International Journal of Psychophysiology 61 (2006) 98 – 112 www.elsevier.com/locate/ijpsycho

Task modulation of the effects of brightness on reaction time and response force Piotr Jas´kowski a,*, Dariusz Waodarczyk b a

Department of Cognitive Psychology, University of Finance and Management, Pawia 55, 01-030 Warsaw, Poland b Department of Biophysics, Karol Marcinkowski University of Medicine of Poznan´, Poland Received 15 April 2004; received in revised form 12 April 2005; accepted 10 July 2005 Available online 28 September 2005

Abstract Van der Molen and Keuss [van der Molen, M.W., Keuss, P.J.G., 1979. The relationship between reaction time and intensity in discrete auditory tasks. Quarterly Journal of Experimental Psychology 31, 95 – 102; van der Molen, M.W., Keuss, P.J.G., 1981. Response selection and the processing of auditory intensity. Quarterly Journal of Experimental Psychology 33, 177 – 184] showed that paradoxically long reaction times (RT) occur with extremely loud auditory stimuli when the task is difficult (e.g. needs a response choice). It was argued that this paradoxical behavior of RT is due to active suppression of response prompting to prevent false responses. In the present experiments, we demonstrated that such an effect can also occur for visual stimuli provided that they are large enough. Additionally, we showed that response force exerted by participants on response keys monotonically grew with intensity for large stimuli but was independent of intensity for small visual stimuli. Bearing in mind that only large stimuli are believed to be arousing this pattern of results supports the arousal interpretation of the negative effect of loud stimuli on RT given by van der Molen and Keuss. D 2005 Elsevier B.V. All rights reserved. Keywords: Reaction time; Stimulus intensity; Response force

1. Introduction 1.1. Negative effect of loud stimuli on RT Simple reaction time (RT) is known to decrease as a function of stimulus intensity, approaching an asymptote for the most intense stimuli. These changes of simple RT have usually been assigned to early stages of sensory processing: it is assumed that intensity influences the duration of early, perceptual stages, and not later, motor-related stages (for reviews see Jas´kowski, 1996, 1999; Miller et al., 1999b). This opinion is questionable. To verify this view, the effect of intensity on simple RT and the latency of early components of event related potentials (ERP) could be compared. Using this approach, Vaughan et al. (1966), Wilson and Lit (1981) and Jas´kowski et al. (1990) found that visual intensity had

* Corresponding author. E-mail address: [email protected] (P. Jas´kowski). 0167-8760/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpsycho.2005.07.010

identical effect on SRT and on ERP latency. However, Jas´kowski et al. (1994a) found a larger effect of intensity on simple RT than on ERP latency for auditory stimuli, and more recently comparable results were reported for visual stimuli by Kammer et al. (1999). This logic, however, implies that latencies of ERP components and RT are directly comparable, which has been criticized by Meyer et al. (1988). Another approach to locate stimulus intensity effects was applied by van der Molen and Keuss’ experiments (Keuss and van der Molen, 1982; van der Molen and Keuss, 1979, 1981; van der Molen and Orlebeke, 1980). They showed that for auditory stimuli the relation between RT and loudness depended on the participants’ task. While for the simple and go/no-go task RT monotonically decreased with intensity, the relation was flatter or even U-shaped when the task was more difficult (e.g. choice or Simon task): for extremely loud auditory stimuli (above 85 dB) paradoxically long RTs were found. The results of these experiments suggest at least that loudness affects more distal processes. The logic behind this

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

inference is straightforward. If stimulus intensity affects only very early stages of processing, then it should affect RT independently of the tasks participants perform. Any changes across tasks in the effect of intensity on RT may be interpreted as evidence in favor of the hypothesis that intensity also affects later stages of processing. In their early writing, van der Molen and Keuss (1979) suggested that strong auditory signals elicit a startle reflex which has to be suppressed in order to prevent a degradation of performance when the participant’s task is difficult. Later, they referred, after Sanders (1980), to the concept of immediate arousal. Sanders (1983) claimed that van der Molen and Keuss’ effect could be accounted for in the framework of his model of stress which assumes that the main chain of processing stages (computational stages) is supplied by three energetic interdependent resources: arousal, activation, and effort. Arousal is considered as a transient response to the input, related to the stage of feature extraction. Activation is readiness to respond and is related to the state of response preparation. Arousal is assumed to directly affect activation, with the degree of this influence being regulated by effort. In particular, if the task is more complex or needs precision (like choice reaction) effort is able to reduce the influence of arousal to avoid unacceptable error rates (Sanders, 1983). Thus, the input signals increase the arousal level leading to automatic increase of activation and shortening of simple RTs. If, however, the response needs a choice, the effect of immediate arousal on activation evoked by input signals has to be suppressed, which may be implemented by disconnecting arousal from activation. A shortcoming of this hypothesis is that it does not make clear why RTs are longer for the loudest than for moderately loud stimuli. Indeed, when intensity increases then RT should still asymptotically decrease due to the shortening of early computational stages, even when RT does not reap profit from increased arousal anymore. To account for the RT increase one may assume that in choice tasks the suppression of arousal overshoots leading to a suboptimal level of readiness. Such a mechanism was proposed by van der Molen and Keuss (Keuss and van der Molen, 1982; van der Molen and Keuss, 1981). 1.2. Effect of brightness on RT With visual stimuli, the shape of the RT – intensity relationship is usually independent of the complexity of the tasks the participants have to perform (Azorin et al., 1995; Egeth, 1977; Everett et al., 1985; Pins and Bonnet, 1996; Schwarz et al., 1977; Schweickert et al., 1988; Van Duren and Sanders, 1988). For example, Pins and Bonnet (1996) showed that although the asymptotic value of RT depended on task complexity, the variable part of RT changed with intensity identically, irrespective of whether simple or choice reactions were required and whether identification or categorization was needed.

99

The lack of the difference with visual stimuli as compared to the effect with auditory stimuli as found by van der Molen and Keuss has usually been ascribed to the non-arousing properties of visual stimuli. This view is based on experiments where the effect of foreperiod duration on RT was measured for different intensities. With auditory stimuli, the cost of lengthening the foreperiod from 1 to 5 s was reduced for loud tones as compared to soft tones (Bertelson, 1969; Niemi, 1979; Sanders and Wertheim, 1973). Sanders and Wertheim (1973) ascribed this finding to the arousal characteristics of auditory stimuli: arousing properties of imperative stimuli are less effective when arousal is already high, i.e. just after warning stimuli. In contrast, RTs to visual stimuli of different intensities suffer equally from increasing the foreperiod (Bertelson, 1969; Niemi, 1979; Sanders and Wertheim, 1973). These results with visual stimuli seem to imply that brightness affects only early computational stages that are shared in both tasks. As a consequence, the effect of brightness should be identical for these tasks. Sanders (1975) and Niemi and Lehtonen (1982) showed that visual intensity interacted with foreperiod duration when visual stimuli were large and/or very bright similarly as in case of auditory stimuli. These findings suggested that such visual stimuli possess also arousal properties. Niemi and Lehtonen (1982) used stimuli of size 29  32- and of luminance 130 cd/m2. Maximal luminance of stimuli used by Sanders was equal to 650 cd/m2. No information was given about the size of stimuli. If visual stimuli indeed have arousing properties, large and bright stimuli should produce a pattern of results similar to that obtained by van der Molen and Keuss, i.e. a U-shaped RT – intensity relationship. As mentioned, it has usually been shown that relationship between RT and brightness is independent of task difficulty (Azorin et al., 1995; Egeth, 1977; Everett et al., 1985; Pins and Bonnet, 1996; Schwarz et al., 1977; Schweickert et al., 1988; Van Duren and Sanders, 1988). However, in such studies usually small stimuli of moderate luminances were used. For example, in the extensive study by Pins and Bonnet (1996) stimuli were as small as 30  7.5 min of arc and at most 2  3- (in their Experiment 4) and covered a rather narrow intensity range (for simple RT stimulus intensities ranged from 0.18 to 58.2 cd/m2 only, and from 0.28 to 3.09 cd/m2 only in choice tasks). Schweickert et al. (1988) used a wider range of luminance but their stimuli were rather small (0.4  1.8-) and their maximal luminance was as small as 30 cd/m2. There are, however, at least two studies where some effect of task on the RT –brightness relationship was found. Stanovitch and Pachella (1977) reported that the effect of luminance on RT was smaller when stimulus –response compatibility was low than when it was high. Even more interesting is a study by Kaswan and Young (1965) who found a U-shaped RT – luminance curve for their most difficult tasks. Unfortunately, their data were rather noisy and the authors provided no statistical analysis to show that the effect was significant.

