to study the multiple channel hypothesis of human vision. The results were compared with those of other investigators employing sine-wave gratings, perceptual response measures and the evoked potentials technique.
INTRODUCTION
The purpose of this study was to determine whether simple visual reaction times to the presentation of a grating pattern are affected by manipulation of the parameters of contrast and spatial frequency of the stimulus. The impetus for the investigation was provided by recent advances in knowledge about the mechanisms of pattern vision. It is known that the perceptual response to presentation of a grating depends on its contrast, provided that no change in overall luminance occurs. Specifically, Fiorentini and Maffei11 have shown that apparent contrast is linearly and positively related to the log of stimulus contrast. Moreover, for a grating of constant luminance, the contrast threshold, as well as supra-threshold sensitivity, are determined by the
112 spatial frequency of the stimulus, i.e., the width of the bars. Sensitivity is high for a medium range of spatial frequencies and decreases at higher and lower frequencies 9. Psychophysiological evidence suggests that such selective sensitivity depends on the characteristics of the central visual structures, the latter being an expression of the properties of dimension-detector mechanisms, each of which is tuned to a narrow range of frequencies 8. The behavior of evoked potentials, presumably an expression of the activation of visual occipital structures, shows a similar dependence on contrast as well as on the spatial frequency of the grating patterns. In both humans and animals, the zeroamplitude contrast for evoked potentials coincides with the perceptual threshold ,5,6 and their amplitude increases at supra-threshold contrast as a linear function of the log of the object contrast6, v. In humans, moreover, simultaneous presentation of two spatial frequencies results in a significant increase in the amplitude of the evoked potentials, suggesting the simultaneous activation of different structures, each of which is responsive to a spatial frequency 6. In animals the dependence of evoked potentials on contrast and spatial frequency correponds to the activity of specific groups of cells in the visual cortex. It has been shown that different neural populations exist, each of which is tuned to a preferred spatial frequency 4,1~. They encode the information about supra-threshold contrast, changing their frequency of discharge as a linear function of the log of object contrast. Psychophysical data provide evidence that the brain of man contains similar cells ~. These findings indicate that this type of stimulus is elementary enough to bring into action the simplest structural mechanism involved in processing the visual information. Any modification of the parameters of the grating pattern appears to be followed by a specific functional change in the neural substrates. Evoked potentials presumably follow the electrical changes occurring in these structures. Finally, the correlation of the perceptual response with the behavior of evoked potentials suggests their direct dependence upon these simple mechanisms. Hence it is conceivable that the properties of cortical feature-detectors, whose activation is correlated with changes in evoked potentials, determine the properties of the perceptual response. Thus, it is of interest to investigate whether a behavioral response derived from the processing of visual information, i.e., reaction time (RT) to the detection of a grating, correlates with changes in the parameters of the stimulus. It seems reasonable to expect that RT would show the same dependence on contrast and spatial frequency as does the perceptual response, if both depend on the same processing mechanisms. Information on this point may provide deeper insight on the transfer of information from the visual cortex to the motor areas. The use of RT to determine whether functional changes taking place in the visual structures are transferred to the motor area was suggested by the studies of Donchin and Lindsley 1° and Morrell and Morrel113 who showed that intraindividual variability in simple RT to unpatterned flashes of constant luminance is related to cortical activity as defined by the amplitude of the corresponding evoked potential.
