Acta Psychologica 116 (2004) 245–262 www.elsevier.com/locate/actpsy
Being prepared on time: on the importance of the previous foreperiod to current preparation, as reflected in speed, force and preparation-related brain potentials Rob H.J. Van der Lubbe a,b,*, Sander A. Los c, Piotr Jaskowski b,d, Rolf Verleger b a
Psychological Laboratory, Helmholtz Instituut, Utrecht University, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands b Department of Neurology, University of L€ubeck, Ratzeburger Allee 160, 23538 L€ubeck, Germany c Cognitive Psychology, Vrije Universiteit, De Boelelaan 1111, 1081 HV Amsterdam, The Netherlands d Department of Psychophysiology, Kazimierz Wielki University, ul. Staffa 1, 85-867 Bydgoszcz, Poland Received 20 June 2003; received in revised form 17 March 2004; accepted 17 March 2004 Available online 27 April 2004
Abstract How do participants adapt to temporal variation of preparatory foreperiods? For reaction times, specific sequential effects have been observed. Responses become slower when the foreperiod is shorter on the current than on the previous trial. If this effect is due to changes in motor activation, it should also be visible in force of responses and in EEG measures of motor preparation, the contingent negative variation (CNV) and the lateralized readiness potential (LRP). These hypotheses were tested in a two-choice reaction task, with targets occurring 500, 1500, or 2500 ms after an acoustic warning signal. The reaction time results showed the expected pattern and were accompanied by similar effects on a fronto-central CNV and the LRP. In contrast, the increase of response force with brief current foreperiods did not depend on previous foreperiods. Thus, EEG measures confirm that sequential effects on RT are at least partially due to changes in motor activation originating from previous trials. Effects found on response force may be related to general response readiness rather than activation of motor-hand areas, which may explain the absence of a sequential effect on force in the current experiment. 2004 Elsevier B.V. All rights reserved. *
Corresponding author. Address: Psychological Laboratory, Helmholtz Instituut, Utrecht University, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands. Tel.: +31-30-2534582; fax: +31-30-2534511. E-mail address:
[email protected] (R.H.J. Van der Lubbe). 0001-6918/$ - see front matter 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.actpsy.2004.03.003
246
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
PsycINFO classification: 2330; 2340 Keywords: Foreperiod; Preparation
1. Introduction Recently, there has been growing interest in the influence of the temporal anticipation of a stimulus on the speed and force of the response to that stimulus. The anticipatory activity that is reflected by these measures is generally believed to be intentional in nature, because it can be affected by instructional cues, which tell the participant with some degree of reliability when the stimulus is due to appear (see Nobre, 2001 for a review). However, recent developments suggest that the speed of responding is also affected by an unintentional mechanism during stimulus anticipation, which reflects effects of the temporal layout of events that occurred on preceding trials (Los, Knol, & Boers, 2001; Los & Van den Heuvel, 2001). The purpose of the present article is to examine whether similar influences are reflected on the force of the response, and also on preparation-related brain potentials. This inquiry should also help to further elucidate the complex relationship that has been observed between the speed and force of responding (e.g. see Jaskowski, Van der Lubbe, Wauschkuhn, Wascher, & Verleger, 2000; Mattes & Ulrich, 1997; Mattes, Ulrich, & Miller, 1997). To study anticipatory behavior experimentally, many studies have varied the foreperiod (FP), the time interval between a warning signal and a subsequent target. When the FP varies from trial to trial with an equal probability of occurrence (a rectangular distribution), the classical finding is that mean reaction time (RT) decreases as FP increases (the foreperiod effect; Bertelson & Tisseyre, 1968; Woodrow, 1914; Wundt, 1887). A less well known but possibly related finding is that the force exerted on response buttons also decreases as FP increases (e.g. Jaskowski & Verleger, 1993; Mattes & Ulrich, 1997). The latter finding may indicate that the FP effect is located at a motoric level of processing, which is supported by studies using the additive factors method (Sanders, 1980), reflex amplitudes (Brunia, 1983), transcranial magnetic stimulation (Hasbroucq et al., 1999), functional magnetic resonance imaging, and positron emission tomography (Coull & Nobre, 1998). In spite of this evidence, the locus of the FP effect has recently become the subject of controversy. M€ uller-Gethmann, Ulrich, and Rinkenauer (2003) examined the influence of FP duration on the time of onset of the lateralized readiness potential (LRP), which is generally considered as a valid index of the starting point of motor preparation (De Jong, Wierda, Mulder, & Mulder, 1988; Gratton, Coles, Sirevaag, Eriksen, & Donchin, 1988; see Eimer, 1998 for a review). They observed that FP duration, when varied between blocks of trials, affects the interval between target onset and LRP onset, and not the interval between LRP onset and the release of the response, which suggests a premotoric locus of the FP effect (see Osman & Moore, 1993). In a simultaneous and independent study, Tandonnet, Burle, Vidal, and Hasbroucq (2003) confirmed this finding with respect to the LRP, but still favored a motoric locus of the FP effect in view of the results of a fine grained analysis where
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
247