100

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

The main aim of the present study was to check if van der Molen and Keuss-like effects can also be observed for visual stimuli when they are large and cover a wide range of luminance. In Experiment 1 we tested this hypothesis with stimuli as large as 8.5  8.5- with a luminance over 300 cd/m2. 1.3. Response force Besides RT, we measured response force (RF, i.e. the maximal force that participants exert on the response key). This was done for three reasons: (1) According to one of the contemporary models RF is predominately determined by arousal (for reviews see Jas´kowski et al., 2000; Mattes et al., 2002; Miller et al., 1999a). There is some evidence that general arousal or stress affects RF. The higher the level of arousal, the larger is RF. Thus, responses were stronger under more stressful conditions like time pressure, i.e. when the time for response was limited (Jas´kowski et al., 1994b, 2000; Van der Lubbe et al., 2001), knowledge of results, i.e. when RT results were fed back after every trial (Jas´kowski and Waodarczyk, 1997) and occasional delivery of task-irrelevant electrical shocks (Jas´kowski et al., 1994c). Correspondingly, suboptimal activation induced by sleep deficit leads to delay of RTs and weakening of RF (Waodarczyk et al., 2002). There are also some indications that stimulus-induced transient arousal affects RF. It was shown several times that in simple RT task participants responded not only faster to loud than to soft tones but also more forcefully (Jas´kowski et al., 1995; Miller et al., 1999a; Ulrich et al., 1998). As mentioned, it is commonly believed that loud tones elicit transient arousal. For example, Sanders (1983) argued that arousal is transmitted to the motor stages via the activation system, thus outside the information-processing system. Similar bypass mechanism is assumed by Miller et al. (1999a) to account for the relation between RF and loudness. Some other arguments were provided recently by Mattes et al. (2002) who showed that in a go/no-go task stronger responses were elicited by less frequent stimuli. One possibility to account for this finding discussed by Mattes et al. is that rare stimuli evoke a transient increase of arousal level. Some indirect support for this idea is, according to Mattes et al., neuronal activity evoked by infrequent stimuli like mismatch negativity (Na¨a¨ta¨nen, 1995) and P300 (Verleger, 1988, 1998). (2) Arousal models predict that RF should increase with loudness but much less with luminance, since auditory signals are considered more arousing than visual stimuli as discussed above. While the effect of loudness on RF was quite robust (Jas´kowski et al., 1995; Miller et al., 1999a,b; Ulrich et al., 1998), the relation for visual stimuli is not so clear. Jas´kowski et al. (1995) and Miller et al. (1991) found no effect of brightness on RF. However, in Angel’s (1973) pioneering study as well as in a more recent study by Ulrich et al. (1998) effects of both luminance and loudness on RF

were found. Reasons for these discrepancies are not clear. One possibility is that the stimuli used by Jas´kowski et al. (1995) and by Miller et al. (1991) were less arousing that those used by Angel (1973) and by Ulrich et al. (1998). Our results of Experiment 2, where the effect of stimulus luminance on RF and RT was compared for large and small stimuli, indicate that this is a very probable reason for the discrepant results reported by different authors. (3) Assuming that arousal is responsible for more forceful responses we expected a gradual increase of RF with intensity brightness for the simple-response task as found by Ulrich et al. (1998) given that large visual stimuli posses an arousing property. Of special interest are, however, the changes of RF in the choice task. The suppression hypothesis by van der Molen and Keuss suggests an inverted U-shaped relation because for very bright stimuli arousal has to be suppressed to avoid wrong responses and, consequently, participants should respond weaker in this task.

2. Experiment 1 In Experiment 1 we used relatively large (8.5-) and bright (over 300 cd/m2) visual stimuli. A general prediction based on van der Molen and Keuss’ papers (Keuss and van der Molen, 1982; van der Molen and Keuss, 1979; van der Molen and Orlebeke, 1980; van der Molen and Keuss, 1981) is that a monotonically decreasing RT – intensity relation should be found for simple or go/no-go tasks whereas a U-shaped relation should be found when a choice is required. Furthermore, the higher the response selection demands, the more curvilinear the RT –intensity function. The last conclusion was based on the experiment with Simon-like task (van der Molen and Keuss, 1981). A high or low tone was presented monaurally to the left or right ear. The participants’ task was to react with their left or right hand according to pitch while ignoring the tone’s presentation side. In the so-called correlated condition, a cue was presented before the tone which indicated that the stimulus content corresponds with the presentation side. This way the selection demands were remarkably reduced in respect to uncorrelated conditions where the cue was uninformative. The curvilinearity of the RT – intensity curve was higher under the uncorrelated condition. Bearing this in mind, we used the Simon task to keep response-selection demands as high as possible. We additionally manipulated speed – accuracy trade-off to even more enhance van der Molen and Keuss’ effect. This manipulation was motivated by a study of van der Molen and Orlebeke (1980). They showed that while the choice RT – intensity curve for correct responses was U-shaped, the relation turned out to be monotonous for incorrect responses. They assumed that when a choice is required subjects have to suppress the instantaneous increase of readiness evoked by loud stimuli to avoid impulsive and chaotic responses.

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

Increasing error rate with increasing loudness proves that this suppression becomes more and more difficult: apparently the response choice stage is more and more frequently bypassed. We reasoned that we can manipulate these two competitive response mechanisms by speed – accuracy trade-off. It is conceivable that if speed is enhanced participants would likely take the liberty of making more reflex-like responses bypassing the choice stage. This should lead to more ‘‘lucky guesses’’, i.e. correct responses delivered before the choice process is completed. This, in turn, should make the RT – intensity relation more monotonic. Conversely, with emphasis put on accuracy, the bypassing mechanism was expected to be suppressed and a U-shaped relation should occur. It should be noted that this prediction implicitly assumes that accuracy/speed instruction can exert its effect early enough. Indeed, van der Molen and Keuss (1979) suggested that it is the response selection which is retarded at the loudest tones because no re-increase of RT was found for a go/no-go task. Some further support for this hypothesis came from an observation that RT – intensity relation is more markedly curvilinear for incompatible than for compatible trials in a Simon-like task (van der Molen and Keuss, 1981). Moreover, Rinkenauer et al. (2004) have convincingly showed that SAT instruction affected not only that portion of RT which starts after onset of lateralized readiness potential (LRP) (as argued by Osman et al., 2000 and Van der Lubbe et al., 2001) but also earlier processes. 2.1. Method 2.1.1. Participants 20 naive participants drawn from the student population of University of Poznan´ (10 males and 10 females aged 19 – 24) took part in the experiment. All were naive to the purposes of the experiment and took part in a psychophysical experiment for the first time. All had normal vision by self-report. 2.1.2. Stimuli and apparatus Visual stimuli were generated by means of two LED arrays. They consisted of 64 (8  8) red (660 nm) ultra-light LEDs and their centers were located 5.1- to the left and to the right of the fixation point (a red LED, 3.0 cd/m2). From the observation distance of 50 cm, their surfaces covered 8.5-  8.5-. The LEDs in an array were controlled by a microprocessor system which supplied every LED with a short current pulse. Luminance was changed by varying both duration and amplitude of the pulses. With this system luminance could be varied in the range 1:104 identically for every turned-on LED. Refresh rate was 110 Hz. The stimuli were two patterns of LEDs arranged as letters ‘‘A’’ or ‘‘O’’. The letters were ‘‘drawn’’ black on red, i.e. the LEDs consisting of a letter were switched off while the remaining LEDs were on. 8 luminance levels (0.12, 0.39, 1.15, 3.83, 10.7, 35.1, 108, 327 cd/m2) were used. Stimuli lasted 250 ms.

