Sequential inhibitory control processes assessed through simultaneous EEG–fMRI

NeuroImage 94 (2014) 349–359

Contents lists available at ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/ynimg

Sequential inhibitory control processes assessed through simultaneous EEG–fMRI Sarah Baumeister a,1, Sarah Hohmann a,1, Isabella Wolf a,c, Michael M. Plichta b, Stefanie Rechtsteiner a, Maria Zangl b,d, Matthias Ruf c, Nathalie Holz a, Regina Boecker a, Andreas Meyer-Lindenberg b, Martin Holtmann a,e, Manfred Laucht a, Tobias Banaschewski a, Daniel Brandeis a,f,g,h,⁎ a

Department of Child and Adolescent Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim/Heidelberg University, Mannheim, Germany Department of Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim/Heidelberg University, Mannheim, Germany c Department Neuroimaging, Central Institute of Mental Health, Medical Faculty Mannheim/Heidelberg University, Mannheim, Germany d Department of Cognitive and Clinical Neuroscience, Central Institute of Mental Health, Medical Faculty Mannheim/Heidelberg University, Mannheim, Germany e Child and Adolescent Psychiatry, Ruhr-University Bochum, Bochum, Germany f Department of Child and Adolescent Psychiatry, University of Zurich, Zurich, Switzerland g Center for Integrative Human Physiology, University of Zurich, Zurich, Switzerland h Neuroscience Center Zurich, University of Zurich and ETH Zurich, Zurich, Switzerland b

a r t i c l e

i n f o

Article history: Accepted 16 January 2014 Available online 25 January 2014 Keywords: Inhibition Simultaneous EEG–fMRI N2 P3 NoGo

a b s t r a c t Inhibitory response control has been extensively investigated in both electrophysiological (ERP) and hemodynamic (fMRI) studies. However, very few multimodal results address the coupling of these inhibition markers. In fMRI, response inhibition has been most consistently linked to activation of the anterior insula and inferior frontal cortex (IFC), often also the anterior cingulate cortex (ACC). ERP work has established increased N2 and P3 amplitudes during NoGo compared to Go conditions in most studies. Previous simultaneous EEG–fMRI imaging reported association of the N2/P3 complex with activation of areas like the anterior midcingulate cortex (aMCC) and anterior insula. In this study we investigated inhibitory control in 23 healthy young adults (mean age = 24.7, n = 17 for EEG during fMRI) using a combined Flanker/NoGo task during simultaneous EEG and fMRI recording. Separate fMRI and ERP analysis yielded higher activation in the anterior insula, IFG and ACC as well as increased N2 and P3 amplitudes during NoGo trials in accordance with the literature. Combined analysis modelling sequential N2 and P3 effects through joint parametric modulation revealed correlation of higher N2 amplitude with deactivation in parts of the default mode network (DMN) and the cingulate motor area (CMA) as well as correlation of higher central P3 amplitude with activation of the left anterior insula, IFG and posterior cingulate. The EEG–fMRI results resolve the localizations of these sequential activations. They suggest a general role for allocation of attentional resources and motor inhibition for N2 and link memory recollection and internal reflection to P3 amplitude, in addition to previously described response inhibition as reflected by the anterior insula. © 2014 Elsevier Inc. All rights reserved.

Introduction Inhibitory control has been a prominent topic in neuroimaging (as reviewed by Swick et al. (2011)) and has been defined to include the ability to suppress actions that are inappropriate in a certain context and that interfere with goal-driven behaviour (Aron, 2007). The concept of inhibitory control has been investigated in numerous studies using a variety of cognitive tasks and neuroimaging methods. However, only

⁎ Corresponding author at: Central Institute of Mental Health, J5, 68159 Mannheim, Germany. E-mail address: [email protected] (D. Brandeis). 1 SB and SH contributed equally to this work. 1053-8119/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neuroimage.2014.01.023

few studies have focused on the combination of more than one imaging modality. The most commonly used tasks to investigate inhibitory control are Go/Nogo, stop signal and conflict tasks. During Go/NoGo tasks, subjects are usually asked to respond to one type of stimulus, while withholding their response to another type of stimulus. Withholding a response following NoGo stimuli is more difficult when Go trials are frequent. Go/NoGo and stop signal tasks have frequently been treated interchangeably (Aron et al., 2004). Conflict- or interference tasks (as e.g. the Stroop task) have also been used extensively by neuroscientists to assess inhibitory control. These kinds of tasks require subjects to discriminate between task relevant and task irrelevant stimuli or stimulus dimensions and, in turn, inhibit the reaction to the task irrelevant aspects. Critics have suggested that many of the experimental effects