motoric activation was studied independently for the left and right hemisphere. Specifically, they observed that the interval between the onset of a positive amplitude on the ipsilateral hemisphere and the release of the response was unaffected by FP, but that the interval between the onset of a negative amplitude on the contralateral hemisphere and the release of the response was reliably affected by FP. The authors argued that because ipsilateral positivity determines LRP onset, this measure obscures the critical effect of FP on the onset of contralateral negativity. Apart from this controversy on the locus of the FP effect, there has recently been interest in the mechanism underlying the effects of FP on RT and response force. The standard model of the FP effect holds that the classical FP–RT function is due to an increasing ‘‘subjective expectancy’’ about the moment of target presentation, which is assumed to guide nonspecific motor activation to that moment (N€a€at€anen, 1971). Motor activation is low in case of low expectancy for the target, and, therefore, its distance to the motor threshold to execute the response is relatively large. As a result, low expectancy will lead to slow responding, because of the relatively large distance that motor activation should bridge to reach the motor threshold. Recently, Mattes and Ulrich (1997) (see also Mattes et al., 1997) extended this model to explain both the effects found on reaction time and response force. In particular, they suggested that a large difference between initial activation and the motor threshold should lead to a larger overshoot above the motor threshold, and the larger the overshoot, the more forceful the response. Thus, in case of low ‘‘expectancy’’ for a short FP, the time needed to reach the motor threshold will be long, and the overshoot above the threshold will be large, which results in slow and forceful responses. An important specification of the classical FP–RT function is that RT is not only dependent on the FP of the current trial, but also on the FP of the immediately preceding trial (e.g. Drazin, 1961; Los et al., 2001; Los & Van den Heuvel, 2001; Woodrow, 1914). Specifically, relative to an intertrial repetition of FPs, responding will be slow when the FP on the previous trial was longer than the FP on the current trial, but of about the same speed when the FP on the previous trial was shorter than the FP on the current trial. Importantly, but often ignored, these sequential effects underlie the main FP effect on RT. The average for trials with the shortest FP consists of a large proportion of trials with a longer FP on a previous trial, which delays responding, whereas the average for trials with the longest FP consists of a large proportion of trials with a shorter FP on a previous trial, which leaves RT relatively unaffected. Apart from the effect of the preceding trial, there may also be some residual influence of even earlier preceding trials (e.g. Los et al., 2001). These findings indicate that one should focus on sequential effects to grasp the FP effect on RT. Los and colleagues (Los, 1996; Los et al., 2001; Los & Van den Heuvel, 2001) proposed that sequential effects may be caused by unintentional changes in motor activation based on the principles of trace conditioning, thereby facilitating responses at the moments at which the target occurred on previous trials. Alternatively, it may be proposed that the sequential effects are a result of subjective intentions (i.e. intentional motor preparation) that vary from trial to trial (see also Jentzsch & Sommer, 2002a, 2002b, for the relevance of this distinction in explaining patterns of sequential effects in choice–response tasks without FPs). That is, participants simply may have
248
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
a preference to respond at the same moment as in the previous trial. The asymmetry of sequential effects can be accounted for by assuming that on intertrial alternations of FP, participants are still capable to correct their wrong temporal preference when FP turns out to be longer than on the previous trial, but not vice versa (Niemi & N€ a€ at€ anen, 1981; Requin, Brener, & Ring, 1991). Support for the possible relevance of intentional preparation comes from studies in which targets were preceded by temporal cues that indicated the likely moment of appearance of the target (Coull, Frith, B€ uchel, & Nobre, 2000; Coull & Nobre, 1998; Miniussi, Wilding, Coull, & Nobre, 1999; Nobre, 2001). Responses were faster after valid than after invalid temporal cues. However, Los and Van den Heuvel (2001) presented evidence for an unintentional account of the sequential effect, showing that the sequential effect was present in case of infrequent invalid temporal cues. Regardless of the origin of sequential effects, a slight modification of the model of Mattes and Ulrich (1997) may explain the corresponding force data: When ‘‘expectancy’’ changes on a trial-by-trial basis either because of unintentional or intentional processes, effects found on RT and response force may be explained by the distance between motor activation and the motor threshold, and the amount of overshoot above the motor threshold. We will call this model the extended overshoot model in the following. Thus, the extended overshoot model predicts that any manipulation that affects the level of preparation should have corresponding effects on RT and response force. That is, any manipulation that increases the level of preparation should decrease both RT and force, whereas any manipulation that decreases the level of preparation should increase both RT and force. A first goal of the current study was to examine whether sequential effects are indeed expressed in response force. A second goal was to examine whether sequential effects have a motoric locus, which seems plausible given that the main FP effect derives from sequential effects. An alternative possibility is that sequential effects are due to changes in response bias over time at a decisional level. Specifically, the threshold to select a response may be