101

Each trial started with a warning signal. This was a flash of two small squares (one on each array) formed by the 4 LEDs (9.0 cd/m2) located in the center of the arrays. The duration of the foreperiods (interval between the warning and imperative stimuli) was sampled from an exponential distribution with a mean of 700 ms plus a constant period of 700 ms. The warning signal of the next trial was presented 3 s after participant’s response. Participants sat in an experimental chamber with their straight index fingers resting on the force-sensitive response keys during the whole session. A mechano-to-electrical converter was built into each key. These electrical response signals were amplified, sampled by an A/D converter, and fed to the computer. Response signals were sampled at a rate of 500 Hz starting 20 ms before stimulus onset and continuing for 1500 ms. To reduce learning effects, participants were given a short practice session, before the experiment. The results obtained in this session were excluded from further analysis. 2.1.3. Procedure In the simple-response task participants made the same response for each stimulus irrespective of letter identity and presentation side. In the Simon task, the left/right response key was assigned to ‘‘A’’/’’O’’ irrespective of stimulus side. The total number of stimuli was 288 (18 replications  2 sides  8 intensities) in the simple-response task and 576 (= 288 compatible and 288 incompatible trials) in the Simon task. Stimulus intensities, presentation sides, and intensities were arranged in random order within each block. The sequence of the two blocks corresponding to Simon task and simple task alternated between participants (AB, BA). Before each block, the participants were informed what kind of task was to be performed. Participants were unaware that the force of their responses was being measured. The session lasted about 70 min. 2.2. Speed –accuracy trade-off Participants were randomly assigned to one of two groups. One group worked with a payoff equation that emphasized speed, the other worked with an equation which emphasized accuracy. Before the session each participant was told about this emphasis and about the payoff system. After each trial, the bonus was calculated according to the following equations: for the accuracy group :

0:1IpIð1150  RTÞ 2000  1:5Ið1  pÞ Polish zlotyð ¼ 0:25$Þ

for the speed group :

0:1IpIð600  RTÞ 1000  0:075Ið1  pÞ Polish zloty;

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

where p = 1 for a correct response and p = 0 for an incorrect response, and RT denotes reaction time in ms. At the end of a given block, participants were informed about the resultant bonus (being the sum of the trial bonuses). Depending on results participants could earn about 10 –40 zloty (40 zloty = 1/30 of the average monthly salary in Poland). 2.3. Data analysis All parameters were derived from the force – time functions. Responses were classified as correct when a force of 2 N was exceeded between 100 ms and 1000 ms after stimulus onset on the required side and not on the other side. Three dependent measures were analyzed: the percentage of correctly responded trials (PC); the RT defined as the time from onset of stimulus to the moment at which the force exceeded 2.0 N relative to the baseline measured 20 ms before stimulus onset; and third, response force (RF) defined as the maximum value of the exerted force. RT and RF were determined in each trial and averaged across all correctly responded trials. The three parameters were determined separately for each stimulus intensity in each task. To evaluate data statistically analyses of variances were performed with two withinsubjects factors (intensity and task) and one betweensubjects factor (trade-off). All p-values obtained from analysis of variance were adjusted using GreenhouseGeisser coefficients. As the Simon-compatible and Simonincompatible levels were mixed within blocks, a separate analysis was performed to compare the effect of compatibility in the Simon task. Only effects of compatibility will be reported from this separate analysis 2.4. Results 2.4.1. Premature response and misses Trials in which RTs were shorter than 100 ms were defined as premature responses. As there were few such responses they were not further analyzed. Trials in which RTs were longer than 1000 or responses were not given at all were considered as misses. There were only 1.49% of misses. The differences between percent of misses among different conditions were very small and insignificant. 2.4.2. Percentage of correct choices A response made with the proper hand was considered as a correct choice. Percentages of correct choices (PC) were analyzed only for Simon task. They were higher in the accuracy than in the speed group (99.4 vs. 94.6; F(1,18) = 23.1, MSE = 8.0, p < 0.001). Significant was also the effect of intensity ( F(7,126) = 4.4, MSE = 1.6, p = 0.003, see Fig. 1): PC decreased monotonically when intensity increased. PC was higher in compatible than incompatible trials (98 vs. 96, F(1,18) = 8.8,

100

98

Percent correct

102

96

94

92

90 0.1

10

1000

Luminance (cd/m2) Fig. 1. Percent correct as a function of luminance for the simple-response and the Simon task for both groups (Experiment 1). Empty symbols denote the data of the speed-instruction group, filled symbols of the accuracyinstruction group.

MSE = 13.0, p = 0.002). However, this effect was mainly due to smaller PC for incompatible trials in the speed group (comp. 96, incomp. 93). For the accuracy group PCs for compatible and incompatible trials were equal (99%). This is supported by the significant interaction between compatibility and trade-off ( F(1,18) = 7.3, MSE = 0.8, p = 0.015). 2.4.3. Reaction time RTs were longer for the Simon than for the simpleresponse task (585 vs. 286 ms, F(1,18) = 384.8, MSE = 16204, p < 0.001) and in the accuracy than the speed group (507 vs. 479 ms, F(1,18) = 13.6, MSE = 23066, p = 0.002). Moreover, RTs decreased with intensity ( F(7,126) = 49.5, MSE = 625.6, p < 0.001). The significant interaction between task and intensity indicated ( F(7,126) =18.0, MSE = 684, p < 0.001) that this relation was different for both tasks (Fig. 2): while simple RTs decreased monotonically with intensity, the curves were U-shaped for the Simon task. A separate ANOVA performed only for the three highest intensities for the Simon task showed that RT indeed re-increased in this range ( F(2,36) = 4.2, MSE = 173, p = 0.033). The overall effect of intensity (maximal RT  minimal RT) was almost twice as large for simple RTs than for the Simon task (66 vs. 35 ms). No other interaction was found to be significant, in particular the effects of intensity did not interact with speed – accuracy trade-off. The separate ANOVA for the Simon task indicated that RTs were shorter for compatible than for incompatible trials (560 vs. 570 ms, F(1,18) = 9.5, MSE = 678, p = 0.006). Moreover, the RT –intensity curves had different shapes for both types of trials (interaction intensity  compatibility:

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

360

660

simple

Simon

320

620

280

580

240

540

200 0.1

10

1000

0.1

10

Reaction time (ms)

Reaction time (ms)

103

500 1000

Luminance (cd/m2) Fig. 2. Reaction times as function of luminance for the simple-response and the Simon task for both groups (Experiment 1). Empty symbols denote the data of the speed-instruction group, filled symbols of the accuracy-instruction group.

F(7,126) = 2.49, MSE = 2670, p = 0.02): the U-shape of the curve for compatible arrangement seems to be more pronounced (see Fig. 3). No other interaction of compatibility was significant. 2.4.4. Response force Participants responded more forcefully in the Simon than in simple-response task (see Fig. 4, note different RF axes; 12.3 vs. 10.5 N; F(1,18) = 5.9, MSE = 43.3, p = 0.026). Moreover, RFs increased with luminance ( F(7,126) = 4.8,

MSE = 1.9, p < 0.001). Fig. 4 suggests that this increase was larger for the Simon than the simple-response task but the interaction intensity  task was insignificant ( F(7,126) = 1.35, MSE = 1.5, p = 0.246). The effect of trade-off was only marginally significant in spite of the large absolute differences between conditions (see Fig. 4, 13.0 N for speed and 10.5 N for accuracy, F(1,18) = 4.16, MSE = 598, p = 0.056). No other interaction was found. The separate ANOVA for the Simon task indicated no effect of compatibility.