113 Hence it is possible that RT is affected by the modification of cortical activity following the presentation of the stimulus. In the study to be reported, grating patterns of different spatial frequency and contrast were used to determine whether RT is affected by changes in both parameters of the stimulus. In addition, the effect of the simultaneous presentation of two spatial frequencies was analyzed. METHODS
Stimuli and apparatus Patterned bar stimuli with essentially square-wave luminance distributions were presented tachistoscopically. The patterns varied in contrast or spatial frequency or both parameters and were identical with respect to average luminance, as described below. Stimulus cards were graphically prepared from photographic material. Each consisted of a square patterned area surrounded by an unpatterned field whose luminance was equal to the space-average luminance of the pattern. The patterned area had 26-ram sides, subtending a visual angle of 2 degrees at a viewing distance of 75 cm. Three patterns of different spatial frequency were used. Two of them contained one spatial frequency, respectively one cycle/degree and 3 cycles/degree. The third pattern consisted of the 2 frequencies with half of the one cycle/degree stimulus located in the upper half of the card and half of the 3 cycles/degree stimulus in the lower half. Consequently, the total area of the combination stimulus was identical to the stimuli with one spatial frequency. The bars were horizontally oriented in order to avoid selective processing of either spatial frequency by one hemisphere during the presentation of the third pattern. The bars were made from different shades of gray photographic paper in order to present each pattern at different levels of contrast. Three different levels of contrast were used, 0.12, 0.26, and 0.37 (as defined below). Therefore, 9 stimulus cards were designed, each spatial frequency for each of the 3 levels of contrast. Particular attention was given to the stimulus construction process in order to control contrast while maintaining space-average luminance constant for all patterns. Various shades of gray were produced by exposing large sheets of Kodak Portralure (matte finish) to a distant point source of light for varying exposure times and then developing according to standard instructions. After measuring brightness with a Pritchard Spectra Photometer under the constant luminance condition employed in the experiment, sample pairs were chosen which produced the desired contrast at the constant space-average luminance level. The actual patterns were produced by cutting out the bar pattern in the lighter sample and glueing it to the darker one. From the values of luminance of the light bars (LM) and of the dark bars (Lm), the actual contrast (C) was determined by the equation C = LM--Lm/ LM+ Lm where LM+ Lm/2 is equal to the mean luminance of the pattern. The standard deviation from the average luminance of the population of the stimuli did not exceed
114 3 ~ of the mean. No systematic variation in luminance was related to either the different spatial frequencies or the different levels of contrast. The stimuli were presented in a 2-channel tachistoscope (Scientific Prototype Model 800-F), using standard G E F6T5D fluorescent bulbs. The normal distance between the bulbs was decreased, the normal distance from bulbs to stimulus was increased, and parabolic reflectors were mounted behind the bulbs so as to produce more parallel light rays orthogonal to the stimulus surface. This procedure was employed to minimize shadows between the bars resulting from one set of bars being one paper thickness above the other set. The fixation field, whose luminance and color were adjusted to equal the space average luminance and color of the stimulus, was used as pre- and post-exposure field in order to minimize changes in the overall luminance during presentation of the stimuli. It consisted of a gray card with a black dot in the center of a square defined by 4 corner brackets. Each side of the square subtended 2 degrees of visual angle. The stimulus was flashed for 50 msec. The average luminance of both the fixation and the stimulus field was set at 1.5 ft. Lambert. During each experimental session, the luminance of the fields was repeatedly measured in order to compensate for eventual variations of the performance of the fluorescent bulbs. When changes had occurred during an experimental session, they were found not to exceed ~: 0.05 ft. Lamberts from the pre-set value.
Subjects The subjects were 13 university students (11 men, 2 women) who were strongly right-handed as defined by their indication of consistent preference for the use of the right hand on a variety of activities listed in a handedness inventory. All the subjects scored within the normal range on tests of phoria (Bausch and Lomb Orthorater) and had an uncorrected visual acuity of at least 20/25. All but one subject (S) had no previous experience on an RT task and were naive about the purpose of the experiment.
