The norepinephrine system and its relevance for multi-component behavior

Author’s Accepted Manuscript The norepinephrine system and its relevance for multi-component behavior Moritz Mückschel, Krutika Gohil, Tjalf Ziemssen, Christian Beste www.elsevier.com

PII: DOI: Reference:

S1053-8119(16)30553-5 http://dx.doi.org/10.1016/j.neuroimage.2016.10.007 YNIMG13504

To appear in: NeuroImage Received date: 13 July 2016 Revised date: 8 September 2016 Accepted date: 3 October 2016 Cite this article as: Moritz Mückschel, Krutika Gohil, Tjalf Ziemssen and Christian Beste, The norepinephrine system and its relevance for multicomponent behavior, NeuroImage, http://dx.doi.org/10.1016/j.neuroimage.2016.10.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

2

The norepinephrine system and its relevance for multi-component behavior

3

Moritz Mückschel1,2,*, Krutika Gohil1, Tjalf Ziemssen2, Christian Beste1,3

1

1

4 5 6 7 8 9 10 11 12

2

3

Cognitive Neurophysiology, Department of Child and Adolescent Psychiatry, Faculty of Medicine, TU Dresden, Germany

MS Centre Dresden, Centre of Clinical Neuroscience, Department of Neurology, Faculty of Medicine, TU Dresden, Germany

Experimental Neurobiology, National Institute of Mental Health, Klecany, Czech Republic

*Corresponding author at: Cognitive Neurophysiology, Department of Child and Adolescent Psychiatry, Faculty of Medicine of the TU Dresden, Germany, Schubertstrasse 42, D-01309 Dresden, Germany. Tel.: +49 351 458 7072; fax: +49 351 458 7318. [email protected]

13 14 15 16

1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Abstract The ability to execute several actions in a specific temporal order to achieve an overarching goal, a process often termed action cascading or multi-component behavior, is essential for everyday life requirements. We are only at the beginning to understand the neurobiological mechanisms important for these cognitive processes. However, it is likely that the locus coeruleus-norepinephrine (LC-NE) system may be of importance. In the current study we examine the relevance of the LC-NE system for action cascading processes using a system neurophysiological approach combining high-density EEG recordings and source localization to analyze event-related potentials (ERPs) with recordings of pupil diameter as a proximate of LC-NE system activity. N= 25 healthy participants performed an action cascading stopchange paradigm. Integrating ERPs and pupil diameter using Pearson correlations, the results show that the LC-NE system is important for processes related to multi-component behavior. However, the LC-NE system does not seem to be important during the time period of response selection processes during multi-component behavior (reflected in the P3) as well as during perceptual and attentional selection (P1 and N1 ERPs). Rather, it seems that the neurophysiological processes in the fore period of a possibly upcoming imperative stimulus to initiate multi-component behavior are correlated with the LC-NE system. It seems that the LC-NE system facilitates responses to task-relevant processes and supports task-related decision and response selection processes by preparing cognitive control processes in case these are required during multi-component behavior rather than modulating these processes once they are operating.

22 23 24 25

Keywords: cognitive control, sensorimotor integration, executive function, EEG, vegetative factors, pupil, source localization

2

1

1. Introduction

2 3 4 5 6 7 8 9 10 11 12

To cope with most everyday tasks, one has to execute several actions in a specific temporal order to achieve an overarching multi-component goal. Doing so, we are quite often required to interrupt an ongoing action and turn to an alternative action. During such multi-component behavior we heavily depend on action cascading processes, which are defined as the ability to generate, process, and execute separate task goals and responses in an expedient temporal order to produce an efficient goal-directed multi-component behavior (Dippel and Beste, 2015; Duncan, 2010; Mückschel et al., 2015, 2014, Stock et al., 2015, 2014). These processes have been shown to differ in their demands on cognitive control and response selection processes, depending on whether stimuli signaling to execute one specific behavior are presented simultaneously, or with a temporal gap in between. This makes it possible to finish one behavioral subprocess in the chain of actions before executing another one.

13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

A number of neurobiological factors have already been shown to modulate these processes. For example, using transcranial vagus nerve stimulation (tVNS), known to modulate the GABAergic and the norepinephrine (NE) system (Raedt et al., 2011; Van Leusden et al., 2015), it has been shown that action cascading processes become faster (Steenbergen et al., 2015). Both, the GABAergic and the NE system are therefore likely to play a role, but from the data by Steenbergen et al. (2015) it cannot be concluded which of the two modulated systems (i.e. GABA or NE) are central to modulations in action cascading. While the relevance of the GABAergic system for these processes has been underlined using magnetic resonance spectroscopy (Yildiz et al., 2014) there is no knowledge about the role of the NE system. The NE system may be of particular relevance since it has been suggested that one property of the locus coeruleus-NE function (LC-NE) is to modulate task-related decision or selection processes (for review: Aston-Jones and Cohen, 2005) with phasic LC-NE responses probably facilitating responses to task-relevant processes (Aston-Jones and Cohen, 2005; Nieuwenhuis et al., 2005). Exactly such processes are important for action cascading and multi-component behavior (Mückschel et al., 2014), which makes it likely that phasic LC-NE responses are related to this important executive control function.

29 30 31 32 33 34 35 36 37 38 39 40 41 42

The goal of the current study is to investigate the relevance of the LC-NE system for multicomponent behavior using a Stop-Change task. This goal is pursued in a system neurophysiological approach using high-density EEG recordings to analyze event-related potentials (ERPs) and their related neuronal sources combined with recordings of pupil diameter. The pupil diameter has been suggested not only to be modulated by tonic, but also by phasic LC-NE activity (Hou et al., 2005; Murphy et al., 2011). Therefore, recordings of the pupil diameter are frequently used when interested in LC-NE functions in relation to cognitive functions in humans (Gilzenrat et al., 2010; Hong et al., 2014; Jepma and Nieuwenhuis, 2010; Murphy et al., 2011). Opposed to a pharmacological manipulation of the NE system, the continuous recording of the pupil diameter in combination with ERPs offers the advantage to examine which cognitive subprocess and at which precise time point during information processing in a given task the NE system is of importance. The latter aspect is of particular relevance to examine a possible role of phasic LC-NE responses for dissociable subprocesses during multi-component behavior.

43 44

On a neurophysiological (EEG) level, the P3 ERP has frequently been suggested to reflect decision processes between stimulus evaluation and motor responding, i.e. action selection 3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

processes (e.g. Twomey et al., 2015; Verleger et al., 2015). In line with this, response selection mechanisms during action cascading have been shown to be reflected by P3 ERP amplitude modulations having their source in anterior cingulate (ACC), inferior parietal and inferior frontal cortices (Beste et al., 2014a; Dippel and Beste, 2015; Stock et al., 2015). Especially ACC regions show stronger activation in situations when response selection mechanisms can operate in a suboptimal (parallel) fashion, i.e. when multiple response options are handled at the same time (e.g. Mückschel et al., 2014). These situations are linked to higher P3 amplitudes during multi-component behavior. The reason is that the P3 likely reflects inhibition processes and changing processes. This is especially the case when people handle response options at the same time (Dippel and Beste, 2015). As has been shown before (Beste et al., 2014b), the more participants attempt to simultaneously process the “stop-goal” and the “change-goal” in a Stop-Change task used to examined multi-component behavior, the stronger is the interference between these goals at a strategic response selection bottleneck (Verbruggen et al., 2008). Inhibitory control processes to manage the stopping of a response are likely to be intensified due to this strong interference. Such an intensification of response inhibition efforts has frequently been shown to be related to higher P3 amplitudes (Huster et al., 2013a).