660

2.5. Discussion

Reaction time (ms)

620

580

540

500 0.1

10

1000

Luminance (cd/m2) Fig. 3. Reaction times as function of luminance for compatible and incompatible trials in the Simon task (Experiment 1), pooled across the two groups. Empty symbols denote incompatible trials, filled symbols compatible trials.

The general pattern of results was very similar to that obtained by van der Molen and Keuss (Keuss and van der Molen, 1982; van der Molen and Keuss, 1979; van der Molen and Orlebeke, 1980; van der Molen and Keuss, 1981) for auditory stimuli. First of all, intensity had different effects on RTs in the simple-response task and choice task (Simon paradigm). The RT –intensity curve was flatter for the Simon than for the simple-response task. Moreover, a significant increase of RT was found for the highest intensities in case of the Simon task. This is at odds with previous findings concerning the effect of brightness on RT: no effect of task difficulty on the RT – intensity relation was usually reported for visual stimuli (Azorin et al., 1995; Egeth, 1977; Everett et al., 1985; Pins and Bonnet, 1996; Schwarz et al., 1977; Schweickert et al., 1988; Van Duren and Sanders, 1988). In contrast to these previous studies we applied, however, conditions which were intended to enhance possible arousing properties of visual stimuli, i.e. we used relatively large and bright stimuli. For example, in the extensive study by Pins and Bonnet (1996) stimuli were

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

104

14.0

11.0 simple Simon

13.5

10.0

13.0

9.5

12.5

9.0

12.0

8.5

11.5

8.0 0.1

10

1000

0.1

10

Response force (N)

Response force (N)

10.5

11.0 1000

Luminance (cd/m2) Fig. 4. Peak force as a function of intensity for the simple-response and the Simon task in the two groups (Experiment 1). Empty symbols denote the data of the speed-instruction group, filled symbols of the accuracy-instruction group.

as small as 30  7.5 min of arc and at most 2  3- (in their Experiment 4) and covered a rather narrow intensity range compared to the present study (for simple RT stimulus intensities ranged from 0.18 to 58.2 cd/m2 only, and even less for choice tasks: from 0.28 to 3.09 cd/m2). With regard to RF, we expected within the frameworks of the arousal-suppression model to find a monotonic increase of RF for the simple-response task whereas this tendency should be stopped or even reversed for the more demanding tasks. On the contrary, RFs continued to grow for the brightest stimuli. The increase of RF with intensity was independent of the task. Therefore, if we assume that RF reflects immediate arousal, this finding contradicts the idea of the suppression of arousal for more demanding tasks. A gradual increase of RF with intensity was found by Angel (1973) and Ulrich et al. (1998) for a simple task, but Jas´kowski et al. (1995) reported no changes of RF with brightness. Thus, the present results seem to support Angel (1973) and Ulrich et al. (1998) findings rather than Jas´kowski et al. (1995). However, the conditions used in Experiment 1 and other studies differed remarkably. In particular, different luminance ranges and stimulus sizes were used. Namely, Jas´kowski et al. (1995) used rather small targets (0.19-) with luminance ranging from 0.3 cd/m2 to 2000 cd/m2, whereas Ulrich et al. (1998) applied unusual bright stimuli ranging from 220 cd/m2 to 22 000 cd/m2 (for comparison, the maximal luminance of a typical computer monitor is about 150 cd/m2). Furthermore, their targets were also substantially larger than the ones used by Jas´kowski et al. (5.4-). Thus, the reason for the different results obtained by different authors is that RF depends on brightness only if the target is sufficiently large and/or bright. RF was larger for the Simon task than for the simple task, which is at odds with some previous results. RF was found to

be independent of tasks in previous studies from our group (Jas´kowski et al., 2003; Van der Lubbe et al., 2001) and in a recent study by Miller et al. (1999a). Miller et al.’s subjects performed simple, go/no-go and choice tasks for auditory stimuli of different intensities.1 Even more surprising, they found no effect of task on the RT– intensity relation, in contrast to the present study and to previous findings by van der Molen and Keuss (1979, 1981) and Keuss and van der Molen (1982) as well as Waodarczyk et al. (2002). We are not able to explain why these results are at odds. To boost the effect of task on the RT– intensity relation, we applied speed – accuracy trade-off. One can expect that, with emphasis laid on accuracy, the mechanisms preventing accidental responses should be elicited rendering the RT – intensity relation more curvilinear than when speed was emphasized. Participants from the speed group made more errors and responded faster than participants from the other group. Moreover, the more demanding the task, the more effective was the trade-off as measured by the RT difference between speed and accuracy groups. These findings prove that the induction of speed –accuracy trade-off was successful. However, the instruction had no effect on the relation between RT and stimulus intensity.

3. Experiment 2 Results of Experiment 1 suggest that a negative effect of very strong stimuli on RT in demanding tasks is not 1 It should be noted that in no-go task RFs were found to be larger than for other tasks (Ulrich et al., 1999). But this effect should probably be assigned to the response- or stimulus-probability effect than to the taskcomplexity effect (Mattes et al., 2002).

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

3.1. Method 3.1.1. Participants 3 women and 7 men (mean age 22.5, S.D. 1.8) served as participants. They were recruited from the student population of Adam Mickiewicz University of Poznan´, were naive to the purposes of the experiment and took part in a psychophysical experiment for the first time. All had normal vision by self-report. 3.1.2. Stimuli and apparatus As in Experiment 1, visual stimuli were generated by means of two LED arrays. The stimuli were two neighboring LEDs arranged vertically or horizontally. These ‘‘lines’’ (1.7- in length) were ‘‘drawn’’ either black on red (‘‘large stimuli’’, i.e. the two LEDs forming a ‘‘line’’ were switched off while the remaining 62 LEDs were on) or red on black (‘‘small stimuli’’, i.e. the two LEDs forming a ‘‘line’’ were switched on while the remaining 62 LEDs were off). In other words, only 2 LEDs were flashed in case of small targets and all but two were flashed in case of large targets. Small and large stimuli were presented in a random sequence. Luminance levels and stimulus duration was identical as in Experiment 1. 3.1.3. Procedure In the simple-response task participants made the same response for each stimulus irrespective of stimulus orientation and the presentation side. In the Simon task the left/ right response key was assigned to horizontal/vertical orientation or vice versa irrespective of stimulus side. The total number of stimuli in a session was 256 (8 replications  2 sides  8 intensities  2 target sizes) in the simple-response task and 512 in the Simon task (= 256 compatible and 256 incompatible trials). Two sessions were performed for every participant on two consecutive days. Statistical analysis was the same as in Experiment 1.

3.2. Results 3.2.1. Premature response and misses Trials in which RTs were shorter than 100 ms were defined as premature responses. As there were few such responses (< 0.2%) they were not further analyzed. Trials in which RTs were longer than 1000 or responses were not given at all were considered as misses. There were only 2.0% of misses. The only effect on percent of misses found was the interaction between task and stimulus size ( F(1,9) = 6.3, MSE = 0.08, p = 0.033): for simple reactions, percent of misses was smaller for large than for small stimuli (2.42% vs. 1.25%); no such an effect of size was found for the Simon task (1.56% vs. 1.64%). 3.2.2. Percent correct Percent correct choices in the Simon task (see Fig. 7) was higher for small than for large targets (93.0 vs. 86.6, F(1,9) = 22.5, MSE = 14.6, p = 0.001). Moreover, responses were more erroneous for incompatible than for compatible trials (86.5. vs. 93.0; F(1,9) = 10.2, MSE = 3.3, p = 0.011) but there was no interaction of compatibility with other factors (Fig. 5). Insignificant was also the interaction between task and intensity which was found to be significant in Experiment 1. 3.2.3. Reaction time RTs (see Fig. 6) were longer for the Simon than for the simple-response task (260 vs. 502 ms, F(1,9) = 628, MSE = 7449, p < 0.001) and for the small than large targets (374 vs. 387 ms, F(1,9) = 25.9, MSE = 508, p = 0.001). Target size had different effects in both tasks (see Fig. 6 100

98

Percent correct

restricted to auditory stimuli but can also be observed for visual stimuli. Previous studies failed to find any effect of task complexity on the RT – intensity relation in case of visual stimuli. In Experiment 2 we examined a possible reason for this discrepancy. We suggest, following Niemi and Lehtonen (1982), that the crucial factor is the stimulus size. Niemi and Na¨a¨ta¨nen (1981) claimed that, unlike small stimuli, large stimuli posses arousing property like auditory stimuli. This suggestion was further explored in Experiment 2 in which the effects of brightness on RT and RF were compared for large and small stimuli. First, we expected that the negative effect of very bright stimuli should disappear for small stimuli or at least be weaker than for large stimuli. Second, task demands should exert no effect on RT for small stimuli. Moreover, RF should increase with brightness for large stimuli but not for small stimuli.