set lower from the moment at which the target occurred on a previous trial, without any effect on motor processing proper. Therefore, we focused on two measures that can be derived from the ongoing EEG (electroencephalogram) prior to the imperative stimulus. The first measure is a combination of slow waves that together have been denoted as the contingent negative variation (CNV; Walter, Cooper, Aldridge, McCallum, & Winter, 1964). The second measure is the lateralized readiness potential, which provides a highly specific index for (pre)motor activation. The early part of the CNV wave is assumed to reflect aspects of the warning stimulus (see Brunia, 1999), whereas the late part is thought to reflect anticipation of the target and also, but predominantly at anterior sites, motor preparation (the movement preceding negativity: MPN, see Jentzsch & Leuthold, 2002; Van Boxtel, 1994; Van Boxtel & Brunia, 1994; Verleger, Wauschkuhn, Van der Lubbe, Jaskowski, & Trillenberg, 2000). In spite of the successful application of the CNV in the study of nonspecific preparation, interpreting its amplitude remains somewhat ambiguous. It may indeed reflect motoric preparation, but also nonmotoric anticipation of the target. Therefore, using the LRP as an additional measure might enable more specific conclusions. The LRP is the difference of
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
249
activity between the sites contra- and ipsilateral to the response hand, for the sites overlying the primary motor areas. Because any activity unrelated to the response side is cancelled out by this subtraction, the LRP can be considered as a highly specific index for motor activation. Several studies observed that the LRP deviates from baseline before the response, and peaks at about the moment of executing the response (e.g. De Jong et al., 1988; Gratton et al., 1988), which is most clearly visible in response-locked averages (e.g. Van der Lubbe, Jaskowski, Wauschkuhn, & Verleger, 2001). Previous studies revealed that precues indicating the required response hand but not yet the precise response produce an LRP during the FP (Leuthold, Sommer, & Ulrich, 1996). As a consequence, when the required response hand is known but not yet the precise response, then the LRP can already provide information about variations in motor activation over time. Thus, finding an effect on the CNV and the LRP at a probable moment of target presentation would support the view that sequential effects have a motoric locus. 1 In the current study, three different equiprobable FPs (500, 1500, or 2500 ms) were employed that varied randomly from trial to trial, which should result in sequential effects of the FP on RT, being most pronounced for the shortest FP. Based on the extended overshoot model, responses were expected to be more forceful when the FP was longer on the previous trial than on the current trial, thereby also accounting for the overall decrease of response force when the FP on the current trial increases, as in that case the chance that the FP was longer on the previous trial decreases. Similarly, the amplitude of the anterior CNV was expected to be larger when the target on the previous trial occurred at that same critical moment than when it occurred at a later moment, reflecting increased preparation at that specific moment. A comparable effect on the LRP would indicate that at least part of this preparation effect has a motoric locus.
2. Methods 2.1. Participants Informed consent was obtained from 15 participants from the student population of L€ ubeck who were paid DM 42 (approximately A22) for their cooperation. All participants (three male, mean age: 23.2 years, one left handed) had normal or corrected to normal vision, and had no history of neurological disorders. Three participants were excluded from the EEG analyses because too many artifacts (>50%) were found in their recordings, which left 12 participants (three male, mean age: 23.0 years, one left handed).
1 We did not focus on the start of the stimulus- and the response-locked LRPs to determine whether the locus of sequential effects is in motor or premotor stages or earlier (see Osman & Moore, 1993), as this procedure seems less appropriate when the required response hand is already known, and the LRP starts deviating from zero preceding the onset of the target (see Leuthold et al., 1996).
250
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
2.2. Stimuli and procedure Participants were seated on a comfortable armchair in a separate chamber. Visual 00 stimuli were presented on a 14 Multisync monitor by means of a PC, with an observation distance of approximately 1.2 m. A central fixation cross (0.6 · 0.6) was continuously present. A tone (1000 Hz, about 70 dB) of 200 ms duration served as warning signal, presented by an external speaker placed behind the monitor. After an equiprobable FP of 500, 1500, or 2500 ms, the letter L or R (0.6 · 0.6), varying randomly from trial to trial, appeared at the center of the screen. Intertrial intervals were sampled from an exponential distribution with a mean of 700 ms. A constant period of 2800 ms was added to avoid any possible confounding of preparatory effects by activity needed to finish the preceding trial. 2.3. Task The task contained two hand conditions, each consisting of 450 trials, to enable the determination of an effect on the lateralized readiness potential (LRP). In the left hand condition, participants had to press a left or right button with their middle (in case of L) or index (in case of R) finger, and in the right hand condition they were instructed to press with their middle (in case of R) or index (in case of L) finger. The order of conditions was counterbalanced between participants. A short pause was given after every series of approximately 110 trials. Participants were instructed to respond as fast and as accurately as possible. 