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

Importantly, it has been suggested that the P3 is an electrophysiological correlate of the LC phasic NE response (Murphy et al., 2011; Nieuwenhuis et al., 2005). We therefore hypothesize that especially these response selection processes (reflected by the P3) related to medial frontal cortices are correlated with LC-NE responses. Higher NE concentrations, as reflected by a larger pupil diameter (Hou et al., 2005; Phillips et al., 2000), are supposed to be related to more efficient response selection processes during multi-component behavior. However, the "LC-P3" theory has more recently been modified (Warren et al., 2011; Warren and Holroyd, 2012). According to these modifications LC bursts impact cortical activity earlier than during the time frame of the P3, i.e. ~250 ms post stimulus, and it is the LC refractory period that coincides with the P3 generation. An alternative hypothesis may therefore be that the LC-NE system is specifically correlated with neurophysiological processes prior the P3. Previous studies suggest that aside from response selection processes, also perceptual gating and attentional selection mechanisms, reflected by P1 and N1 ERP (e.g. Herrmann and Knight, 2001) play a role in action cascading in that they contribute to an efficient unfolding of response selection processes (Gohil et al., 2015; Yildiz et al., 2014). Also for these processes the LC-NE system activity has been described to be important (Corbetta et al., 2008; Lisi et al., 2015; Petersen and Posner, 2012; Posner and Petersen, 1990). Therefore, we hypothesize that processes from perceptual gating and attentional selection level to processes at the response selection level also correlate with the pupil diameter as a proximate of the LC-NE system activity.

38 39

2. Materials and Methods

40

2.1 Participants

41 42 43 44

A sample of 25 healthy subjects, 9 male and 16 female, aged 20 to 32 years (22.78 ± 2.76) took part in the experiment. All participants were students. All of the participants were righthanded and had no history of psychiatric or neurologic diseases. Each participant gave written informed consent before beginning the experiment. After the experiment, each of them was 4

1 2

reimbursed with 10 €. The study was approved by the ethics committee of the Faculty of Medicine of the TU Dresden and accords with the Declaration of Helsinki.

3

2.2 Task

4 5 6 7 8 9

A modified version of the Stop-Change paradigm introduced by Verbruggen et al. (2008) was used for this study (see Figure 1) (Gohil et al., 2016a; Mückschel et al., 2015). All participants were seated at a distance of 60 cm from a 24 inch computer monitor in a dimly lit room (illuminance 45.2 lux). The participant's head was fixated using a chin rest. The participants responded using one out of four keys on a response pad in front of them. The stimuli were presented using the software “Presentation” (Neurobehavioral Systems Inc.).

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

The task lasted about 25 minutes and included 864 trials. Two third of the trials were “GO” trials, one third “STOP-CHANGE” (SC) trials. All trials were presented in a pseudorandomized order. The stimuli were presented on a black background (0.4 cd/m²). Four vertically arranged white border circles with a diameter of 7 mm were presented inside a white rectangular frame (55x16 mm). Each circle was separated by a white horizontal line (width 13 mm, line thickness 1 mm). This task array was shown at the beginning of each trial. After 250 ms, one of the four circles was filled with white color indicating the target (120.1 cd/m²). In the GO condition, the middle horizontal line was the “reference line”. Participants were asked to indicate with their right hand whether the filled white circle was located below or above the reference line. If the white filled circle was located above the reference line, the correct response was to press the top right key with the right middle finger. If the circle was below the reference line, the correct response was to press the lower right button using the right index finger. If participants did not respond within 1000 ms after the onset of the target, a speed up sign (German word “Schneller!” which translates to “Faster!”) was presented above the stimulus array and displayed until the trial ended with a button press. In SC trials the GO stimulus was followed by a STOP stimulus, indicated by a red frame replacing the white bordered frame surrounding the stimulus array (118.8 cd/m²). Participants were instructed to refrain from responding if they see the STOP stimulus. The variable Stop-signal delay (SSD) was initially set to 250 ms but adapted to the individual task performance by means of a staircase algorithm (Logan and Cowan, 1984; Verbruggen et al., 2008). For every correct inhibition the SSD decreased by 50 ms, for every false alarm the SSD was increased by 50 ms. The lower limit of the SSD was 50 ms and the upper limit was 1000 ms. The staircase procedure was used to yield a 50% probability of successful inhibition for SC trials. The stop signal reaction time (SSRT) consequently can be estimated by subtracting the mean SSD from the mean go RT (Verbruggen and Logan, 2009). The STOP signal was always followed by a tactile CHANGE signal. In half of the SC trials, the CHANGE signal was presented with a stop-change delay (SCD) of 300 ms (SCD300 trials). In the other half of the SC trials, the SCD was set to 0 ms (SCD0 trials). The change signal was presented using an air puff device. The air puff was delivered on two different positions: On the chest of the participant or below the navel. The puff on the chest indicated to use the top horizontal line as reference line for this trial. The lower puff represented the bottom horizontal line and the central reference line was indicated by a simultaneous air puff at both locations. Participants were instructed to respond to this CHANGE signal with their left hand. If the white circle was below the reference line indicated by the CHANGE signal, the correct response was to press the lower left key with the left middle finger. If the white circle was above the reference line, the correct response was pressing the upper left key with the left index finger. The speed up 5

1 2

sign was displayed if participants did not respond within 2000 ms after the onset of the CHANGE stimulus until the trial ended by pressing a response button.

3

2.3 EEG and pupil diameter recording and analysis

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

High-density EEG recording was acquired using a QuickAmp amplifier (Brain Products, Inc.) with 60 Ag–AgCl electrodes at standard scalp positions in an equidistant electrode setup. The reference electrode was located at electrode Fpz. All electrode impedances were kept below 5 kΩ. The data were recorded with 1 kHz and then down-sampled offline to 256 Hz. Afterwards the EEG was filtered using filter boundaries between 0.5 and 20 Hz (and 50 Hz notch filter; each 48db/oct; IIR-filter). In a subsequent manual inspection of the data gross technical artifacts were removed. Recurrent artifacts like eye blinks, horizontal and vertical movements as well as pulse artefacts were removed by means of independent component analysis (ICA) (infomax algorithm). The number of computed independent components for the ICA was determined by the number of valid EEG channels minus 1. After the removal of artifacts, the EEG data was segmented according to the two different SCD conditions (i.e. SCD0 and SCD300). The segmentation was performed in relation to the occurrence of the stop signal. Only correct trials were included in the data analysis. After the data was epoched, an automated artifact rejection was applied. The rejection criteria included a voltage of more than 150 μV/ms, a value difference of more than 150 μV in a 250 ms interval, or activity below 0.1 μV in a 100 ms interval. In order to eliminate the reference potential from the data, a current source density (CSD) transformation was run (Perrin et al., 1989). In addition to removing the reference potential, the CSD also serves as a spatial filter (Nunez and Pilgreen, 1991) which helps to identify the electrodes that best reflect activity related to cognitive processes. Finally, a baseline correction was made for the time window from −800 till −600 ms. Similar to previous studies on this paradigm (e.g. Mückschel et al., 2014), this baseline was chosen to have a pre-stimulus baseline before the presentation of the Go stimulus. We quantified the P1, N1, and P3 event-related potentials (ERPs). Electrodes were chosen by visual inspection on the basis of the scalp topographies. The visual P1 and N1 were quantified at electrodes P7 and P8 (P1: 80– 160 ms and N1: 150-300 ms post-stimulus, respectively) and the P3 was quantified at electrode Cz (SCD0: 150-380 ms; SCD300: P3 for STOP process 180-400 ms and P3 for the CHANGE process 470-670 ms). We were not able to determine tactile P1 and N1 ERPs. All ERP components were quantified relative to the pre-stimulus baseline. For all components, we quantified peak amplitude and latency on the single-subject level. The choice of these electrodes was statistically validated using the procedure described in Mückschel et al. (2014). This procedure confirmed the choice of the above-mentioned electrodes used to quantify the ERPs.