105

96

94

92

90 0.1

10

1000

Luminance (cd/m2) Fig. 5. Percent correct as a function of luminance for the simple-response and the Simon task and for the two target sizes (Experiment 2). Empty symbols denote data of small targets, filled symbols of large targets.

Reaction time (ms)

350

550

300

500 simple

250

200 0.1

1000 0.1

10

Simon

450

10

400 1000

Reaction time (ms)

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

106

Luminance (cd/m2) Fig. 6. Reaction time as a function of luminance for the simple-response and the Simon task for the two target sizes (Experiment 2). Empty symbols denote data of small targets, filled symbols of large targets.

compatible than incompatible trials, the three-way interaction was insignificant. Also insignificant turned out to be the interaction intensity  compatibility. 3.2.4. Response force Participants responded more forcefully in the Simon than the simple-response task (Fig. 8—note different RF axes, 11.9 vs. 9.6 N; F(1,9) = 19.0, MSE = 22, p = 0.002), to large than small targets (11.0 vs. 10.6, F(1,9) = 12.3, MSE = 1.3,

560 540

Reaction time (ms)

interaction task  size, F(1,9) = 317, MSE = 345, p < 0.001). RTs were longer to large stimuli than to small stimuli in the Simon task (526 vs. 477 ms; separate ANOVA: F(1,9) = 157.3; p < 0.001) and shorter in the simple-response task (247 vs. 272 ms; separate ANOVA: F(1,9) = 104.8; p < 0.001). Moreover, RTs depended on intensity ( F(7,63) = 106.7, MSE = 234, p < 0.001). The significant interaction between task and intensity ( F(7,126) = 4.8, MSE = 264, p = 0.007) indicated, however, that these relations were different for both tasks. As Fig. 8 shows, simple RTs decreased monotonically with intensity, nearly parallel for both target sizes. The overall change of simple RT was about 60 ms. The situation was quite different for the Simon task. For the small target, the curve was monotonic, with the overall change of RT between lowest and highest luminance being approximately equal to the change for simple RT. For the large target, the curve was U-shaped with an overall change of 17 ms and a maximal difference of 27 ms. These observations are supported by the significant interaction size  intensity (larger effect of intensity for the small than large target; F(7,63) = 8.7, MSE = 411, p = 0.001) and the three-way interaction (task  size  intensity, F(7,63) = 3.2, MSE = 204, p = 0.029). However, neither a Tukey test nor an ANOVA performed only for the four highest luminances showed any significant re-increase of RTs. The separate ANOVA for the Simon task (see Fig. 7) indicated that RTs were shorter for compatible than incompatible trials (490 vs. 514 ms, F(1,9) = 13.8, MSE = 3188, p = 0.005). The overall compatibility effect was slightly larger for the small than for large target (29 vs. 17 ms, F(1,9) = 6.8, MSE = 378, p = 0.028). Although, similarly to Experiment 1, the re-increase of RTs for the highest intensities looked more pronounced for

520 500 Simon 480 460 440 420 0.1

10

1000

Luminance (cd/m2) Fig. 7. Reaction times as function of luminance for compatible and incompatible trials in the Simon task and for small and large targets, pooled across the two groups. (Experiment 2). Empty symbols denote incompatible trials, filled symbols compatible trials.

12.0

14.0

11.5

13.5

11.0

13.0 Simon

10.5

10.0

12.0

simple

9.5

9.0 0.1

12.5

107

Response force (N)

Response force (N)

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

11.5

10

1000

0.1

11.0 1000

10

Luminance (cd/m2) Fig. 8. Peak force as a function of luminance for the simple-response and the Simon task for the two target sizes (Experiment 2). Empty symbols denote data of small targets, filled symbols of large targets.

The separate ANOVA for the Simon task (see Fig. 9) indicated no effect of compatibility.

p = 0.007) and to brighter stimuli ( F(7,63) = 5.3, MSE = 1.5, p = 0.016). Moreover, the effect of target size was slightly greater for the Simon than the simple-response task (interaction size  task F(1,9) = 1.9, MSE = 1.4, p = 0.054). Target size differentially affected the relation RF – intensity, which was steeper for large than small targets (interaction size  intensity F(7,63) = 2.5, MSE = 0.3, p = 0.039). ANOVAs performed for each size separately showed no effect of intensity for the small target ( p = 0.12) and a significant effect for the large target ( F(7,63) = 5.77, MSE = 0.44, p = 0.011). The interaction of task  intensity and the threefold interaction were insignificant.

3.3. Discussion In Experiment 1, we found a clear effect of stimulus intensity on RF and a modulating effect of task difficulty on the shape of the RT – intensity relation. Both these effects might be elicited by the arousal properties of strong and large visual targets. Experiment 2 was designed to check this possibility by comparing the effect of brightness on RT and RF for large and small visual stimuli. This experiment yielded two main findings. First, task demands clearly

14.0

Response force (N)

13.5

13.0 large 12.5 small 12.0

11.5

11.0 0.1

10

1000

0.1

10

1000

Luminance (cd/m2) Fig. 9. Peak force as a function of luminance for compatible and incompatible trials in the Simon task for small and large targets (Experiment 2). Empty symbols denote incompatible trials, filled symbols compatible trials.

108

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

affected shapes of the RT – intensity curves for large targets like in Experiment 1, whereas these curves were almost identical between tasks for small targets. Second, the effect of intensity on RF was likewise found for large targets only, whereas RF –intensity relations were virtually flat for small targets. Both these findings can be accounted for by assuming different arousing properties of large and small targets. This is exactly what we expected. Although we found clear modulating effect of task on RT – intensity relation, re-increase of RTs for the brightest stimuli and Simon task was found in contrast to results of Experiment 1. It should be noted that under some special conditions van der Molen and Keuss reported also monotonic rather than U-shaped curves even for choice tasks. Monotonic relations were found when a preknowledge about the stimulus intensity was given before each trial and for very long foreperiod (30 s). Such an ‘‘instability’’ of the RT – intensity relation might indicate that the shape of the relation for difficult tasks is prone to the influences of some other factors like motivation and general stress induced by the experimental situation. A factor, which could preclude re-increase of RTs, might be mixing of large and small stimuli (e.g. by lowering the overall arousal). Alternatively, the luck of a clear reincrease might be due to low statistical power of Experiment 2. Indeed, to keep the session length reasonable we had to lower the number of trial for a given condition (16 in case of simple RT and 32 in case of Simon task). Bearing in mind that RTs in Simon task were relatively variable, it is not surprising that the re-increase we found (Fig. 7) did not reach significance level. Of considerable interest is that RTs were longer for the large than small target in case of the Simon task, whereas the reversed relation was noted in the simple task. This effect could also be accounted for by assuming arousing properties of large stimuli. In a difficult task, arousing properties of the stimulus should result in delayed RT because inhibitory processes have to be triggered to prevent premature and inaccurate responses. In contrast, arousing properties are welcome when no choice is necessary and when the response can be initiated without complete stimulus recognition. Alternatively, shortening of RTs for the simple task could partially be due to shorter perceptual latency for central than peripheral stimulation (Payne, 1966, 1967; Rains, 1963). Indeed, the inner LEDs of the large target were located more centrally than the two LEDs switched on in case of the small target. This might facilitate reactions, as detection of the light emitted by the border LEDs was sufficient to respond in case of the simple task. So far we used the term brightness and stimulus intensity interchangeably. This usage could be somewhat misleading especially for readers representing more psychophysiological tradition. Indeed, Barry and James (1981), for example, manipulated stimulus intensity by changing the target size. In psychophysical tradition, however, brightness reflects