350

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

observed during conflict correspond to the facilitated processing of task relevant stimuli rather than to active inhibition (Aron, 2007; Egner and Hirsch, 2005). However, for the common flanker task, where a central target stimulus is usually surrounded by other, irrelevant stimuli which are either compatible, neutral, or incompatible with the target stimulus or response (Eriksen and Eriksen, 1974), cortical motor inhibition appears to play an important role (Klein et al., 2013). The signal of the electroencephalogram (EEG) is directly related to electric neuronal activity and shows a high temporal resolution within the millisecond range, while its spatial precision depends on restrictive assumptions or is limited to imprecise, blurred localizations of the distributed cortical and sometimes subcortical regions involved in cognitive processes (Pascual-Marqui et al., 2009). EEG studies have provided a large database on highly time-resolved neurophysiologic processes during response inhibition in Go/NoGo and stop signal tasks (see review by Huster et al. (2013)). The most prominent findings are enhanced inhibitory or conflict-related components of the event-related potentials (ERPs) during NoGo or Stop trials (the NoGo N2 and the NoGo P3). The N2 is a negative potential increased in frontocentral regions during NoGo compared to Go trials (Review by Folstein and Van Petten (2008)). This association between the amplitude of the NoGo N2 and successful response inhibition has also been reported by Falkenstein et al. (1999). However, the N2 is also increased by conflict processing without inhibition (e.g.Randall and Smith (2011)), and therefore not specific for inhibitory processes. In contrast, the subsequent NoGo P3 seems to be more consistently linked to response inhibition (Bekker et al., 2004; Bruin et al., 2001; Donkers and van Boxtel, 2004), and is characterized by a frontocentral positivity which has been extensively demonstrated in healthy adults (Bokura et al., 2001; Bruin et al., 2001; Fallgatter et al., 1997; Pfefferbaum et al., 1985). The NoGo P3 is increased in participants responding fast when compared to slow responders (Smith et al., 2006). Some other studies found that the amplitude of the NoGo P3 is increased in relation to the level of response preparation (Bekker et al., 2004; Bruin et al., 2001; Smith et al., 2007). For the Eriksen flanker task (Eriksen and Eriksen, 1974), incompatible arrays, where the reaction to the flanking stimuli has to be inhibited, lead to a frontocentral increase of the N2 amplitude when compared to compatible arrays (incongruent vs. congruent condition). In contrast, there is almost no effect on the P3 positivity (Gehring et al., 1992; Kopp et al., 1996; van Veen and Carter, 2002). However, an increase in the latency of the P3 was reported in the incongruent condition (Ridderinkhof and van der Molen, 1995). Both, N2 and P3 have been subject to intensive research in clinical populations. Differences in amplitude or anteriorization of the P3 can be detected e.g. in patients with ADHD and are interpreted as a representation of a persistent neurophysiological deficit, while results for NoGo-N2 are more heterogeneous showing no significant reduction of N2 amplitude in ADHD in many studies (Albrecht et al., 2012; Brandeis et al., 2002; Dhar et al., 2010; Doehnert et al., 2010; Fallgatter et al., 2004, 2005; Valko et al., 2009). While EEG source localization allows high temporal resolution but in most cases only an approximate or blurred localization of inhibition- or conflict-related activity (Fallgatter et al., 1997; Strik et al., 1998), functional magnetic resonance imaging (fMRI) provides consistently high spatial resolution and enables a more precise localization of brain regions engaged during cognitive processes. However, the latter offers only low temporal precision, as the blood-dependent oxygen-level-dependent (BOLD) response is rather slow. Neuroimaging studies have investigated inhibitory control using a variety of tasks and contrasts, but the contrast of successfully inhibited NoGo trials compared to Go trials is most commonly reported. Various studies have yielded activation in areas such as the anterior cingulate cortex (ACC), dorsolateral prefrontal cortex (DLPFC) and inferior frontal cortex (IFC) for this contrast (Menon et al., 2001; Nee et al., 2007). Some studies also report activation in the lingual gyrus and caudate (Menon et al., 2001). However, a recent meta-analysis