36 37 38 39 40 41 42 43 44 45 46

Pupil data was recorded using a RED 500 eye tracking device and the software iView X (SensoMotoric Instruments GmbH) using a sampling rate of 500 Hz. During the experiment, the participants head was fixed in a chin rest. The eye tracking device was calibrated prior to starting the task using a 9 point calibration. Eye blinks were automatically interpolated by the eye tracking recording software. For all subjects, the pupil diameter data in millimeter (mm) of both eyes was analyzed; i.e. the mean pupil diameter of both eyes was used for data analysis. Pupil reactions from the left and right eye did not differ (p > .95) and were strongly correlated (r > .98). Pupil diameter data and EEG data were synchronized using the EYE-EEG extension (Dimigen et al., 2011) for EEGLab (Delorme and Makeig, 2004) (http://www2.hu-berlin.de/eyetracking-eeg). For analysis, the sampling rate was downsampled to 256 Hz and a low pass filter was applied (IIR filter with 20 Hz at a slope of 48 6

1 2 3 4 5 6

dB/oct each). Remaining artifacts, for example artifacts related to eye blinks, were removed manually by means of a raw data inspection. Segmentation and baseline correction were conducted analog to the EEG data (i.e. the same baseline from -800 to -600 ms was used). For pupil diameter data, we quantified the minimum peak (Lx min) of the initial decrease 0-600 ms after the locking point (i.e. STOP stimulus presentation) and the maximum peak (Lx max) of the diameter increase 1000-2000 after the locking point.

7

2.4 Source localization analysis

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

To examine sources relating to amplitude modulations in ERPs in the SCD0 and SCD300 condition, source localization was conducted using sLORETA (standardized low resolution brain electromagnetic tomography; Pascual-Marqui, 2002). sLORETA provides a single linear solution to the inverse problem without a localization bias (Marco-Pallarés et al., 2005; Pascual-Marqui, 2002; Sekihara et al., 2005). It has been mathematically proven that sLORETA provides reliable results without localization bias (Sekihara et al., 2005). There is also evidence of EEG/fMRI and EEG/TMS studies underlining the validity of the sources estimated using sLORETA (e.g. (Dippel and Beste, 2015; Sekihara et al., 2005). For sLORETA, the intracerebral volume is partitioned into 6239 voxels at 5 mm spatial resolution. The standardized current density at each voxel is calculated in a realistic head model (Fuchs et al., 2002) using the MNI152 template. For the statistics the sLORETA-builtin voxel-wise randomization tests with 2000 permutations, based on statistical nonparametric mapping (SnPM) were performed. Voxels with significant differences (p < .01, corrected for multiple comparisons) between contrasted conditions were located in the MNI-brain www.unizh.ch/keyinst/NewLORETA/sLORETA/sLORETA.htm.

23

2.5 Statistics

24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

For the descriptive data the mean and standard error of the mean is given. To test all variables included in the analysis for normal distribution a Shapiro-Wilk test was used. The behavioral data (reaction times (RTs) for GO, SCD0 and SCD300 trials as well as error rates in percent) and the neurophysiological data (ERP amplitude and latency) were analyzed in repeated measures ANOVAs using the factor “SCD interval” as within-subject factor. When necessary, the factor “electrode” was also added as within-subject factor. Greenhouse-Geisser correction was used throughout the analyses and post-hoc tests were Bonferroni corrected when necessary. For all variables that did not meet the assumptions of normal distribution the nonparametric Friedman test was used followed by the Wilcoxon signed-rank test with Bonferroni correction for post-hoc analysis. To integrate the pupil diameter data and the EEG data, Pearson correlations were calculated between the pupil diameter data and the EEG data. To this end, the trials in the EEG and pupil diameter data were first averaged for each single participant. Correlations were then calculated across participants on the basis of this averaged data (i.e. correlation were not calculated within subject on a trial by trial basis). Using Pearson correlations we correlated the amplitude values at each time bin on the ERP curve with the amplitude values at each time point on the pupil diameter curve. Each time bin had a length of 4 ms. This is the length of each sampling point after down-sampling the data to a resolution of 256 Hz (refer EEG and pupil diameter processing steps of the data above).

42

3. Results

43

3.1 Behavioural data

7

1 2 3 4 5 6 7 8 9 10 11 12 13

For the RT data, the Friedman test revealed that RTs differed significantly between trial types (χ²(2) = 32.24; p < .001). Bonferroni corrected post-hoc analysis with Wilcoxon signed-rank tests were conducted (significance level of p < 0.017). RTs were significantly smaller in GO trials (679 ms ± 33) than in SCD0 trials (997 ms ± 43; Z = -4.076; p < .001; r = .815) and SCD300 trials (814 ms ± 42; Z = -2.381; p = .017; r = .476). RTs in SCD0 and SCD300 also differed significantly (-4.372; p < .001; r = .874). The Friedman test showed that besides RTs also the error rates differed between trial types (χ²(2) = 50; p < .001). As revealed by post-hoc Wilcoxon signed-rank tests with Bonferroni correction applied (p < .017), error rates differed significantly between GO and SCD0 trials (Z = -4.372; p < .001; r = .874), GO and SCD300 trials (Z = -4.373; p < .001; r = .875) as well as SCD0 and SCD300 trials (Z = -4.374; p < .001; r = .875). Participants committed more errors in SCD0 trials (41.1 % ± 12) than in SCD300 trials (20.5 % ± 22) and were most accurate in GO trials (5.1 % ± 7). The mean stop signal reaction time (SSRT) was 275 ms (± 65).

14 15

3.2 EEG data and pupil diameter data

16

The EEG data and pupil diameter data are shown in Figure 2.