subjective feeling of how bright a target is. It is measurable by using the magnitude-estimation method in which participants assign different ranks to stimuli of different perceived brightness. Systematic measurements of this kind were made by Mansfield (1973). He showed that once the target area exceeded a critical value perceived brightness did not depend on the size anymore being a function of luminance only. According to his results, the critical target area is as small as 0.17- (in diameter). Only for very small targets perceived brightness was found to be a function of the product of target area and luminance (luminous flux). It is, however, clear from our data that RF depends on combined effect of luminance and size. Of interest would be to see if RF changes as a function of luminous flux. Although we did not vary target area systematically, we could plot RF as a function of luminous flux bearing in mind that the large stimuli had 31 times larger lighting area than the small stimuli. The curves for large and small targets should overlap. We found that they did not (not shown). It means that the relationship between RF, target luminance and target size must be more complicated.

4. General discussion 4.1. Van der Molen and Keuss’ effect for visual stimuli In two experiments we explored the intensity effect on reaction time and response force for visual stimuli. The point of departure for this study was van der Molen and Keuss’ finding (Keuss and van der Molen, 1982; van der Molen and Keuss, 1979; van der Molen and Orlebeke, 1980; van der Molen and Keuss, 1981) that task difficulty may modify the RT –intensity relation in case of auditory stimulation: while for a simple task RTs gradually decreased as a function of intensity, a re-increase of RTs was observed for the loudest tones when a choice task was to be accomplished. This effect was accounted for by arousing properties of loud auditory stimuli. It was assumed that in case of simple tasks immediate arousal evoked by strong auditory signals has a beneficial effect on reaction times. If, however, the task is more complex (needs a choice) and stimuli are very loud subjects try ‘‘to suppress the impulsive prompting of a motor response’’ (van der Molen and Orlebeke, 1980, p. 475) to avoid too many premature and/or wrong responses. Our main goal was to check if a similar pattern of results occurs for visual stimuli provided that they are arousing. Based on Niemi and Lehtonen (1982) findings we expected a modulatory effect of task on the RT –intensity relation when visual targets are large enough. Indeed, the shape of RT –intensity relation depended on task demands when the stimuli were relatively large. For small targets the relation was monotonic and almost identical with that for the simple task. Furthermore, stimulus size did not affect the relation when a simple task was to be done. The relation changed

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

dramatically once participants have to respond to large targets and the task was demanding. In such a case, besides flattening of RT –intensity curve, we observed a re-increase of RTs for the brightest stimuli, similar to that observed by van der Molen and Keuss for auditory stimuli. This is the first demonstration of the van der Molen and Keuss’ effect for visual stimuli. 4.2. The effect of arousal The U-shaped intensity –RT relation for large and not for small targets provides an important support for van der Molen and Keuss’ claim that arousal is responsible for the RT re-increase for very intense stimuli. Indeed, visual stimuli commonly used in experimental psychology evoke no arousal and, therefore, no effect of task on RT – intensity relation was found so far. Arousing properties were, however, reported for large (Niemi and Lehtonen, 1982) or extremely bright stimuli (Sanders, 1975). In accordance with this claim we found the effect of task on the RT – intensity relation only for large but not for small stimuli. Taken together, van der Molen and Keuss’ results as well as the present findings provided converging evidence that in case of demanding tasks arousal can modify the RT – intensity curve. How does arousal exert its effect on RT? Following Sanders (1977), van der Molen and Orlebeke (1980) seem to accept that arousal affects readiness to respond. In Sanders’ later writing this idea is expressed as follows (Sanders, 1983, pp. 89– 90) ‘‘the rise in arousal triggers a signal to the activation system, enhancing response readiness’’. Also Niemi and Na¨a¨ta¨nen (1981) and Na¨a¨ta¨nen (1971) considered immediate arousal as a factor influencing preparation/readiness. Usually, response preparation is modeled as the distance between motor readiness and a threshold called action limit (e.g. Niemi and Na¨a¨ta¨nen, 1981). If readiness increases, the distance to action limit decreases. The overt response is executed if motor readiness crosses a threshold level. The more advanced is motor preparation, the smaller becomes the distance between the current value of readiness and the motoraction limit. Therefore, better motor preparation means smaller distance between motor readiness and the action limit. Therefore, reaction time is short for well-prepared responses. Accordingly, one can assume that readiness increases instantaneously whenever an arousing stimulus (e.g. a loud tone) is presented. Such an instantaneous increase of readiness must somehow be suppressed if a choice is required. This, in turn, means a remarkable increase of the distance between motor activation and action limit, leading to RT lengthening. This model might successfully account for the U-shaped RT – intensity relationship reported for arousing stimuli by van der Molen and Keuss and in the present study. But as we will show in the following, this proposal seems to be inconsistent with force data.

109

4.3. Response force 4.3.1. Motor readiness and response force Na¨a¨ta¨nen’s (1971) model of motor readiness was recently extended by Mattes et al. (1997) to account for RF results. They assumed that RF is directly related to the maximal overshoot of motor activation over the action limit. Furthermore, they assumed that the larger the distance from the current motor activity at the moment of response initiation to the action limit, the larger the overshoot. In other words, in case of a badly prepared response the distance between readiness prior to stimulus onset is large and activation overshoots the action limit by a large amount, resulting in a forceful response. In contrast, if a response is well prepared the distance to the action limit is small and ‘‘the required increment of activation can be calibrated well, producing only a small activation overshoot and consequently a less forceful response’’ (Mattes et al., 2002, p. 479). This nicely fits their results on the effect of response probability on RT as they found that participants responded faster and weaker when the response probability was high. 4.3.2. Arousal and response force However, as mentioned in the Introduction, some evidence indicates that both general and immediate arousal can affect RF. Unfortunately, Mattes et al. (1997; see also Mattes et al., 2002) provided no conception of how to model the problem of the arousal effects on RF. If we assume, as it was done in the preceding section, that arousal affects readiness by decreasing the distance between motor activation and action limit, the arousing stimuli should be associated with weaker responses. This is obviously not the case.2 In the following we draft another model in which we supplemented an elaborated version of Mattes et al.’s overshoot idea by a possible mechanism of arousal effects. The assumptions of the model are the following. (i) Arousing factors are assumed to increase general internal noise which entails larger variability of all motor processes. For simplification we will assume that arousing factors affect only variability of the action-limit level. As Jas´kowski et al. (2000) pointed out, a majority of RF results could be accounted for by an assumption that immediate and general arousal lead to an increase of overall variability of neural events. Such a mechanism was originally proposed by Van Galen and de Jong (1995; see also Van Galen and van Huygevoort, 2000; Van Gemmert and Van Galen, 1997) to account for changes of axial pen pressure during aiming movements in response to increased mental load, physical factors or task demands (Van den Heuvel et al., 1998; Van Galen and de Jong, 1995; Van Galen and van Huygevoort, 2000; Van Gemmert and Van Galen, 1997). Van Galen and 2 It should be mentioned that Mattes et al. (2002) recently reported data which were inconsistent with the original formulation of the model. They showed, in contrary to the predictions of the model, that also stimulus probability affects RT. This forced a post-hoc reformulation of the model.