by Swick et al. (2011) found activation in the anterior insula to be largest and most significant. Further activation for NoGo compared to Go trials was found in the right middle frontal gyrus, right inferior parietal lobule/precuneus and left middle frontal cortex. The same meta-analysis also compared Go/NoGo to Stop signal tasks, and found overlapping activation in the right insula and superior frontal gyrus, while areas more activated during the Go/NoGo task comprised the right middle and superior frontal gyrus, as well as the right inferior parietal lobule and precuneus. The multimodal approach of simultaneous EEG and fMRI recording integrates the advantages of both imaging modalities. Although important insights have been obtained by separate sequential recordings in the same subject (Halder et al., 2007; Vitacco et al., 2002), simultaneous recordings are needed to ensure the identical cognitive state and to capture spontaneous variation and trial-by-trial coupling of electrical and hemodynamic activity. The aim of integrating both imaging modalities is gaining both high spatial and temporal resolution in the same subject (Debener et al., 2005). Although a number of studies have used this approach, only one has so far investigated inhibitory control (Huster et al., 2011). Using cross-modal correlation and independent component analysis (ICA) to integrate ERP with fMRI data in a stop-signal task they found the stop-related N2/P3 complex to be correlated with activation in the rostral anterior midcingulate cortex (aMCC), pre SMA, anterior insula, putamen and globus pallidus, while the Go-related N2/P3 complex was associated with activation in the ventral anterior and posterior MCC, the left postcentral region, the SMA and deactivation in the occipital gyrus. The present study combined fMRI and ERP measures to investigate spatio-temporal aspects of inhibitory control through single trial parametric modulation as suggested by previous EEG–fMRI studies (Benar et al., 2007; Eichele et al., 2005). In contrast to the previous study by Huster et al. (2011), the N2 and P3 were treated as separate components to model their successive, independent parametric modulations of the fMRI in order to gain new insights into the time course and the specific inhibitory characteristics of their BOLD correlates. Still, our EEG-informed fMRI analysis also represents an asymmetric approach to simultaneous EEG–fMRI integration, while fMRI-informed EEG source analysis, or symmetric multimodal data fusion represent alternative approaches as reviewed in Huster et al. (2012). Material and Methods Subjects The full sample consisted of 23 right-handed healthy subjects (12 male, 11 female) aged between 20 and 35 years (M = 24.70, SD = 4.29). Due to technical difficulties and insufficient EEG-data quality, six subjects had to be excluded from EEG- and combined analyses. The EEG sample therefore consisted of 17 subjects (9 male, 8 female) aged between 20 and 35 years (M = 24.71, SD = 4.15). All subjects gave written informed consent prior to their participation and had normal or corrected-to-normal vision. The study was approved by the Ethics Committee of the Medical Faculty of the Ruprecht-Karls-University Heidelberg. Experimental paradigm In this Flanker/NoGo task (Blasi et al., 2006; Bunge et al., 2002; Meyer-Lindenberg et al., 2006) stimuli consisted of an array of five shapes including a central target arrow pointing either left or right, flanked by two shapes (arrows, squares or Xs) on each side. Subjects were instructed to press a button corresponding to the central arrow when the flankers were also arrows (congruent and incongruent conditions) or boxes (neutral condition), but not when they were Xs (NoGo condition). Flanking arrows were pointing either in the same (congruent) or opposite (incongruent) direction as the central arrow, thus allowing to investigate conflict processing and interference. A total of 145 stimuli (33 NoGo, 31 neutral, 40 incongruent, 41 congruent)