17 18 19 20 21 22 23 24 25 26 27 28 29

For P1 amplitudes at electrode P7 and P8 all variables were normally distributed, as revealed by a Shapiro-Wilk test (S-W > .937; df = 25; p > .127). The ANOVA revealed a main effect of trial type (F(1,24) = 60.34; p < .001; η² = .72). P1 amplitudes were higher in SCD300 trials (5.12 µV/m² ± 0.54) compared to SCD0 trials (4.52 µV/m² ± 0.62). All other effects were not significant (F < 0.88; p > .36). The Friedman test showed that the P1 latencies differed between electrodes and trial type (χ² = 10.120; p = .018). For post-hoc analysis, Wilcoxon sign-ranked tests were calculated with Bonferroni correction applied (confidence level p < .008). P1 latencies were shortest in SCD300 trials at electrode P8 (112.66 ± 2.74), followed by SCD0 at P8 (120.47 ± 3.11) as well as SCD 300 at P7 (120.78 ± 3.93) and longest in SOA0 trials at electrode P7 (129.22 ± 3.83). Though, the P1 latency differences were only significant between SCD0 trials at electrode P7 and SCD300 trials at electrode P8 (Z = 2.875; p = .004; r = .575). All other differences were not significant or did not survive Bonferroni correction (Z > -2.230; p > .026).

30 31 32 33 34 35 36 37 38 39 40 41

The N1 amplitude data was distributed normally as revealed by a Shapiro-Wilk test (S-W > .926; df = 25; p > .070). A significant main effect of trial type (F(1,24) = 34.87; p < .001; η² = .59) indicated that N1 amplitudes were smaller in SCD0 trials (-8.38 µV/m² ± 0.66) in comparison with SCD300 trials (-5.88 µV/m² ± 0.55). The main effect of electrode was not significant (F(1,24) = 1.25; p = .28). There was an interaction effect of electrode and trial type (F(1,24) = 9.51; p = .005; η² = .28). Post-hoc t-tests with the SCD effect (SCD0 amplitudes minus SCD300 amplitudes; confidence level p < .025) showed that the SCD effect is higher for the P7 electrode (-3.25 µV/m² ± 0.51) compared to the P8 electrodes (-1.76 µV/m² ± 0.46); t(24) = -3.08; p = .005; d = .617. The Friedman test revealed no significant effects for N1 latencies (χ² = 3.603; p = .308). The source localization analysis using sLORETA shows that the amplitude differences in the N1 between the SCD0 and the SCD300 condition were due to activation differences in the cuneus (BA19) (Figure 2D).

42 43

For the P3 amplitudes, the Shapiro-Wilk test revealed that all variables were distributed normally (S-W > .949; df = 25; p > .234). As shown by a significant main effect of trial type 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14

(F(1,24) = 58.70; p < .001; η² = .71), the Stop-process P3 amplitudes at electrode Cz differed significantly between SCD0 (14.71 µV/m² ± 1.33) and SCD300 (5.20 µV/m² ± 0.54). Additionally, the mean Change process P3 amplitude of the SCD300 trials (8.53 µV/m² ± 0.93) differed significantly from the P3 of the SCD0 trials (F(1,24) = 55.45; p < .001; η² = .70). For P3 latency, the Friedman test revealed that the latencies differed significantly (χ²(2) = 49.515; p < .001). Post-hoc Bonferroni corrected Wilcoxon signed-rank tests (confidence level p < .017) showed that latency differed significantly between SCD0 P3 and SCD300 Stop P3 (Z = -4.288; p < .001; r = .858), SCD0 P3 and SCD300 Change P3 (Z = -4.381; p < .001; r = .876) as well as between SCD300 Stop P3 and Change P3 (Z = -4.373; p < .001; r = .875). The Change P3 in SCD300 trials (535.16 ms ± 7.52) peaked significantly later than Stop P3 in SCD300 trials (341.88 ms ± 11.69) as well as the P3 in SCD0 trials (229.84 ms ± 6.37). The source localization analysis using sLORETA shows that the amplitude differences in the P3 between the SCD0 and the SCD300 condition were due to activation differences in the anterior cingulate cortex (BA24, BA32) (Figure 2D).

15 16 17 18 19 20 21 22 23 24

Concerning the pupil data (refer Figure 2C), all variables were normally distributed as revealed by a Shapiro-Wilk test (S-W > .923; df = 25; p > .061). The repeated measures ANOVA showed that the initial decrease in pupil diameter just after the presentation of the STOP stimulus (time point 0) was 75.62 ms later in SCD300 trials (301.17 ms ± 35.59) than in SCD0 trials (225.55 ms ± 27.47) (F(1,24) = 12.20; p = .002; η² = .34). There was no SCD effect on pupil diameter (F = 0.57; p = .456). For the latency of the consecutive increase in pupil diameter relative to the baseline (i.e. Lx max) (refer Figure 2C) the ANOVA showed a significant main effect of SCD interval (F(1,24) = 18.82; p < .001; η² = .57). The Lx max emerged 259 ms later in SCD300 trials (1658.28 ms ± 32.13) than in SCD0 trials (1399.14 ms ± 38.38). There was no SCD interval effect on the pupil size (F = 1.13; p = .299).

25

3.3 Integration of behavioral data and electrophysiological data

26 27 28

Pearson correlation analysis revealed no significant correlations of P3 amplitudes and RT data (p > .24). This was not contributed for the error rates because they were affected by the applied staircase procedure.

29 30

3.4 Integration of behavioral data and pupil diameter data

31 32

Pearson correlation analysis revealed no significant correlations of pupil diameter and behavioral data (p > .07).

33 34

3.5 Integration of EEG and pupil diameter data

35 36 37 38 39 40

An important variable also varying from trial to trial is the stop-signal delay (SSD), which was adjusted using the staircase procedure. The initial Go stimulus, shown on every trial, likely induces a pupil response which should evolve with a slow time-course. Because the STOP stimulus was delivered shortly after the Go stimulus, pupil diameter data in the period before the STOP stimulus may be at least partly determined by the SSD. To test this, the SSD was correlated with the pupil diameter data. The results are shown in Figure 3. 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

As can be seen in Figure 3, there were substantial correlations between SSD and pupil diameter in the fore period of the STOP stimulus, i.e. between -400 ms (before) the STOP stimulus presentation and 100 ms after STOP stimulus presentation. Strongest correlations were obtained 100 to 200 ms before STOP stimulus presentation (r = .61; R2 = 0.36). For the subsequent analyses integrating the pupil diameter data with the ERP data, this means that every correlation of the pupil diameter data and the ERP data in the time period -400 ms (before) the STOP stimulus presentation and 100 ms after STOP stimulus presentation is biased by the SSD and therefore cannot be interpreted reliably. However, from 100 ms after STOP stimulus presentation onwards no correlations between SSD and pupil diameter were obtained until 1000 ms after STOP stimulus presentation, in both SCD conditions. Correlations between pupil diameter data and ERP data in this time interval are therefore unbiased by the SSD. In Figure 4 and 5, showing the results of the pupil diameter and ERP data integration, the time periods that are biased due to the SSD are shaded. The results on the Pearson correlations run to integrate the ERP data with the pupil diameter data are shown in Figure 4 for the P1/N1 data and in Figure 5 for the P3 data. In each of these Figures, part A reveals the p-values and part B the degree of correlation (r values).

17 18 19 20 21 22 23 24 25

For the P1/N1 data (averaged over electrodes P7 and P8) the analysis shows positive correlations (between r = .5 (R2 = 0.25) and r = .6 (R2 = 0.36) indicating that larger pupil diameter was correlated with higher amplitude of the ERP at almost every time point in the SCD0 and the SCD300 condition including the N1 ERP in the period which is biased by the SSD. These correlations can therefore not be interpreted. Little to no correlations were obtained for the time course after the presentation of the STOP signal. However, correlations were obtained for the ERP amplitude prior to the P1 time window as well as the time period between 500-1000 ms after the STOP signal presentation and the pupil diameter between 500 to 1000 ms after STOP signal presentation.