Procedure S was seated before the tachistoscope with his head resting against the viewing hood, with the right index finger pressing a reaction time key, and binocularly viewing the stimuli. The session began with a 5-rain period of adaptation to the level of luminance of the stimuli, during which the subject was familiarized with the different patterns and with the RT procedure. S was instructed to look straight at the fixation point at the occurrence of a warning tone and to release a spring loaded key as quickly as possible when any of the 9 patterns appeared. Release of the key stopped a millisecond reaction timer which was started at the onset of the stimulus flash. The interval between the warning tone and onset of the stimulus was randomized among 4 possible delays (1.5, 1.8, 2.1, and 2.4 sec) in order to minimize the effect of habituation on RT. In a single trial block, each stimulus was presented once at each of the different combinations of tone and stimulus delay and the same interval was never repeated two times consecutively. The sequence of the stimuli was previously programmed so that no stimulus
115 TABLE I MEDIAN R T SCORES FOR INDIVIDUAL SUBJECTS AS A FUNCTION OF CONTRASTAND SPATIAL FREQUENCY
Subjects
D.G. N.V. P.B. M.E. S.C. M.D. P.K. T.M. B.B. L.S. D.S. C.S. M.S.
Contrast,0.37
Contrast, 0.26
Contrast, 0.12
Spatial frequency 1 cpd 3 cpd 1÷ 3 cpd
Spatial frequency 1 cpd 3 cpd 1 Jr 3 cpd
Spatial frequency 1 cpd 3 cpd 1+ 3 epd
198.5 194.5 186.5 227.5 216.5 254.5 196.5 259.0 236.5 223.5 231.0 228.5 215.5
194.5 191.0 187.5 242.0 238.5 232.5 197.0 257.0 247.5 232.0 226.5 227.5 232.5
207.0 203.0 200.5 243.5 228.5 258.5 216.0 262.0 266.5 229.0 213.5 229.5 228.5
211.5 213.0 187.5 249.5 233.5 269.0 211.0 257.5 281.5 233.5 217.0 228.5 231.5
192.5 204.0 177.5 247.5 234.0 248.5 216.0 259.5 283.0 222.5 222.5 225.0 237.5
208.5 194.5 192.5 251.5 238.5 248.0 210.5 273.0 268.5 214.0 226.5 219.5 229.5
212.5 200.0 191.5 234.0 231.0 264.0 232.5 294.5 249.0 233.5 217.0 215.5 226.5
209.0 223.0 193.5 250.5 247.0 266.5 196.5 280.0 291.0 213.5 233.5 228.0 256.5
215.5 202.0 194.0 241,5 243.0 258.5 225.5 271.5 272.5 219.0 222.5 229.5 223.5
* cpd = cycle(s)/degree for this Table and Table III.
was repeated two times in a row and each pattern could follow only once a pattern of the same contrast or a pattern of the same spatial frequency. In this way, the effect of ipsimodal presentation on RT was controlled 1. Any trial for which RT was longer than 350 msec or shorter than 100 msec was categorized as a mistrial and rescheduled to the end of a trial block. These were very rare occurrences. Four blocks of 36 trials were given and within a block, each stimulus was repeated 4 times. A total of 144 trials were run for each subject. At the end of each block, Ss were allowed to rest 3-5 min while luminance of both fixation and stimulus fields were measured and readjusted if necessary. Any block did not last more than 20 min and total duration of each session was less than 2 h. RESULTS
The median R T of each subject for each of the 9 stimulus conditions appears in Table I. The individual scores were subjected to a treatment x treatment x subject analysis of variance 14. Treatment A, i.e., the contrast, is present at 3 different levels, namely 0.12, 0.26, 0.37 as well as treatment B, the low spatial frequency pattern (one cycle/degree), the high spatial frequency pattern (3 cycles/degree), and the complex pattern, combining one and 3 cycles/degree. The summary of the statistical analysis is given in Table I1. It is evident that the changes in both level of contrast and the spatial characteristics of the patterns have an effect on reaction time. The statistical analysis shows that the individual median reaction times differ as a function of level of contrast
116 TABLE II SUMMARY ANALYSIS OF VARIANCE ON MEDIAN
Source o f variation
df
Contrast (A) Spatial frequency (B) Subjects (S) A A B A
× B .'< S ~ S ~: B x S
2 2 12 4 24 24 48
R T SCORES FOR CONTRAST AND SPATIAL FREQUENCY
MS
F
417.05 724.48 5242.11 77.85 75.80 117.72 84.20
5.50* 6.15"* 0.92
116 * P < 0.025. ** P < 0.01.