2.4. Recording and data processing EEG was recorded from 19 Ag/AgCl electrodes (FMS, Munich) located at standard electrode positions of the 10/20 system (Pivik et al., 1993). An electrode affixed at the nose was used as the reference. Using the nose has some advantages as compared to the employment of linked mastoids, as lateralized effects cannot be caused by the reference electrode, and artifacts arising from the neck muscles are avoided. To control for ocular artifacts, the electrooculogram (EOG) was recorded bipolarly both vertically from above and below the left eye (vEOG) and horizontally from the outer canthi of both eyes (hEOG). EEG and EOG were amplified and filtered by a Nihon-Kohden Neurotop amplifier (TC ¼ 5.0 s, lowpass 35 Hz). Electrode resistance was kept below 10 kX. Response force was recorded continuously from isometric weight elements that had to be pressed with the middle and index finger of one hand. The data (EEG, EOG, and response force) were recorded at a rate of 100 Hz after receiving triggers from a control computer from 95 ms before to 3200 ms after presenting the warning signal. Trials with artifacts were excluded from further analyses. EEG was corrected for EOG artifacts by using the procedure of Verleger, Gasser, and M€ ocks (1982). A low-pass filter with a cut-off at 14.7 Hz (Glaser & Ruchkin, 1976) was applied. Importantly, activity will decrease over time because of the high-pass filter of the amplifier (TC ¼ 5.0 s), leading to an underestimation of the actual amplitude, which implies a confounding for the main FP effect. Analytical cor-
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
251
rection (Elbert & Rockstroh, 1980) was used to remove this effect. This method iteratively calculates the unwanted decrease at a given time based on the already corrected output at the preceding time point, and then corrects this decrease (see Joyce, Gorodnitsky, Teder-S€ alej€ arvi, King, & Kutas, 2002, for a critical evaluation of this method). 2.5. Data analysis For all statistical analyses, correction for degrees of freedom was performed by using Huynh–Feldt epsilon, whenever necessary. 2.5.1. Response parameters RT was determined on each trial as the moment at which response force exceeded 2 N, and was averaged across middle and index finger responses and across left and right hand blocks. Trials with forces larger than 2 N on the incorrect finger, and responses faster than 150 ms were defined as errors and were excluded from further RT and response force analyses. Response force was defined as the maximum value of force output on a trial. The mean RTs, proportions of errors (PEs), and response forces were evaluated statistically by analysis of variance (ANOVA) with repeated measures with current FP (500, 1500, and 2500 ms) and previous FP (500, 1500, 2500 ms) as within subjects factors. 2.5.2. EEG parameters Event related potentials (ERPs) were computed for the Fz, Cz, and Pz electrodes by averaging EEGs (stimulus-locked) for all trials without artifacts. We focused on changes in CNV activity due to the previous FP. Averaged activity was determined in intervals from )50 to +50 ms around each possible moment of target occurrence to examine whether activity was affected by the FP in the previous trial when the FP in the current trial was the same as or longer than in the previous trial. This analysis was performed separately per possible moment of target occurrence (500, 1500, or 2500 ms), and included the factors electrode, previous FP and current FP where the number of levels for current FP decreased from 3 to 1 when the critical moment increased from 500 to 2500 ms. The LRP was determined by averaging response hand related lateralization for the left hand (C4–C3) and the right hand (C3–C4) conditions. The baseline was determined as the average activity in the )90 to 0 ms interval preceding the warning signal and we only determined activity until 1000 ms after the warning signal to include more trials. We restricted our statistical analyses to lateralized activity in a )20 to +20 ms window around the first possible moment of target occurrence (i.e. 500 ms), as the sequential effect on RT was predicted to be largest at this moment. 2
2
Note that a possible effect on the LRP does not reflect selection of the appropriate response: responses were given with two fingers of the same hand in a given block, therefore response selection cannot affect the asymmetry between the two hemispheres, measured by the LRP.
252
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
3. Results 3.1. Behavioral measures Very few premature responses (0.02%) and too slow responses or misses (0.27%) were observed. Mean response times, error percentages, and force amplitudes are displayed in Figs. 1 and 2 as a function of current and previous FP. The averaged force curves for the participants that were included in the EEG analyses are shown in the lower panels of Fig. 3a–c. Responses were slow when the current FP was short (a main effect of the current FP) F ð2; 28Þ ¼ 12:9, e ¼ 0:68, p ¼ 0:001, and when the previous FP was long (a main effect of the previous FP), F ð2; 28Þ ¼ 61:3, e ¼ 0:94, p < 0:001. Most important, these two factors interacted, F ð4; 56Þ ¼ 14:8, e ¼ 1:0, p < 0:001, reflecting that the influence of the previous FP decreases as the current FP increases (see Fig. 1). The analysis of PEs showed no significant effects, all F s < 1:0. Response force was
440 FP n - 1: 0.5 s FP n - 1: 1.5 s
Reaction time (ms)
430
FP n - 1: 2.5 s 420
410
400
390
Errors (%)
10 8 6 4
0.5
1.5
2.5
Foreperiod on trial n (s) Fig. 1. Reaction times (RT) and proportion of errors as a function of FP in the current trial n and FP in the previous trial n 1.
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
253
Response force (N)
9. 0 8. 8 8. 6 8. 4 8. 2 8. 0 7. 8
FP n - 1: 0.5 s FP n - 1: 1.5 s FP n - 1: 2.5 s
0.5
1.5
2.5
Foreperiod on trial n (s) Fig. 2. Response force as a function of FP in the current trial n and FP in the previous trial n 1.