26 27 28 29 30 31 32 33 34 35 36 37 38

For the P3, the data analysis revealed the following: Significant inverse correlations (between r = -.4 (R2 = 0.16) and r = -.5 (R2 = 0.25) were also obtained for the pupil diameter between about 500 and 1100 ms and ERP amplitudes in the period before STOP stimulus presentation and in the time range just after the descending slope of the P3 potential; i.e. between 450 and 1000 ms in the SCD0 condition as well as 600 and 1000 ms in the SCD300 condition. Interestingly, in the time window spanning the ascending and descending slope of the P3 potential no significant correlations with the pupil diameter were obtained in the SCD0 and SCD300 condition. Further correlations were obtained between the pupil diameter in the period prior to the presentation of the STOP signal and the amplitude of the P3 from 250 ms post STOP stimulus presentation onwards in the SCD0 condition and in the SCD300 condition (refer Figure 5A) (r = -.45 (R2 = 0.20) and r = -.6 (R2 = 0.36)). However, this correlation was biased by the SSD. There were generally no significant within-subject correlations, i.e. between single trials ERP and pupil diameter data (p > .4).

39

4. Discussion

40 41 42 43 44 45

In the current study we examined the importance of the LC-NE system for multi-component behavior (action cascading). To this end we performed a system neurophysiological study combining high-density EEG recordings and source localization analyses with recordings of the pupil diameter while participants performed a stop-change paradigm. The behavioral data are in line with numerous previous results on this task, i.e. that RTs on the CHANGE stimuli were longer in the SCD0, than in the SCD300 condition. This modulation likely reflects 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

processing resource restrictions in response selection mechanisms (Mückschel et al., 2014; Verbruggen et al., 2008b). The neurophysiological data showed that the P3 was larger in the SCD0 than in the SCD300 condition and this difference was related to activity changes in anterior cingulate cortex (ACC) in source localization analyses (see also Gohil et al., 2015; Mückschel et al., 2014). The P3 modulation observed likely reflects a strong interference between STOP and CHANGE action goals at a strategic response selection bottleneck (e.g. Verbruggen et al., 2008). Likely, inhibitory control processes, needed to manage the stopping and changing of responses, are intensified when the STOP and the CHANGE processes are simultaneously demanding response selection resources, which has been shown to be related to higher P3 amplitudes (overview: Huster et al., 2013b). The modulations in the P1/N1 amplitudes suggests that early perceptual gating and attention selection processes, which are typically reflected by these ERPs (e.g. Gohil et al., 2015; Herrmann and Knight, 2001) are intensified since the SCD0 condition requires simultaneous multisensory integration (Gohil et al., 2016b, 2015) of concurrent visual and tactile stimulus information, which is known to influence early ERPs (Molholm et al., 2002; Murray et al., 2005).

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Concerning the pupil diameter data as a proximate for the activity/relevance of the LC-NE system for action cascading processes and experimental variations in the complexity of these processes (i.e. SCD interval variations) the analyses show that there were no differences in the amplitude parameter (i.e. the pupil diameter), but in the latency of the maximal pupil diameter change between 1200 and 1600 ms after STOP stimulus presentation. The finding that the peak of the pupil diameter occurred approximately 250 ms later in the SCD300 condition than in the SCD0 condition well reflects manipulations of the SCD interval. This latency effect likely reflects the delayed activation of the change process in the SOA 300 condition. Though the NE system has been found to be critically involved in higher order cognitive flexibility processes (Kehagia et al., 2010; but see Steenbergen et al., 2015), the relation between NE system activity and P3 modulations may not be as strong as for example assumed by the Nieuwenhuis et al. (2005). This is further underlined by the finding that the pupil diameter data was not correlated with behavioral performance. As discussed below, the analysis shows that the pupil diameter was also not correlated with the ERP in the time frame of the P3. However, we could not find a link between P3 ERP and behavioral performance, though previous studies, yet with much larger sample size, showed that the P3 is predictive for behavioral performance (e.g. Beste et al., 2014; Mückschel et al., 2014; Stock et al., 2014).

33

4.1 ERP and pupil diameter correlations

34 35 36 37 38 39 40 41 42 43 44 45 46

The link between LC-NE system activity and ERPs was further analyzed by correlating P3 ERP data at electrode CZ and the pupil diameter (refer Figure 5). Here, during the time interval of 200 to 500 ms after STOP stimulus presentation no correlations with the pupil diameter data were obtained in both SCD condition. In both SCD conditions, however, significant correlations between the EEG data before STOP stimulus presentation and the pupil diameter data in the period between 500 and 1000 ms after STOP signal presentation were observed. Similar correlations were obtained for EEG data for the time period approximately 500 ms after STOP stimulus presentation, which should be unbiased by the SSD. Joshi et al. (2016) showed that the pupil diameter as an index of LC-NE activity typically has a response latency of about 400-500 ms with a peak around 1000 ms. Given this pupil response latency, the correlation patterns suggest that there was no direct link between the activity of the LC-NE system and the processes reflected by the P3 during the time period in which the P3 reached its maximum. If the LC-NE system would directly modulate the P3 11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

processes, significant correlations should have emerged in the pupil diameter time range starting from 1000 ms after the STOP signal. Rather, it seems that the neurophysiological processes in the fore period of a possibly upcoming STOP stimulus were linked to the LC-NE system activity, as indicated by the pupil diameter correlation pattern 500 to 1000 ms after the onset of the STOP stimulus. After the presentation of the GO stimulus the participants can expect the occurrence of the STOP stimulus in 33% of cases. To efficiently process the STOP and related CHANGE stimuli it is advantageous to prepare cognitive control processes needed for stopping and changing (i.e. action cascading) processes. These processes are reflected by the P3 ERP (see also: Dippel and Beste, 2015; Mückschel et al., 2014). This fits well with the literature on the functional role of the LC-NE system put forward by Nieuwenhuis et al. (2005) stating that the P3 is an electrophysiological correlate of the LC phasic response and enhances processing in target areas (Nieuwenhuis et al., 2005). Though, because the time intervals showing correlations were almost the same for the SCD0 and the SCD300 condition and did not vary with the stop-change delay, it seems that it is primarily the STOP process rather than the CHANGE process that is likely to be prepared by the LC-NE activity. Corroborating this interpretation, other studies have shown that pharmacological modulations of the LC-NE system affect neurophysiological processes reflected by the P3 ERP component during stopping (Logemann et al., 2013). The results support findings that the modulation by the NE system affects action monitoring (e.g. Geva et al., 2013; Riba et al., 2005) and that the pupil diameter may be an index of the cognitive control state (Gilzenrat et al., 2010). The results may be interpreted that the LC-NE system facilitates responses to task-relevant processes and supports task-related decision and response selection processes. This, however, may be achieved by preparing cognitive control processes for the case that these are required during multi-component behavior but to a lesser extent by modulating these processes once they are operating. The results from the correlation analysis using posterior electrodes (P1/N1) support the interpretation of a preparatory role of the LC-NE system. As can be seen in Figure 4, correlations unbiased by the SSD were evident in similar time ranges as the correlations with the P3 data; neuronal (EEG) activity before the presentation of the STOP stimulus was correlated with pupil diameter in the time range between 500 ms to 1000 ms after STOP stimulus presentation.