(P < 0.025) as well as of the various patterns (P < 0.01). The interaction between spatial frequency and contrast was not significant, implying an additive role of the two factors. Thus, it can be concluded that the effect o f the different levels o f contrast is present for each spatial frequency and the influence o f spatial frequency does not vary across each level o f contrast. Table III presents the mean o f the median scores for each stimulus averaged across subjects. The mean difference in reaction times to stimuli characterized by a c o m m o n level o f either contrast or spatial frequency as well as their level o f significance are also presented in this table. TABLE I11 MEANS OF MEDIAN
R T SCORES FOR 3 LEVELS OF CONTRAST AND 3 SPATIAL FREQUENCY CONDITIONS WITH
PAIRED COMPARISONS FOR EACH VARIABLE
Spatial frequency
Contrast level 0.37
0.26
0.12
1 cpd 3 cpd 1 cpd and 3 cpd
220.6 232.6 228.5
223.5 228.8 230.9
228.7 237.6 232.2
Mean
227.3
227.7
233.1
Mean
Contrast levels
RTdifferences
224.6 233.0 230.5
0.124).37 0.12-0.26 0.26-0.37
5.8 msec*** 5.4 msec** 0.4msec (NS)
Spatial frequency conditions 1 cpd and 3 cpd --.. 1 cpd 3 cpd - - 1 cpd and 3 cpd 3 cpd - - 1 cpd * P < 0.05.
** P < 0.025. *** P
111
REACTION TIME TO SQUARE-WAVE GRATINGS AS A FUNCTION OF SPATIAL FREQUENCY, COMPLEXITY AND CONTRAST
A N T O N I O TARTAGL1ONE, DAVID P. G O F F AND A R T H U R L. BENTON
Neurosensory Center and Department of Neurology, University of lowa, lowa City, Iowa 52242 (U.S.A.) and Clinica delle Malattie Nervose e Mentali dell'Universitgt, 16132 Genoa (Italy) (Accepted June 7th, 1975)
SUMMARY
Visuomotor reaction time (RT) of normal human subjects to briefly presented square-wave bar patterns were measured. The pattern varied with respect to magnitude and number of spatial frequencies (one cycle/degree of visual angle, 3 cycles/ degree, and both combined) with each spatial frequency stimulus presented at 3 levels of contrast. Space-average luminance was held constant. RT increased with the higher spatial frequency and was intermediate in value for the combination stimulus at each level of contrast. RT increased with decreasing contrast levels for each spatial frequency stimulus. There was no interaction between the two parameters. The findings indicate that a behavioral response measure can be employed to study the multiple channel hypothesis of human vision. The results were compared with those of other investigators employing sine-wave gratings, perceptual response measures and the evoked potentials technique.
INTRODUCTION
The purpose of this study was to determine whether simple visual reaction times to the presentation of a grating pattern are affected by manipulation of the parameters of contrast and spatial frequency of the stimulus. The impetus for the investigation was provided by recent advances in knowledge about the mechanisms of pattern vision. It is known that the perceptual response to presentation of a grating depends on its contrast, provided that no change in overall luminance occurs. Specifically, Fiorentini and Maffei11 have shown that apparent contrast is linearly and positively related to the log of stimulus contrast. Moreover, for a grating of constant luminance, the contrast threshold, as well as supra-threshold sensitivity, are determined by the
112 spatial frequency of the stimulus, i.e., the width of the bars. Sensitivity is high for a medium range of spatial frequencies and decreases at higher and lower frequencies 9. Psychophysiological evidence suggests that such selective sensitivity depends on the characteristics of the central visual structures, the latter being an expression of the properties of dimension-detector mechanisms, each of which is tuned to a narrow range of frequencies 8. The behavior of evoked potentials, presumably an expression of the activation of visual occipital structures, shows a similar dependence on contrast as well as on the spatial frequency of the grating patterns. In both humans and animals, the zeroamplitude contrast for evoked potentials coincides with the perceptual threshold ,5,6 and their amplitude increases at supra-threshold contrast as a linear function of the log of the object contrast6, v. In humans, moreover, simultaneous presentation of two spatial frequencies results in a significant increase in the amplitude of the evoked potentials, suggesting the simultaneous activation of different structures, each of which is responsive to a spatial frequency 6. In animals the dependence of evoked potentials on contrast and spatial frequency correponds to the activity of specific groups of cells in the visual cortex. It has been shown