the largest when the current FP was short, F ð2; 28Þ ¼ 5:1, e ¼ 0:76, p ¼ 0:022, but was neither affected by the previous FP, F ð2; 28Þ ¼ 2:1, e ¼ 1:0, p ¼ 0:15, nor was there an interaction between the previous and the current FP, F ð4; 56Þ ¼ 1:3, e ¼ 0:77, p ¼ 0:3. A separate analysis for the current FP of 500 ms, for which the effect of the previous FP should be the largest, in line with the RT data, revealed no significant effect, F ð2; 28Þ ¼ 0:3, e ¼ 0:8, p ¼ 0:5. 3.2. CNV CNVs are displayed in Fig. 3a–c (based on the data of 12 participants), for the frontal, central, and parietal electrodes as a function of previous and current FP. The analysis of averaged activity around 500 ms after the warning signal (the first moment at which the target could appear, see Fig. 3) with the factors electrode (Fz, Cz, Pz), current FP (500, 1500, or 2500 ms), and previous FP (500, 1500, or 2500 ms) revealed a main effect of previous FP, a main effect of electrode, and an interaction between previous FP and electrode, F s > 10:3, p < 0:002. At the Fz electrode, CNV amplitude was larger when the previous FP was 500 ms ()5.7 lV), than when it was 1500 ms ()3.8 lV) or 2500 ms ()3.6 lV), F ð1; 11Þ > 15:3, p < 0:002. At the Cz electrode, CNV amplitude was also larger when the previous FP was 500 ms ()4.8 lV) than when it was 1500 ms ()2.5 lV) or 2500 ms ()2.2 lV), F ð1; 11Þ > 32:3, p < 0:001. Analyses for the Pz electrode revealed no significant effect of previous FP, F ð2; 22Þ ¼ 2:1, p ¼ 0:16. Analysis of averaged activity around 1500 ms after the warning signal with the factors electrode (Fz, Cz, Pz), current FP (1500 and 2500) and previous FP (500, 1500, 2500) revealed a main effect of electrode, F ð2; 22Þ ¼ 7:9, e ¼ 0:82, p ¼ 0:005,
254
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
a) ERPs and averaged force curves with a current Foreperiod of 2500 ms
amplitude (µV)
-10 Fz
-5 0 5 10 -10
amplitude (µV)
Cz
-5 0 5 10
amplitude (µV)
-10 Pz
-5 0 5
response force (N)
10 10 8
Fpn-1(500)
6
Fpn-1(1500)
4
Fpn-1(2500)
2 0
WS
500
1000
1500 t(ms)
2000
2500
3000
Fig. 3. (a–c) ERPs on the frontal (Fz), central (Cz), and the parietal (Pz) electrodes, and averaged force curves after the acoustic warning stimulus (WS) presented at t ¼ 0. After an FP of 2500 (a), 1500 (b), or 500 ms (c) the target was presented. The ERPs and force curves are indicated as a function of the FP in the previous trial (FPn1 ). The moments at which targets could occur are indicated by dashed lines. The peak amplitudes of the force curves are lower than the determined peak amplitudes on a single trial basis due to trial to trial jittering, individual differences, and the exclusion of three participants from the EEG analyses.
no effect of previous FP, and an interaction between electrode and previous FP, F ð4; 44Þ ¼ 3:9, e ¼ 0:57, p ¼ 0:019. Separate analyses per electrode revealed a trend toward an effect of previous FP for the Fz electrode, F ð2; 22Þ ¼ 3:1, e ¼ 0:7, p ¼ 0:087, with a small positivity (0.2 lV) when the previous FP was 2500 ms and negativity when it was 1500 ()1.7 lV) or 500 ms ()1.6 lV), and also a trend toward
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
255
b) ERPs and averaged force curves with a current Foreperiod of 1500 ms -10 amplitude (µV)
Fz -5 0 5 10 -10 amplitude (µV)
Cz -5 0 5 10 -10 amplitude (µV)
Pz -5 0 5 10
response force (N)
10 8
Fpn-1(500)
6
Fpn-1(1500)
4
Fpn-1(2500)
2 0 WS
200
400
600
800
1000 1200 1400 1600 1800 2000 2200 t(ms)
Fig. 3 (continued)
an effect for the Cz electrode, F ð2; 22Þ ¼ 3:2, e ¼ 0:9, p ¼ 0:069, with smallest negativity when the previous FP was 2500 ms ()2.6 lV), and larger negativity when it was 1500 ()4.1 lV) or 500 ms ()4.2 lV).
256
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
c) ERPs and averaged force curves with a current Foreperiod of 500 ms
-10 amplitude (µV)
Fz
-5 0 5 10 -10
amplitude (µV)
Cz
-5 0 5 10 -10
amplitude (µV)
Pz
-5 0 5 10
response force (N)
10 8
Fpn-1(500)
6
Fpn-1(1500)
4
Fpn-1(2500)
2 0 WS
200
400
600 t(ms)
800
1000
1200
Fig. 3 (continued)
Analysis on activity when the current FP was 2500 ms (see Fig. 3) revealed a main effect of electrode, F ð2; 22Þ ¼ 3:5, e ¼ 0:99, p ¼ 0:048, and a significant interaction between electrode and previous FP, F ð4; 44Þ ¼ 9:2, e ¼ 0:77, p < 0:001. Separate analyses per electrode revealed no effect of previous FP, F ð2; 22Þ < 1:9.
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
257
3.3. LRP The LRPs and hEOGs (determined relative to response side), as a function of the previous FP averaged across current FP, from )90 to 700 ms after onset of the warning signal, are displayed in Fig. 4. The data turned out to be rather noisy, possibly because the LRP was determined in a procedure where the responding hand was varied between blocks of trials rather than on a trial-by-trial basis. For this reason, we restricted our analysis of this data to testing a specific hypothesis concerning the LRP amplitude in a 40 ms time window surrounding the earliest possible moment of target presentation (i.e. from 480 to
The LRP and hEOG as a function of previous FP -20 hEOG
amplitude (µV)
-10
0
10
20 -2 LRP Fpn-1(500) Fpn-1(1500)
amplitude (µV)
-1
Fpn-1(2500)
0
1 WS
250
500
750
t(ms)
Fig. 4. The LRPs as a function of the previous FP, averaged across current FP, from 90 ms before the acoustic warning stimulus (WS) until 700 ms after the warning stimulus. At t ¼ 500 ms, the first possible moment of target presentation is indicated.