31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

The above interpretation that the LC-NE serves functions to prepare the cognitive control system relates to the "modified LC-P3" theory by Warren et al. (Warren et al., 2011; Warren and Holroyd, 2012). Due to this account LC bursts impact cortical activity earlier than during the time frame of the P3, about 250 ms post stimulus, and it is the LC refractory period that coincides with the P3 generation. According to this modified account especially infrequent events elicit phasic NE release (Warren and Holroyd, 2012). The finding that correlations between ERPs and pupil diameter were strongest shortly after the relevant stimulus (~300 ms) and spared for the time interval in which the P3 occurred is in line with this account. However, this modified account proposes that this relates to the N2 ERP, but in the paradigm applied no N2 is evident (Mückschel et al., 2014). The negativity seen before the P3 (refer e.g. figure 2B) cannot be regarded as a N2 since it has an occipital negativity in the topography (and not a fronto-central distribution) and is also too early for a typical N2. It may rather reflect a remote effect of the occipital N1 on the STOP stimulus (Mückschel et al., 2014). Yet, the N2 may be superimposed by processes related to the selection of the appropriate response on the CHANGE stimuli. Such processes have repeatedly been shown to be reflected by the P3 (Twomey et al., 2015; Verleger et al., 2015).

12

1 2 3 4 5 6

It needs to be noted that the correlations obtained were between-subject and not withinsubject correlations. However, the absence of a significant correlation measured between subjects does exclude the possibility that variations in pupil size at the level of the single trial (thus within subjects) could be predictive of performance. This is in line with other recent results (Chmielewski et al., 2016) also showing no within-subject correlations, likely because of the low signal-to-noise ratio of ERP and pupil diameter data in single trials.

7

4.2 Conclusion

8 9 10 11 12 13

In summary, the study shows that the LC-NE system is important for processes related to multi-component behavior. Yet, the influence of the LC-NE system on the P3 appears to be less pronounced than postulated by previous studies. The LC-NE system may facilitate responses to task-relevant processes and support task-related decision and response selection processes. This may be obtained by preliminary preparation of cognitive control processes for later use but not by direct modulation of these processes once they are active.

14 15

Acknowledgements

16 17

This work was supported by grants from the Deutsche Forschungsgemeinschaft (DFG) BE4045/10-2 and SFB 940 project B8 to C.B.

18 19

Literature

20 21 22

Aston-Jones, G., Cohen, J.D., 2005. An integrative theory of locus coeruleus-norepinephrine function: adaptive gain and optimal performance. Annu. Rev. Neurosci. 28, 403–450. doi:10.1146/annurev.neuro.28.061604.135709

23 24 25

Beste, C., Stock, A.-K., Epplen, J.T., Arning, L., 2014a. On the relevance of the NPY2receptor variation for modes of action cascading processes. NeuroImage 102, Part 2, 558–564. doi:10.1016/j.neuroimage.2014.08.026

26 27 28

Beste, C., Stock, A.-K., Epplen, J.T., Arning, L., 2014b. On the relevance of the NPY2receptor variation for modes of action cascading processes. NeuroImage. doi:10.1016/j.neuroimage.2014.08.026

29 30 31 32

Chmielewski, W.X., Mückschel, M., Ziemssen, T., Beste, C., 2016. The norepinephrine system affects specific neurophysiological subprocesses in the modulation of inhibitory control by working memory demands. Hum. Brain Mapp. n/a-n/a. doi:10.1002/hbm.23344

33 34 35

Corbetta, M., Patel, G., Shulman, G.L., 2008. The Reorienting System of the Human Brain: From Environment to Theory of Mind. Neuron 58, 306–324. doi:10.1016/j.neuron.2008.04.017

13

1 2 3

Delorme, A., Makeig, S., 2004. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J. Neurosci. Methods 134, 9–21. doi:10.1016/j.jneumeth.2003.10.009

4 5 6

Dimigen, O., Sommer, W., Hohlfeld, A., Jacobs, A.M., Kliegl, R., 2011. Coregistration of eye movements and EEG in natural reading: Analyses and review. J. Exp. Psychol. Gen. 140, 552–572. doi:10.1037/a0023885

7 8

Dippel, G., Beste, C., 2015. A causal role of the right inferior frontal cortex in the strategies of multi-component behaviour. Nat. Commun. doi:10.1038/ncomms7587

9 10 11

Duncan, J., 2010. The multiple-demand (MD) system of the primate brain: mental programs for intelligent behaviour. Trends Cogn. Sci. 14, 172–179. doi:10.1016/j.tics.2010.01.004

12 13 14

Fuchs, M., Kastner, J., Wagner, M., Hawes, S., Ebersole, J.S., 2002. A standardized boundary element method volume conductor model. Clin. Neurophysiol. Off. J. Int. Fed. Clin. Neurophysiol. 113, 702–712.

15 16 17

Geva, R., Zivan, M., Warsha, A., Olchik, D., 2013. Alerting, orienting or executive attention networks: differential patters of pupil dilations. Front. Behav. Neurosci. 7, 145. doi:10.3389/fnbeh.2013.00145

18 19 20

Gilzenrat, M.S., Nieuwenhuis, S., Jepma, M., Cohen, J.D., 2010. Pupil diameter tracks changes in control state predicted by the adaptive gain theory of locus coeruleus function. Cogn. Affect. Behav. Neurosci. 10, 252–269. doi:10.3758/CABN.10.2.252

21 22 23

Gohil, K., Dippel, G., Beste, C., 2016a. Questioning the role of the frontopolar cortex in multi-component behavior – a TMS/EEG study. Sci. Rep. 6, 22317. doi:10.1038/srep22317

24 25 26

Gohil, K., Hahne, A., Beste, C., 2016b. Improvements of sensorimotor processes during action cascading associated with changes in sensory processing architecture-insights from sensory deprivation. Sci. Rep. 6, 28259. doi:10.1038/srep28259

27 28

Gohil, K., Stock, A.-K., Beste, C., 2015. The importance of sensory integration processes for action cascading. Sci. Rep. 5, 9485. doi:10.1038/srep09485

29 30 31

Herrmann, C.S., Knight, R.T., 2001. Mechanisms of human attention: event-related potentials and oscillations. Neurosci. Biobehav. Rev. 25, 465–476. doi:10.1016/S01497634(01)00027-6

32 33 34

Hong, L., Walz, J.M., Sajda, P., 2014. Your eyes give you away: prestimulus changes in pupil diameter correlate with poststimulus task-related EEG dynamics. PloS One 9, e91321. doi:10.1371/journal.pone.0091321

35 36 37 38

Hou, R.H., Freeman, C., Langley, R.W., Szabadi, E., Bradshaw, C.M., 2005. Does modafinil activate the locus coeruleus in man? Comparison of modafinil and clonidine on arousal and autonomic functions in human volunteers. Psychopharmacology (Berl.) 181, 537–549. doi:10.1007/s00213-005-0013-8 14