that different neural populations exist, each of which is tuned to a preferred spatial frequency 4,1~. They encode the information about supra-threshold contrast, changing their frequency of discharge as a linear function of the log of object contrast. Psychophysical data provide evidence that the brain of man contains similar cells ~. These findings indicate that this type of stimulus is elementary enough to bring into action the simplest structural mechanism involved in processing the visual information. Any modification of the parameters of the grating pattern appears to be followed by a specific functional change in the neural substrates. Evoked potentials presumably follow the electrical changes occurring in these structures. Finally, the correlation of the perceptual response with the behavior of evoked potentials suggests their direct dependence upon these simple mechanisms. Hence it is conceivable that the properties of cortical feature-detectors, whose activation is correlated with changes in evoked potentials, determine the properties of the perceptual response. Thus, it is of interest to investigate whether a behavioral response derived from the processing of visual information, i.e., reaction time (RT) to the detection of a grating, correlates with changes in the parameters of the stimulus. It seems reasonable to expect that RT would show the same dependence on contrast and spatial frequency as does the perceptual response, if both depend on the same processing mechanisms. Information on this point may provide deeper insight on the transfer of information from the visual cortex to the motor areas. The use of RT to determine whether functional changes taking place in the visual structures are transferred to the motor area was suggested by the studies of Donchin and Lindsley 1° and Morrell and Morrel113 who showed that intraindividual variability in simple RT to unpatterned flashes of constant luminance is related to cortical activity as defined by the amplitude of the corresponding evoked potential.
113 Hence it is possible that RT is affected by the modification of cortical activity following the presentation of the stimulus. In the study to be reported, grating patterns of different spatial frequency and contrast were used to determine whether RT is affected by changes in both parameters of the stimulus. In addition, the effect of the simultaneous presentation of two spatial frequencies was analyzed. METHODS
Stimuli and apparatus Patterned bar stimuli with essentially square-wave luminance distributions were presented tachistoscopically. The patterns varied in contrast or spatial frequency or both parameters and were identical with respect to average luminance, as described below. Stimulus cards were graphically prepared from photographic material. Each consisted of a square patterned area surrounded by an unpatterned field whose luminance was equal to the space-average luminance of the pattern. The patterned area had 26-ram sides, subtending a visual angle of 2 degrees at a viewing distance of 75 cm. Three patterns of different spatial frequency were used. Two of them contained one spatial frequency, respectively one cycle/degree and 3 cycles/degree. The third pattern consisted of the 2 frequencies with half of the one cycle/degree stimulus located in the upper half of the card and half of the 3 cycles/degree stimulus in the lower half. Consequently, the total area of the combination stimulus was identical to the stimuli with one spatial frequency. The bars were horizontally oriented in order to avoid selective processing of either spatial frequency by one hemisphere during the presentation of the third pattern. The bars were made from different shades of gray photographic paper in order to present each pattern at different levels of contrast. Three different levels of contrast were used, 0.12, 0.26, and 0.37 (as defined below). Therefore, 9 stimulus cards were designed, each spatial frequency for each of the 3 levels of contrast. Particular attention was given to the stimulus construction process in order to control contrast while maintaining space-average luminance constant for all patterns. Various shades of gray were produced by exposing large sheets of Kodak Portralure (matte finish) to a distant point source of light for varying exposure times and then developing according to standard instructions. After measuring brightness with a Pritchard Spectra Photometer under the constant luminance condition employed in the experiment, sample pairs were chosen which produced the desired contrast at the constant space-average luminance level. The actual patterns were produced by cutting out the bar pattern in the lighter sample and glueing it to the darker one. From the values of luminance of the light bars (LM) and of the dark bars (Lm), the actual contrast (C) was determined by the equation C = LM--Lm/ LM+ Lm where LM+ Lm/2 is equal to the mean luminance of the pattern. The standard deviation from the average luminance of the population of the stimuli did not exceed