258
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
520 ms after the onset of the warning stimulus), as a function of the FP that occurred on the preceding trial. This analysis revealed a significant effect of previous FP, F ð2; 22Þ ¼ 4:5, e ¼ 0:86, p ¼ 0:03. Larger contralateral negativity was observed when the previous FP was 500 ms ()0.26 lV) than when it was 1500 or 2500 ms (0.11 lV). The hEOG waveforms show that the effects on the LRP are not due to horizontal eye movements. Finally, we explored whether there is a relation between the LRP amplitudes in our 480–520 ms window and effects found on RT when the current FP was 500 ms. Per participant, we computed the difference in LRP amplitude between conditions in which the previous FP was 500 ms and when it was 2500 ms, and we determined the difference in RT between these conditions. A one-tailed test showed that there was no significant relation between both measures ðp ¼ 0:32Þ.
4. Discussion Since the seminal work of Woodrow (1914) it has been known that the speed of responding is dependent on the time currently available for preparation (the FP), as well as on FPs occurring on immediately preceding trials, as reflected by sequential effects. So, it has been recognized for a long time that the capacity for response preparation is not equally spread over a wider interval. However, only recently it was pointed out that the temporal relations of preceding events may fully account for effects on current preparation (Los et al., 2001). The presented data replicated the classical FP effect and the sequential effects of the preceding FP when FP duration varies from trial to trial (see also Los & Van den Heuvel, 2001; Niemi & N€a€at€anen, 1981). First, responses were faster when the FP on the current trial became longer. Secondly, this effect was only observed when the FP on the current trial was shorter than on the previous trial: Thus, there was no effect of current FP on RT when the previous FP was 500 ms, and there was no difference between current FPs of 1500 and 2500 ms when the previous FP was 1500 ms, because these were the cases when the previous FP was shorter than or equal to the current FP. In terms of the extended overshoot model, these data indicate that at the moment of target occurrence the distance between motor activation and the motor threshold is larger when the target on the previous trial occurred later than on the current trial as compared to when it occurred earlier or at the same moment. Regarding the exerted response force, our data replicate the finding that overall force decreases when the FP on the current trial increases. However, opposed to the prediction of the extended overshoot model, response force showed, if anything, a sequential effect that was opposite to the pattern observed on RT. In particular, Fig. 2 shows no visible effect of the previous FP on response force when the current FP was 500 ms, whereas according to the model the effect of previous FP should be most evident for the shortest current FP. Clearly, the observed main effect of the current FP on response force cannot be explained from sequential effects. This implies a failure of the extended overshoot model. The sequential effects on RT (see also the CNV and the LRP data) suggest that motor activation at target onset is lower when the previous FP was longer than the current FP, and this decreased motor activation should
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
259
therefore lead to more forceful responses, which is not what we found. Another explanation, in line with alternative models on the regulation of response force (see Jaskowski et al., 2000) is that the larger force for short FPs is determined by immediate arousal arising from the acoustic warning stimulus. However, the study by Jaskowski and Verleger (1993) led to the conclusion that preparatory state rather than arousing effects of the warning stimulus are responsible for the FP effect on force. Nevertheless, for the present study, it cannot be excluded that immediate arousal was related to the observed effect on force. Interestingly, there were clear changes of fronto-central CNV amplitude during the FP around the first moment at which the target could be presented. CNV was more negative around 500 ms when the FP in the previous trial had been 500 ms than when it had been either 1500 or 2500 ms. CNV amplitude around the second moment of 1500 ms tended to be more negative when the previous FP had been 500 or 1500 ms than when it had been 2500 ms. Possibly the latter effect could not so clearly be demonstrated because the number of trials with the current FP larger than 1500 ms was smaller than the number of trials that could be used for the analysis with the current FP larger than 500 ms. Importantly, the effects of prior FP around the first and second moment of possible target occurrence are unlikely to be caused by the warning stimulus itself, as there is no reason to expect a change in the processing of this stimulus because of an earlier event. The effect had a fronto-central focus. According to Verleger et al. (2000), who compared CNV before hand movements and before eye movements, this topography accords with activation of hand-motor areas. Furthermore, according to our previous results, this specific topography does not reflect a higher degree of general readiness because, when CNVs were compared between trials with and without speed instruction the differences between these states had a more parietal focus (Verleger et al., 2000). Thus, in any case parietal activation should be involved when general response readiness is increased, and therefore the present topography suggests a more direct associative link from the previous trial to present hand-motor activation, without much energizing side-effects. This might fit the lack of effects on response force, which is presumably sensitive to this energizing component (Wascher et al., 1997) and would also fit the notion that the previousforeperiod effect reflects an automatic, conditioned link to the effector system (Los & Van den Heuvel, 2001). Further support for this interpretation comes from our LRP data. The effect of the previous FP on the LRP around 500 ms after the warning signal indicates that selective activation of the responding hand was indeed higher at that moment when the previous FP was 500 ms than when the previous FP was longer. That this effect was less marked with the LRP than with the CNV might be due to the use of the same hand throughout a block of trials, with the tonic preference of the active hand to the passive hand possibly insufficiently reflected in the phasic LRP, which among other reasons (e.g. individual differences), may also explain the absence of a direct relation between the sequential effect on LRP amplitude and the sequential effect on RT. These data favor the view that the sequential effect of FP on RT, and therefore also the ensuing main effect of FP on RT, has a motoric locus. In conclusion, our RT data support the view that sequential effects are essential for understanding the main FP effect, which is underlined by the sequential effects
260
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
on the CNV and the LRP. The latter effects support the view that the sequential effect on RT has, at least partially, a motoric locus. Furthermore, our finding of a dissociation between effects of the previous FP on RT and response force disfavors the extended overshoot model, at least in case of auditory warning stimuli. Future studies should reveal whether this conclusion also applies to settings with visual warning stimuli, which do not have the arousing properties that are attributed to acoustic stimuli.