1 2 3 4

Huster, R.J., Enriquez-Geppert, S., Lavallee, C.F., Falkenstein, M., Herrmann, C.S., 2013a. Electroencephalography of response inhibition tasks: functional networks and cognitive contributions. Int. J. Psychophysiol. Off. J. Int. Organ. Psychophysiol. 87, 217–233. doi:10.1016/j.ijpsycho.2012.08.001

5 6 7 8 9

Huster, R.J., Enriquez-Geppert, S., Lavallee, C.F., Falkenstein, M., Herrmann, C.S., 2013b. Electroencephalography of response inhibition tasks: Functional networks and cognitive contributions. Int. J. Psychophysiol., Electrophysiological and Neuroimaging Studies of Cognitive Control: Evidence from Go/NoGo and Other Executive Function Tasks 87, 217–233. doi:10.1016/j.ijpsycho.2012.08.001

10 11 12

Jepma, M., Nieuwenhuis, S., 2010. Pupil Diameter Predicts Changes in the Exploration– Exploitation Trade-off: Evidence for the Adaptive Gain Theory. J. Cogn. Neurosci. 23, 1587–1596. doi:10.1162/jocn.2010.21548

13 14 15

Joshi, S., Li, Y., Kalwani, R.M., Gold, J.I., 2016. Relationships between Pupil Diameter and Neuronal Activity in the Locus Coeruleus, Colliculi, and Cingulate Cortex. Neuron 89, 221–234. doi:10.1016/j.neuron.2015.11.028

16 17 18

Kehagia, A.A., Murray, G.K., Robbins, T.W., 2010. Learning and cognitive flexibility: frontostriatal function and monoaminergic modulation. Curr. Opin. Neurobiol., Cognitive neuroscience 20, 199–204. doi:10.1016/j.conb.2010.01.007

19 20

Lisi, M., Bonato, M., Zorzi, M., 2015. Pupil dilation reveals top–down attentional load during spatial monitoring. Biol. Psychol. 112, 39–45. doi:10.1016/j.biopsycho.2015.10.002

21 22

Logan, G.D., Cowan, W.B., 1984. On the ability to inhibit thought and action: A theory of an act of control. Psychol. Rev. 91, 295–327. doi:10.1037/0033-295X.91.3.295

23 24 25

Logemann, H.N.A., Böcker, K.B.E., Deschamps, P.K.H., Kemner, C., Kenemans, J.L., 2013. The effect of noradrenergic attenuation by clonidine on inhibition in the stop signal task. Pharmacol. Biochem. Behav. 110, 104–111. doi:10.1016/j.pbb.2013.06.007

26 27 28

Marco-Pallarés, J., Grau, C., Ruffini, G., 2005. Combined ICA-LORETA analysis of mismatch negativity. NeuroImage 25, 471–477. doi:10.1016/j.neuroimage.2004.11.028

29 30 31 32

Molholm, S., Ritter, W., Murray, M.M., Javitt, D.C., Schroeder, C.E., Foxe, J.J., 2002. Multisensory auditory–visual interactions during early sensory processing in humans: a high-density electrical mapping study. Cogn. Brain Res., Multisensory Proceedings 14, 115–128. doi:10.1016/S0926-6410(02)00066-6

33 34

Mückschel, M., Stock, A.-K., Beste, C., 2015. Different strategies, but indifferent strategy adaptation during action cascading. Sci. Rep. 5. doi:10.1038/srep09992

35 36 37

Mückschel, M., Stock, A.-K., Beste, C., 2014. Psychophysiological Mechanisms of Interindividual Differences in Goal Activation Modes During Action Cascading. Cereb. Cortex 24, 2120–2129. doi:10.1093/cercor/bht066

15

1 2 3

Murphy, P.R., Robertson, I.H., Balsters, J.H., O’connell, R.G., 2011. Pupillometry and P3 index the locus coeruleus–noradrenergic arousal function in humans. Psychophysiology 48, 1532–1543. doi:10.1111/j.1469-8986.2011.01226.x

4 5 6 7

Murray, M.M., Molholm, S., Michel, C.M., Heslenfeld, D.J., Ritter, W., Javitt, D.C., Schroeder, C.E., Foxe, J.J., 2005. Grabbing Your Ear: Rapid Auditory–Somatosensory Multisensory Interactions in Low-level Sensory Cortices Are Not Constrained by Stimulus Alignment. Cereb. Cortex 15, 963–974. doi:10.1093/cercor/bhh197

8 9 10

Nieuwenhuis, S., Aston-Jones, G., Cohen, J.D., 2005. Decision making, the P3, and the locus coeruleus--norepinephrine system. Psychol. Bull. 131, 510–532. doi:10.1037/00332909.131.4.510

11 12 13

Nunez, P.L., Pilgreen, K.L., 1991. The spline-Laplacian in clinical neurophysiology: a method to improve EEG spatial resolution. J. Clin. Neurophysiol. Off. Publ. Am. Electroencephalogr. Soc. 8, 397–413.

14 15 16

Pascual-Marqui, R.D., 2002. Standardized low-resolution brain electromagnetic tomography (sLORETA): technical details. Methods Find. Exp. Clin. Pharmacol. 24 Suppl D, 5– 12.

17 18

Perrin, F., Pernier, J., Bertrand, O., Echallier, J.F., 1989. Spherical splines for scalp potential and current density mapping. Electroencephalogr. Clin. Neurophysiol. 72, 184–187.

19 20

Petersen, S.E., Posner, M.I., 2012. The Attention System of the Human Brain: 20 Years After. Annu. Rev. Neurosci. 35, 73–89. doi:10.1146/annurev-neuro-062111-150525

21 22 23

Phillips, M.A., Szabadi, E., Bradshaw, C.M., 2000. Comparison of the effects of clonidine and yohimbine on spontaneous pupillary fluctuations in healthy human volunteers. Psychopharmacology (Berl.) 150, 85–89.

24 25

Posner, M.I., Petersen, S.E., 1990. The Attention System of the Human Brain. Annu. Rev. Neurosci. 13, 25–42. doi:10.1146/annurev.ne.13.030190.000325

26 27 28 29 30

Raedt, R., Clinckers, R., Mollet, L., Vonck, K., El Tahry, R., Wyckhuys, T., De Herdt, V., Carrette, E., Wadman, W., Michotte, Y., Smolders, I., Boon, P., Meurs, A., 2011. Increased hippocampal noradrenaline is a biomarker for efficacy of vagus nerve stimulation in a limbic seizure model. J. Neurochem. 117, 461–469. doi:10.1111/j.1471-4159.2011.07214.x

31 32 33

Riba, J., Rodríguez-Fornells, A., Morte, A., Münte, T.F., Barbanoj, M.J., 2005. Noradrenergic Stimulation Enhances Human Action Monitoring. J. Neurosci. 25, 4370–4374. doi:10.1523/JNEUROSCI.4437-04.2005

34 35 36

Sekihara, K., Sahani, M., Nagarajan, S.S., 2005. Localization bias and spatial resolution of adaptive and non-adaptive spatial filters for MEG source reconstruction. NeuroImage 25, 1056–1067. doi:10.1016/j.neuroimage.2004.11.051

37 38

Steenbergen, L., Sellaro, R., Stock, A.-K., Verkuil, B., Beste, C., Colzato, L.S., 2015. Transcutaneous vagus nerve stimulation (tVNS) enhances response selection during 16