114 3 ~ of the mean. No systematic variation in luminance was related to either the different spatial frequencies or the different levels of contrast. The stimuli were presented in a 2-channel tachistoscope (Scientific Prototype Model 800-F), using standard G E F6T5D fluorescent bulbs. The normal distance between the bulbs was decreased, the normal distance from bulbs to stimulus was increased, and parabolic reflectors were mounted behind the bulbs so as to produce more parallel light rays orthogonal to the stimulus surface. This procedure was employed to minimize shadows between the bars resulting from one set of bars being one paper thickness above the other set. The fixation field, whose luminance and color were adjusted to equal the space average luminance and color of the stimulus, was used as pre- and post-exposure field in order to minimize changes in the overall luminance during presentation of the stimuli. It consisted of a gray card with a black dot in the center of a square defined by 4 corner brackets. Each side of the square subtended 2 degrees of visual angle. The stimulus was flashed for 50 msec. The average luminance of both the fixation and the stimulus field was set at 1.5 ft. Lambert. During each experimental session, the luminance of the fields was repeatedly measured in order to compensate for eventual variations of the performance of the fluorescent bulbs. When changes had occurred during an experimental session, they were found not to exceed ~: 0.05 ft. Lamberts from the pre-set value.
Subjects The subjects were 13 university students (11 men, 2 women) who were strongly right-handed as defined by their indication of consistent preference for the use of the right hand on a variety of activities listed in a handedness inventory. All the subjects scored within the normal range on tests of phoria (Bausch and Lomb Orthorater) and had an uncorrected visual acuity of at least 20/25. All but one subject (S) had no previous experience on an RT task and were naive about the purpose of the experiment.
Procedure S was seated before the tachistoscope with his head resting against the viewing hood, with the right index finger pressing a reaction time key, and binocularly viewing the stimuli. The session began with a 5-rain period of adaptation to the level of luminance of the stimuli, during which the subject was familiarized with the different patterns and with the RT procedure. S was instructed to look straight at the fixation point at the occurrence of a warning tone and to release a spring loaded key as quickly as possible when any of the 9 patterns appeared. Release of the key stopped a millisecond reaction timer which was started at the onset of the stimulus flash. The interval between the warning tone and onset of the stimulus was randomized among 4 possible delays (1.5, 1.8, 2.1, and 2.4 sec) in order to minimize the effect of habituation on RT. In a single trial block, each stimulus was presented once at each of the different combinations of tone and stimulus delay and the same interval was never repeated two times consecutively. The sequence of the stimuli was previously programmed so that no stimulus
115 TABLE I MEDIAN R T SCORES FOR INDIVIDUAL SUBJECTS AS A FUNCTION OF CONTRASTAND SPATIAL FREQUENCY
Subjects
D.G. N.V. P.B. M.E. S.C. M.D. P.K. T.M. B.B. L.S. D.S. C.S. M.S.
Contrast,0.37
Contrast, 0.26
Contrast, 0.12
Spatial frequency 1 cpd 3 cpd 1÷ 3 cpd
Spatial frequency 1 cpd 3 cpd 1 Jr 3 cpd
Spatial frequency 1 cpd 3 cpd 1+ 3 epd
198.5 194.5 186.5 227.5 216.5 254.5 196.5 259.0 236.5 223.5 231.0 228.5 215.5
194.5 191.0 187.5 242.0 238.5 232.5 197.0 257.0 247.5 232.0 226.5 227.5 232.5
207.0 203.0 200.5 243.5 228.5 258.5 216.0 262.0 266.5 229.0 213.5 229.5 228.5
211.5 213.0 187.5 249.5 233.5 269.0 211.0 257.5 281.5 233.5 217.0 228.5 231.5
192.5 204.0 177.5 247.5 234.0 248.5 216.0 259.5 283.0 222.5 222.5 225.0 237.5
208.5 194.5 192.5 251.5 238.5 248.0 210.5 273.0 268.5 214.0 226.5 219.5 229.5
212.5 200.0 191.5 234.0 231.0 264.0 232.5 294.5 249.0 233.5 217.0 215.5 226.5
209.0 223.0 193.5 250.5 247.0 266.5 196.5 280.0 291.0 213.5 233.5 228.0 256.5
215.5 202.0 194.0 241,5 243.0 258.5 225.5 271.5 272.5 219.0 222.5 229.5 223.5
* cpd = cycle(s)/degree for this Table and Table III.