Acknowledgements Rob van der Lubbe and Piotr Jaskowski were supported by a grant from the Deutsche Forschungs-gemeinschaft (Ve 110/7-1 and 110/7-2), and Rob van der Lubbe was also supported by a grant from The Netherlands Organization for Fundamental Research (NWO: 440-20-000). We thank Ellen Gardziella and Angela Weber for their help in acquiring the data. In addition, we want to thank Werner Sommer and an anonymous reviewer for comments on an earlier draft of this article. An earlier analysis of the current experiment was presented at the eighth meeting of the Cognitive Neuroscience Society (CNS) in New York (see Van der Lubbe, Los, Jaskowski, & Verleger, 2001).
References Bertelson, P., & Tisseyre, F. (1968). The time-course of preparation with regular and irregular foreperiods. Quarterly Journal of Experimental Psychology, 20, 297–300. Brunia, C. H. M. (1983). Motor preparation: Changes in amplitude of achilles tendon reflexes during a fixed foreperiod of one second. Psychophysiology, 20, 658–664. Brunia, C. H. M. (1999). Neural aspects of anticipatory behavior. Acta Psychologica, 101, 213–242. Coull, J. T., Frith, C. D., B€ uchel, C., & Nobre, A. C. (2000). Orienting attention in time: Behavioural and neuroanatomical distinction between exogenous and endogenous shifts. Neuropsychologia, 38, 808– 819. Coull, J. T., & Nobre, A. C. (1998). Where and when to pay attention: The neural systems for directing attention to spatial locations and to time intervals as revealed by both PET and fMRI. The Journal of Neuroscience, 18, 7426–7435. De Jong, R., Wierda, M., Mulder, G., & Mulder, L. J. M. (1988). Use of partial information in responding. Journal of Experimental Psychology: Human Perception and Performance, 14, 682–692. Drazin, D. (1961). Effects of foreperiod, foreperiod variability and probability of stimulus occurrence on simple reaction time. Journal of Experimental Psychology, 62, 43–50. Eimer, M. (1998). The lateralized readiness potential as an on-line measure of central response activation process. Behavior Research Methods, Instruments, and Computers, 30, 146–156. Elbert, T., & Rockstroh, B. (1980). Some remarks on the development of a standardized time constant. Psychophysiology, 17, 504–505. Glaser, E. M., & Ruchkin, D. S. (1976). Principles of neurobiological signal analysis. New York: Academic Press. Gratton, G., Coles, M. G. H., Sirevaag, E. J., Eriksen, C. W., & Donchin, E. (1988). Pre- and poststimulus activation of response channels: A psychophysiological analysis. Journal of Experimental Psychology: Human Perception and Performance, 14, 331–344.