110

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

de Jong (1995) postulated that arousing/stressing factors increase the general neuromotor noise which leads to an unacceptable rate of premature and/or wrong participants’ responses. To prevent such a situation, limb stiffness is increased by contraction of both agonist and antagonist muscles, resulting in an increase of axial pen pressure. (ii) To account for our data we also have to assume that increase of internal noise is very fast. Many models of arousal had to assume fast response of the system for the abrupt changes in the environment. Sanders (1983) introduced a term ‘‘immediate arousal’’ to describe dynamical properties of some arousing factors like stimulus intensity which can affect response speed in a trial-by-trial fashion. Also Van Gammert and van Galen assumed that very load stimuli might evoke an immediate stiffness response, possibly like startle response in animal conditioning. (iii) Response force is directly related to motor activation. (iv) After stimulus onset, motor activation increases, crossing eventually the action limit. We will assume that amount of motor activation over the action limit must be proportional to its variability. In other words, the signal-tonoise ratio (i.e. maximal motor activation over action limit related to the variability of action limit) has to be constant. Hence, we can write ð M  AÞ=rð AÞ ¼ a or RF ¼ f ð M Þ ¼ f ðaIrð AÞ þ AÞ; where r(A) denotes variability of action limit, A is the action limit, M is maximal motor activation, f is a monotonically increasing function relating to motor activation to RF, and a is a constant. (v) We will also relate response preparation to the lowering of the action limit rather than to the level of motor readiness: better preparation means, as before, smaller distance between motor readiness and the action limit. The model provides therefore the straightforward account for two basic empirical facts concerning RF. (1) An increase of arousal level results in more forceful responses. This is because of increase of r(A). (2) Since preparation leads to a reduction of action limit, A, better prepared responses are weaker. In accordance with our previous considerations, we will assume that when a choice is needed the distance between motor readiness and action limit has to be high enough to avoid incorrect responses. If an arousing stimulus is presented, the increase of arousal level results in the related increase of action-limit variability. Accordingly, action limit must be elevated because of a risk of incorrect responses. This leads to two consequences. First, RF increases as both r(A) and A increases. Second, RT may increase because of the elevation of action limit. Note also that RF should be smaller for easy than for complex tasks only if stimuli are arousing. This is because

for a non-arousing situation there is no need to elevate the action limit. Conversely, under arousing conditions, action limit has to be elevated if percent of correct responses must be kept on a reasonable level. In such a situation, RF should be larger for choice tasks than for a simple task. In the present experiments, we found larger RFs for the Simon than detection task and this difference was larger for the big target (more arousing stimulus) than for the small target (less arousing stimulus). Failure to find task effect on RF by other authors (Miller et al., 1999a; Van der Lubbe et al., 2001) could be due to non-arousing stimuli they used. 4.3.3. Speed – accuracy trade-off Our main prediction concerning manipulation of speed – accuracy trade-off was that the van der Molen and Keuss’ effect should vanish or be at least less pronounced under speed condition. This prediction relied on the results reported by van der Molen and Orlebeke (1980) who showed that the re-increase of RT for very loud stimuli does not occur for incorrect responses which were assumed to bypass the choice mechanisms. We expected that this mechanism is more frequently bypassed under speed condition than under accuracy condition leading to higher error rate and lucky guesses whose contribution to the mean should alleviate the curvilinearity of the RT – intensity relations. Instead, we found virtually parallel RT –intensity relations for the both conditions. There are some possibilities why SA instruction failed to affect RT – intensity relationship. The most trivial possibility is that the number of lucky guesses and, consequently, their contribution to the RT mean was too subtle to be detected in our experiment. This possibility seems to be supported by the fact that our speed – accuracy manipulation was relatively ineffective. RT was only by 15% shorter in speed- than accuracy group for the Simon task. Another possibility is that the action limits are more or less equal under both conditions. Although under accuracy condition r(A)/A has to be lower because of the risk of too many errors, speed condition could be by itself more arousing. It is, for example, a common finding that RF is larger under time pressure (Jas´kowski et al., 1994a,b,c, 2000; Van der Lubbe et al., 2001). Also in the present study, participants responded stronger under speed than under accuracy condition. This extra arousal under speed condition might force some additional elevation of action limit which can luckily reach the same level as for accuracy condition. Whether this leads to parallel RT – intensity functions for the both conditions or not depends heavily on the relation between arousal and stimulus strength. 4.4. Summary The main results of the present study can be summarized as follows. (1) RT – intensity relation depends on task difficulty under arousing conditions: for a more demanding

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

task like Simon paradigm RT – intensity curve becomes Ushaped rather than monotonic. This finding, originally demonstrated by Van der Molen and Keuss for auditory stimulation, was extended here to visual stimuli. Additionally, it was shown, in accordance to previous findings, that only large stimuli have such arousing properties. (2) With large visual targets, RF increases monotonously with intensity and this relation is independent of task. (3) In more demanding tasks participants responded more forcefully. (4) Speed –accuracy instruction does not affect the RT/ RF – intensity relation. However, the relation is less markedly U-shaped for fast than for slow responses. All these findings can be accounted for by the outlined model.

Acknowledgements The authors would like to thank Rolf Verleger and Rob van der Lubbe for their helpful comments and suggestions.

References Angel, A., 1973. Input – output relations in simple reaction time experiments. The The Quarterly Journal of Experimental Psychology 25, 193 – 200. Azorin, J.M., Benhaim, P., Hasbroucq, T., Passamd’, C.A., 1995. Stimulus preprocessing and response selection in depression: a reaction time study. Acta Psychologica 89, 95 – 100. Barry, R.J., James, A.L., 1981. Fractionation of phasic responses in dishabituation paradigm. Physiology & Behavior 26, 69 – 75. Bertelson, P., 1969. The time-course of preparation: confirmatory results with visual and auditory warning signal. Acta Psychologica 30, 145 – 154. Egeth, H.E., 1977. Attention and preattention. In: Bower, G.H. (Ed.), The Psychology of Learning and Motivation. Academic Press, New York, pp. 277 – 320. Everett, B.L., Hochhaus, L., Brown, J.R., 1985. Letter-naming as a function of intensity, degradation, S – R compatibility, and practice. Perception & Psychophysics 37, 467 – 470. Jas´kowski, P., 1996. Simple reaction time and perception of temporal order: dissociations and hypotheses. Perceptual and Motor Skills 82, 707 – 730. Jas´kowski, P., 1999. Reaction time and temporal order judgment: the problem of dissociations. In: Aschersleben, G., Bachmann, T., Mu¨sseler, J. (Eds.), Cognitive Contributions to the Perception of Spatial and Temporal Events. Elsevier, Amsterdam, pp. 265 – 287. Jas´kowski, P., Waodarczyk, D., 1997. Effect of sleep deficit, knowledge of results, and stimulus quality on reaction time and response force. Perceptual and Motor Skills 84, 563 – 572. Jas´kowski, P., Pruszewicz, A., Swidzin´ski, P., 1990. VEP latency and some properties of simple motor reaction-time distribution. Psychological Research 52, 28 – 34. Jas´kowski, P., Rybarczyk, K., Jaroszyk, F., 1994a. On the relationship between latency of auditory evoked potentials, simple reaction time and stimulus intensity. Psychological Research 56, 59 – 65. Jas´kowski, P., Verleger, R., Wascher, E., 1994b. Response force and reaction time in a simple reaction task under time pressure. Zeitschrift fur Psychologie 202, 405 – 413. Jas´kowski, P., Wro´blewski, M., Hojan-Jezierska, D., 1994c. Impending electrical shock can affect response force in a simple reaction task. Perceptual and Motor Skills 79, 995 – 1002.