was repeated two times in a row and each pattern could follow only once a pattern of the same contrast or a pattern of the same spatial frequency. In this way, the effect of ipsimodal presentation on RT was controlled 1. Any trial for which RT was longer than 350 msec or shorter than 100 msec was categorized as a mistrial and rescheduled to the end of a trial block. These were very rare occurrences. Four blocks of 36 trials were given and within a block, each stimulus was repeated 4 times. A total of 144 trials were run for each subject. At the end of each block, Ss were allowed to rest 3-5 min while luminance of both fixation and stimulus fields were measured and readjusted if necessary. Any block did not last more than 20 min and total duration of each session was less than 2 h. RESULTS
The median R T of each subject for each of the 9 stimulus conditions appears in Table I. The individual scores were subjected to a treatment x treatment x subject analysis of variance 14. Treatment A, i.e., the contrast, is present at 3 different levels, namely 0.12, 0.26, 0.37 as well as treatment B, the low spatial frequency pattern (one cycle/degree), the high spatial frequency pattern (3 cycles/degree), and the complex pattern, combining one and 3 cycles/degree. The summary of the statistical analysis is given in Table I1. It is evident that the changes in both level of contrast and the spatial characteristics of the patterns have an effect on reaction time. The statistical analysis shows that the individual median reaction times differ as a function of level of contrast
116 TABLE II SUMMARY ANALYSIS OF VARIANCE ON MEDIAN
Source o f variation
df
Contrast (A) Spatial frequency (B) Subjects (S) A A B A
× B .'< S ~ S ~: B x S
2 2 12 4 24 24 48
R T SCORES FOR CONTRAST AND SPATIAL FREQUENCY
MS
F
417.05 724.48 5242.11 77.85 75.80 117.72 84.20
5.50* 6.15"* 0.92
116 * P < 0.025. ** P < 0.01.
(P < 0.025) as well as of the various patterns (P < 0.01). The interaction between spatial frequency and contrast was not significant, implying an additive role of the two factors. Thus, it can be concluded that the effect o f the different levels o f contrast is present for each spatial frequency and the influence o f spatial frequency does not vary across each level o f contrast. Table III presents the mean o f the median scores for each stimulus averaged across subjects. The mean difference in reaction times to stimuli characterized by a c o m m o n level o f either contrast or spatial frequency as well as their level o f significance are also presented in this table. TABLE I11 MEANS OF MEDIAN
R T SCORES FOR 3 LEVELS OF CONTRAST AND 3 SPATIAL FREQUENCY CONDITIONS WITH
PAIRED COMPARISONS FOR EACH VARIABLE
Spatial frequency
Contrast level 0.37
0.26
0.12
1 cpd 3 cpd 1 cpd and 3 cpd
220.6 232.6 228.5
223.5 228.8 230.9
228.7 237.6 232.2
Mean
227.3
227.7
233.1
Mean
Contrast levels
RTdifferences
224.6 233.0 230.5
0.124).37 0.12-0.26 0.26-0.37
5.8 msec*** 5.4 msec** 0.4msec (NS)
Spatial frequency conditions 1 cpd and 3 cpd --.. 1 cpd 3 cpd - - 1 cpd and 3 cpd 3 cpd - - 1 cpd * P < 0.05.
** P < 0.025. *** P