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
261
Hasbroucq, T., Osman, A., Possama€ı, C.-A., Burle, B., Carron, S., Depy, D., Latour, S., & Mouret, I. (1999). Cortico-spinal inhibition reflects time but not event preparation: Neural mechanisms of preparation dissociated by transcranial magnetic stimulation. Acta Psychologica, 101, 243–266. Jaskowski, P., Van der Lubbe, R. H. J., Wauschkuhn, B., Wascher, E., & Verleger, R. (2000). The influence of time pressure and cue validity on response force in an S1–S2 paradigm. Acta Psychologica, 105, 89–105. Jaskowski, P., & Verleger, R. (1993). A clock paradigm to study the relationship between expectancy and response force. Perceptual and Motor Skills, 77, 163–174. Jentzsch, I., & Leuthold, H. (2002). Advance movement preparation of eye, foot, and hand: A comparative study using movement-related potentials. Cognitive Brain Research, 14, 201–217. Jentzsch, I., & Sommer, W. (2002a). The effect of intentional expectancy on mental processing: A chronopsychophysiological investigation. Acta Psychologica, 111, 265–282. Jentzsch, I., & Sommer, W. (2002b). Functional localization and mechanisms of sequential effects in serial reaction time tasks. Perception & Psychophysics, 64, 1169–1188. Joyce, C. A., Gorodnitsky, I. F., Teder-S€alej€arvi, W. A., King, J. W., & Kutas, M. (2002). Variability in AC amplifier distortions: Estimation and correction. Psychophysiology, 39, 633–640. Leuthold, H., Sommer, W., & Ulrich, R. (1996). Partial advance information and response preparation: Inferences from the lateralized readiness potential. Journal of Experimental Psychology: General, 125(3), 307–323. Los, S. A. (1996). On the origin of mixing costs: Exploring information processing in pure and mixed blocks of trials. Acta Psychologica, 94, 145–188. Los, S. A., Knol, D. L., & Boers, R. M. (2001). Looking inside the motor adjustment stage. Conditioning as a basis for nonspecific preparation. Acta Psychologica, 106, 121–145. Los, S. A., & Van den Heuvel, C. E. (2001). Intentional and unintentional contributions to nonspecific preparation during the foreperiod. Journal of Experimental Psychology: Human Perception and Performance, 27, 370–386. Mattes, S., & Ulrich, R. (1997). Response force is sensitive to the temporal uncertainty of response stimuli. Perception & Psychophysics, 59, 1089–1097. Mattes, S., Ulrich, R., & Miller, J. (1997). Effects of response probability on response force in simple RT. The Quarterly Journal of Experimental Psychology, 50A, 405–420. Miniussi, C., Wilding, E. L., Coull, J. T., & Nobre, A. C. (1999). Orienting attention in time. Modulation of brain potentials. Brain, 122, 1507–1518. M€ uller-Gethmann, H., Ulrich, R., & Rinkenauer, G. (2003). Locus of the effect of temporal preparation: Evidence from the lateralized readiness potential. Psychophysiology, 40, 597–611. N€ a€ at€ anen, R. (1971). Non-aging fore-periods and simple reaction time. Acta Psychologica, 35, 316–327. Niemi, P., & N€ a€ at€ anen, R. (1981). Foreperiod and simple reaction time. Psychological Bulletin, 89, 133– 162. Nobre, A. C. (2001). Orienting to instants in time. Neuropsychologia, 39, 1317–1328. Osman, A., & Moore, C. M. (1993). The locus of dual-task interference: Psychological refractory effects on movement-related brain potentials. Journal of Experimental Psychology: Human Perception and Performance, 19, 1292–1312. Pivik, R. T., Broughton, R. J., Coppola, R., Davidson, R. J., Fox, N., & Nuwer, M. R. (1993). Guidelines for the recording and quantitative analysis of electroencephalographic activity in research contexts. Psychophysiology, 30, 547–558. Requin, J., Brener, J., & Ring, C. (1991). Preparation for action. In J. R. Jennings & M. G. H. Coles (Eds.), Handbook of cognitive neurophysiology: Central and autonomic nervous system approaches (pp. 357–448). New York: Wiley. Sanders, A. F. (1980). Some effects of instructed muscle tension on choice reaction time and movement time. In R. S. Nickerson (Ed.), Attention and performance VII (pp. 59–74). Hillsdale, NJ: Erlbaum. Tandonnet, C., Burle, B., Vidal, F., & Hasbroucq, T. (2003). The influence of time preparation on motor processes assessed by surface Laplacian estimation. Clinical Neurophysiology, 114, 2376–2384. Van Boxtel, G. J. M. (1994). Non-motor components of slow brain potentials. Dissertation, Katholieke Universiteit Brabant, Tilburg, The Netherlands.
262
R.H.J. Van der Lubbe et al. / Acta Psychologica 116 (2004) 245–262
Van Boxtel, G. J. M., & Brunia, C. H. M. (1994). Motor and non-motor components of the Contingent Negative Variation. International Journal of Psychophysiology, 17, 269–279. Van der Lubbe, R. H. J., Jaskowski, P., Wauschkuhn, B., & Verleger, R. (2001). Influence of time pressure in a simple response task, a choice-by-location task, and the Simon task. Journal of Psychophysiology, 15, 241–255. Van der Lubbe, R. H. J., Los, S. A., Jaskowski, P., & Verleger, R. (2001). Sequential effects of the foreperiod examined by using psychophysiological indices for motor activation. Journal of Cognitive Neuroscience (Supplement Eight Annual Meeting, 2001, New York), 6B, 46. Verleger, R., Gasser, T., & M€ ocks, J. (1982). Correction for EOG artifacts in event related potentials of the EEG: Aspects of reliability and validity. Psychophysiology, 19, 472–480. Verleger, R., Wauschkuhn, B., Van der Lubbe, R. H. J., Jaskowski, P., & Trillenberg, P. (2000). Posterior and anterior contribution of hand-movement preparation to late CNV. Journal of Psychophysiology, 14, 69–86. Walter, W. G., Cooper, R., Aldridge, V. J., McCallum, W. C., & Winter, A. L. (1964). Contingent negative variation: An electric sign of sensorimotor association and expectancy in the human brain. Nature, 230, 380–384. Wascher, E., Verleger, R., Vieregge, P., Jaskowski, P., Koch, S., & K€ ompf, D. (1997). Responses to cued signals in Parkinson’s disease: Distinguishing between disorders of cognition and of activation. Brain, 120, 1355–1375. Woodrow, H. (1914). The measurement of attention. Psychological Monographs, 17, 1–158. Wundt, W. (1887). Grundz€uge der physiologischen Psychologie [Foundations of physiological psychology]. Leipzig, German: Engelmann.