111

Jas´kowski, P., Rybarczyk, K., Jaroszyk, F., Leman´ski, D., 1995. The effect of stimulus intensity on force output in simple reaction time task in humans. Acta Neurobiologiae Experimentalis 55, 57 – 64. Jas´kowski, P., Wauschkuhn, B., Van der Lubbe, R.H.J., Wascher, E., Verleger, R., 2000. The influence of time pressure and expectancy on response force in an S1 – S2 paradigm. Acta Psychologica 105, 89 – 105. Jas´kowski, P., Werner, I., Verleger, R., 2003. On the translation of some stimulus features to response force. Acta Neurobiologiae Experimentalis 63, 327 – 335. Kammer, T., Lehr, L., Kirschfeld, K., 1999. Cortical visual processing is temporally dispersed by luminance in human subjects. Neuroscience Letters 263, 133 – 136. Kaswan, J., Young, S., 1965. Effect of luminance, exposure duration and task complexity on reaction time. Journal of Experimental Psychology 69, 393 – 400. Keuss, P.J.G., van der Molen, M.W., 1982. Positive and negative effects of stimulus intensity in auditory reaction time tasks: further studies on immediate arousal. Acta Psychologica 52, 61 – 72. Mansfield, R.J.W., 1973. Brightness function: effect of area and duration. Journal of the Optical Society of America 63, 913 – 920. Mattes, S., Ulrich, R., Miller, J.O., 1997. Effects of response probability on response force in simple RT. The Quarterly Journal of Experimental Psychology 50A, 405 – 420. Mattes, S., Ulrich, R., Miller, J., 2002. Response force in RT tasks: isolating effects of stimulus probability and response probability. Visual Cognition 9, 451 – 477. Meyer, D.E., Osman, A., Irwin, D.E., Yantis, S., 1988. Modern mental chronometry. Biological Psychology 26, 3 – 67. Miller, J., Ulrich, R., Pfaff, K., 1991. Visual Stimulus Intensity Does Not Influence Response Force 32nd Annual Meeting of the Psychonomic Society: San Francisco, CA. Miller, J., Franz, V., Ulrich, R., 1999a. Effects of auditory stimulus intensity on response force in simple, go/no-go, and choice RT tasks. Perception & Psychophysics 61, 107 – 119. Miller, J., Ulrich, R., Rinkenauer, G., 1999b. Effects of stimulus intensity on the lateralized readiness potential. Journal of Experimental Psychology. Human Perception and Performance 25, 1454 – 1471. Na¨a¨ta¨nen, R., 1971. Non-aging fore-period and simple reaction time. Acta Psychologica 35, 316 – 327. Na¨a¨ta¨nen, R., 1995. The mismatch negativity: a powerful tool for cognitive neuroscience. Ear and Hearing 16, 6 – 18. Niemi, P., 1979. Stimulus intensity effects on auditory and visual reaction processes. Acta Psychologica 43, 299 – 312. Niemi, P., Lehtonen, E., 1982. Foreperiod and visual stimulus: a reappraisal. Acta Psychologica 50, 73 – 82. Niemi, P., Na¨a¨ta¨nen, R., 1981. Foreperiod and simple reaction time. Psychological Bulletin 89, 133 – 162. Osman, A., Lou, L., Muller-Gethmann, H., Rinkenauer, G., Mattes, S., Ulrich, R., 2000. Mechanisms of speed-accuracy tradeoff: evidence from covert motor processes. Biological Psychology 51, 173 – 199. Payne, W.H., 1966. Reaction time as a function of retinal location. Vision Research 6, 729 – 732. Payne, W.H., 1967. Visual reaction time on a circle about the fovea. Science 155, 481 – 482. Pins, D., Bonnet, C., 1996. On the relation between stimulus intensity and processing time: Pieron’s law and choice reaction time. Perception & Psychophysics 58, 390 – 400. Rains, J.D., 1963. Signal luminance and position effects in human reactions time. Vision Research 3, 239 – 251. Rinkenauer, G., Osman, A., Ulrich, R., Mattes, S., 2004. On the locus of speed-accuracy tradeoff in reaction time: inferences from the lateralized readiness potential. Journal of Experimental Psychology. General 133, 261 – 282. Sanders, A.F., 1975. The foreperiod effect revisited. The Quarterly Journal of Experimental Psychology 27, 591 – 598.

112

P. Jas´kowski, D. Wlodarczyk / International Journal of Psychophysiology 61 (2006) 98 – 112

Sanders, A.F., 1977. Structural and functional aspects of the reaction process. In: Dornic, S. (Ed.), Attention and Performance. Erlbaum, Hilsdale, NJ, pp. 3 – 25. Sanders, A.F., 1980. Stage analysis of reaction processes. In: Stelmach, G.E., Requin, J. (Eds.), Tutorials in Motor Behavior. North-Holland, Amsterdam, pp. 331 – 354. Sanders, A.F., 1983. Towards a model of stress and human performance. Acta Psychologica 53, 61 – 97. Sanders, A.F., Wertheim, M., 1973. The relation between physical stimulus properties and the effect of foreperiod duration on reaction time. The Quarterly Journal of Experimental Psychology 25, 201 – 206. Schwarz, S., Pomerantz, J.R., Egeth, H.E., 1977. State and process limitations in information processing: an additive factors analysis. Journal of Experimental Psychology. Human Perception and Performance 3, 403 – 410. Schweickert, R., Dahn, C., McGuigan, H., 1988. Intensity and number of alternatives in hue identification: Pieron’s law and choice reaction time. Perception & Psychophysics 44, 383 – 389. Stanovitch, K.E., Pachella, R.G., 1977. Encoding, stimulus – response compatibility, and stages of processing. Journal of Experimental Psychology: Human Perception and Performance 3, 411 – 421. Ulrich, R., Rinkenauer, G., Miller, J.O., 1998. Effects of stimulus duration and intensity on simple reaction time and response force. Journal of Experimental Psychology. Human Perception and Performance 24, 915 – 928. Ulrich, R., Mattes, S., Miller, J., 1999. Donder’s assumption of pure insertion: an evaluation on the basis of response dynamics. Acta Psychologica 102, 43 – 75. Van den Heuvel, C.E., Van Galen, G.P., Teulings, H.-L., Van Gemmert, A., 1998. Axial pen force increases with processing demands in handwriting. Acta Psychologica 102, 43 – 75. Van der Lubbe, R.H.J., Jas´kowski, P., Wauschkuhn, B., Verleger, R., 2001. Influence of time pressure in a simple detection task, a choice-bylocation task, and the Simon task. Journal of Psychophysiology 15, 241 – 255. van der Molen, M.W., Keuss, P.J.G., 1979. The relationship between reaction time and intensity in discrete auditory tasks. The Quarterly Journal of Experimental Psychology 31, 95 – 102.

van der Molen, M.W., Keuss, P.J.G., 1981. Response selection and the processing of auditory intensity. The Quarterly Journal of Experimental Psychology 33, 177 – 184. van der Molen, M.W., Orlebeke, J.F., 1980. Phasic heart rate change and the U-shaped relationship between choice reaction time and auditory signal intensity. Psychophysiology 17, 471 – 481. Van Duren, L.L., Sanders, A.F., 1988. On the robustness of the additive factors stage structure in blocked and mixed choice reaction designs. Acta Psychologica 69, 83 – 94. Van Galen, G.P., de Jong, W.P., 1995. Fitts’ law as the outcome of a dynamic noise filtering model of motor control. Human Movement Science 14, 539 – 571. Van Galen, G.P., van Huygevoort, M., 2000. Error, stress and the role of neuromotor noise in space oriented behaviour. Biological Psychology 51, 151 – 171. Van Gemmert, A.W.A., Van Galen, G.P., 1997. Stress, neuromotor noise, and human performance: a theoretical perspective. Journal of Experimental Psychology. Human Perception and Performance 23, 1299 – 1313. Vaughan Jr., H.G., Costa, L.D., Gilden, I., 1966. The functional relation of visual response and reaction time to stimulus intensity. Vision Research 6, 645 – 656. Verleger, R., 1988. Event-related potentials and cognition: a critique of the context updating hypothesis and an alternative interpretation of P3. Behavioral and Brain Sciences 11, 343 – 356. Verleger, R., 1998. Towards an integration of P3 research with cognitive neuroscience. Behavioral and Brain Sciences 21, 150 – 154. Wilson, A.J., Lit, A., 1981. Effects of photopic annulus luminance level on reaction time and on the latency of evoked cortical potential responses to target flashes. Journal of the Optical Society of America 71, 1481 – 1486. Waodarczyk, D., Jas´kowski, P., Nowik, A., 2002. The influence of sleep deprivation and auditory intensity on reaction time and response force. Perceptual and Motor Skills 94, 1101 – 1112.