1 2

action cascading processes. Eur. Neuropsychopharmacol. 25, 773–778. doi:10.1016/j.euroneuro.2015.03.015

3 4 5

Stock, A.-K., Arning, L., Epplen, J.T., Beste, C., 2014. DRD1 and DRD2 genotypes modulate processing modes of goal activation processes during action cascading. J. Neurosci. Off. J. Soc. Neurosci. 34, 5335–41. doi:10.1523/JNEUROSCI.5140-13.2014

6 7 8

Stock, A.-K., Gohil, K., Beste, C., 2015. Age-related differences in task goal processing strategies during action cascading. Brain Struct. Funct. doi:10.1007/s00429-015-10712

9 10 11

Twomey, D.M., Murphy, P.R., Kelly, S.P., O’Connell, R.G., 2015. The classic P300 encodes a build-to-threshold decision variable. Eur. J. Neurosci. 42, 1636–1643. doi:10.1111/ejn.12936

12 13 14

Van Leusden, J.W.R., Sellaro, R., Colzato, L.S., 2015. Transcutaneous Vagal Nerve Stimulation (tVNS): a new neuromodulation tool in healthy humans? Front. Psychol. 6. doi:10.3389/fpsyg.2015.00102

15 16 17

Verbruggen, F., Logan, G.D., 2009. Models of response inhibition in the stop-signal and stopchange paradigms. Neurosci. Biobehav. Rev. 33, 647–661. doi:10.1016/j.neubiorev.2008.08.014

18 19 20

Verbruggen, F., Schneider, D.W., Logan, G.D., 2008. How to stop and change a response: the role of goal activation in multitasking. J. Exp. Psychol. Hum. Percept. Perform. 34, 1212–1228. doi:10.1037/0096-1523.34.5.1212

21 22 23

Verleger, R., Hamann, L.M., Asanowicz, D., Śmigasiewicz, K., 2015. Testing the S–R link hypothesis of P3b: The oddball effect on S1-evoked P3 gets reduced by increased task relevance of S2. Biol. Psychol. 108, 25–35. doi:10.1016/j.biopsycho.2015.02.010

24 25 26

Warren, C.M., Holroyd, C.B., 2012. The Impact of Deliberative Strategy Dissociates ERP Components Related to Conflict Processing vs. Reinforcement Learning. Front. Neurosci. 6, 43. doi:10.3389/fnins.2012.00043

27 28 29

Warren, C.M., Tanaka, J.W., Holroyd, C.B., 2011. What can topology changes in the oddball N2 reveal about underlying processes? Neuroreport 22, 870–874. doi:10.1097/WNR.0b013e32834bbe1f

30 31 32 33

Yildiz, A., Quetscher, C., Dharmadhikari, S., Chmielewski, W., Glaubitz, B., SchmidtWilcke, T., Edden, R., Dydak, U., Beste, C., 2014. Feeling safe in the plane: Neural mechanisms underlying superior action control in airplane pilot trainees—A combined EEG/MRS study. Hum. Brain Mapp. n/a-n/a. doi:10.1002/hbm.22530

34 35 36

Figure Legends

17

1 2 3 4 5 6

Figure 1 Illustration of the Stop-Change task used in this study. GO trials end after the first response to the GO stimulus. SC trials end after the first response to the CHANGE signal. The stop-signal delay (SSD) between the onset of GO stimulus and STOP signal was adjusted using a staircase procedure. The CHANGE signal was presented after a stop-change delay (SCD) of either 0 ms or 300 ms. Different locations and combinations of air puffs were associated with one of the three reference lines.

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Figure 2 The electrophysiological, pupil diameter and sLORETA results. (A) Plot of P1 and N1 ERPs in SCD0 and SCD300 conditions at electrodes P7 and P8 and associated average reference (left) and CSD transformed scalp topographies (right). The grey bar denotes the time range of baseline correction. Time point 0 denotes the STOP stimulus presentation and the dashed line denotes the onset of the CHANGE stimulus. (B) Plot of P3 in SCD0 and SCD300 conditions at electrode Cz and associated scalp topographies (left: average reference; right: CSD transformed) SCD300 topographies are given for the Change P3. The grey bar denotes the time range of baseline correction. Time point 0 denotes the STOP stimulus presentation and the dashed line denotes the onset of the CHANGE stimulus. (C) Plot of pupil diameter data for SCD0 and SCD300 condition. The grey bar denotes the time range of baseline correction. Time point 0 denotes the STOP stimulus presentation and the dashed line denotes the onset of the CHANGE stimulus. (D) Results of source localization analysis using sLORETA. Sagittal view of the activation differences obtained by sLORETA analysis for N1 and P3 ERPs contrasting SCD300 and SCD0 condition. Only significant activation differences (p < .01) as obtained on the basis of randomization tests are depicted. For N1 significant activation differences were found in the cuneus. For P3 significant activation differences were found in the anterior cingulate cortex (ACC).

24 25 26 27

Figure 3 Correlation plots for the SSD with the pupil diameter data in the SCD0 (left) and the SCD300 condition (right). Strong negative correlations are obtained in the time interval from 400 ms prior STOP stimulus presentation to 100 ms thereafter in both SCD conditions.

28 29 30 31 32 33 34 35 36 37 38 39

Figure 4 Correlation plots of averaged P7 and P8 electrodes and the pupil diameter, separately for SCD0 and SCD300 in the time range from -500 ms before until 2000 ms after STOP stimulus presentation. The white lines denote the time point of the stop stimulus presentation. Time intervals in which the correlation between pupil diameter and ERP data is biased by the SSD are shaded. The vertical line plots depict the averaged signal of electrodes P7 and P8 for SCD0 (left) and SCD300 (right). The horizontal plots depict the pupil diameter data for SCD0 (left) and SCD300 (right). The grey shaded lines denote the other SCD condition; i.e. in the column showing the SCD0 condition, the ERP and pupil diameter trace of the SCD0 condition are shown in blue, the ones of the SCD300 condition in grey. In the SCD300 column of the figure this is shown vice versa. (A) Colors denote significant p-values. (B) Colors denote r-values shown for the significant cluster revealed in Figure part A.

40 41 42

Figure 5 Correlation plots of averaged P7 and P8 electrodes and the pupil diameter, separately for SCD0 and SCD300 in the time range from -500 ms before until 2000 ms after STOP 18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

stimulus presentation. The white lines denote the time point of the stop stimulus presentation. Time intervals in which the correlation between pupil diameter and ERP data is biased by the SSD are shaded. The vertical line plots depicts electrode Cz for SCD0 (left) and SCD300 (right). The horizontal plots depict the pupil diameter data for SCD0 (left) and SCD300 (right). The grey shaded lines denote the other SCD condition; i.e. in the column showing the SCD0 condition, the ERP and pupil diameter trace of the SCD0 condition are shown in blue, the ones of the SCD300 condition in grey. In the SCD300 column of the figure this is shown vice versa. (A) Colors denote significant p-values. (B) Colors denote r-values shown for the significant cluster revealed in Figure part A.

17

4. NE system prepares cognitive control processes for multi-component behavior.

Highlight 1. The role of the norepinephrine system for multi-component behavior is examined. 2. Pupil diameter and ERP data are integrated and related to the functional neuroanatomy. 3. The NE-system predicts attentional gating and response selection processes.

19

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5