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

were randomly presented for 800 ms with an interstimulus interval (ISI) which varied between 3 and 12 s. During the ISI a fixation cross was presented. The total duration of the task was 10 min 19 s. Data acquisition Data was acquired recording EEG and fMRI simultaneously. A 3 T Siemens Trio whole body scanner (Siemens Medical Solutions, Erlangen, Germany) with a standard 12 channel headcoil was used to obtain functional images via a BOLD-sensitive T2*-weighted echoplanar sequence (repetition time (TR) = 2210 ms, echo time (TE) = 28 ms, flip angle = 90°). A total of 277 volumes with 36 slices with a thickness of 4 mm were obtained, oriented approximately 20° steeper than the AC–PC-plane (Field of View: 220 mm; Matrix: 64 × 64). Subject’s positions in the scanner were axially shifted about 4 cm to reduce gradient artefacts (Mullinger et al., 2011). The first three volumes were discarded to allow longitudinal magnetization to reach equilibrium. Functional measurement was followed by a T1-weighted anatomical MRI scan (192 sagittal slices, slice thickness 1 mm, FOV = 256 mm × 256 mm, matrix = 256 × 256) for each subject. The EEG was continuously recorded from 60 Ag/AgCl electrodes within an extended 10-20 system along with 2 electrooculogram (EOG) and 2 electrocardiogram (ECG) electrodes using MRcompatible BrainAmp MR plus amplifiers (Brain Products, Gilching, Germany). Data was sampled at 5 kHz with a 32 mV input range to facilitate fMRI gradient correction. Recording reference was F1 while ground electrode was located at F2 and impedances were kept below 20 kΩ. Stimuli were created with Presentation Software Package (Neurobehavioural Systems Inc., Albany, CA, USA) and presented via VisuaStim video goggles (Resonance Technology Inc., Northridge, USA), behavioural data was recorded with response buttons (Current Designs, Philadelphia, USA). Behavioural data analysis Reaction times (RT) and percentage of errors were analysed using SPSS Software package (Version 20, IBM Corp., Armonk, NY, USA). Separate analyses of variance (ANOVA) were conducted to test for expected behavioural effects. Reaction times were expected to be different between conditions, with longer reaction times for the incongruent condition. Percentage of errors was also expected to differ between conditions, with more errors during incongruent and NoGo trials. EEG data analysis FMRI gradient and ballistocardiogramm artefacts were removed using standard template subtraction procedures as implemented in the Brain Vision Analyzer 2 software (Brain Products, Gilching, Germany). EEG data was low-pass filtered at 30 Hz and downsampled to 500 Hz. Eye movement and remaining ballistocardiogramm artefacts were removed using ICA following a conservative approach to minimize signal loss where typically less than 8 out of 59 ICs were discarded. Subsequently the data was re-referenced to average reference and remaining artefacts were marked for exclusion. The continuous EEG was segmented into stimulus-locked ERP epochs of 1250 ms starting 250 ms before the cue and baseline corrected. For the conventional ERP analysis, all artefact free correct trials were averaged per task condition (NoGo, neutral, incongruent, congruent) for each participant and across participants. Paired t-tests were calculated for NoGo vs. neutral, as well as incongruent vs. congruent conditions as implemented in the Brain Vision Analyzer 2, and mean t-values were reported for significant epochs. In addition, mean amplitudes during epochs of significant amplitude differences according to these

351

running t-Tests were exported for every participant, and subjected to a repeated measures ANOVA with two factors: ERP component (N2 and P3) and condition (congruent, incongruent, neutral and NoGo) within SPSS. P3 peaks were identified at electrode Pz within a time window of 300–700 ms after stimulus onset, and latencies were exported for further analysis. For the single trial ERP analysis, we calculated two measures for each experimental condition and response side (neutral left, neutral right, NoGo left, NoGo right, congruent left, congruent right, incongruent left and incongruent right). We quantified both the N2 and P3 ERP components at electrode Cz where the NoGo amplitudes were largest in the average, and where the most prominent statistical NoGo effects were expected. The first measure was the mean voltage of the epoch between 250 and 350 ms after the stimulus (N2). This time window was chosen to represent N2 effects in all experimental conditions. The second measure was the mean voltage of the epoch between 400 and 700 ms after the stimulus, including but not limited to the NoGo P3 time window. To test for significant differences within these specific time-frames, mean amplitudes (across all trials) were also extracted for all conditions and entered into a repeated measures ANOVA with two factors: ERP (N2 and P3) and condition (congruent, incongruent, neutral and NoGo). To test for significant differences within these specific time-frames, mean amplitudes across all trials were also extracted for all conditions and entered into a repeated measures ANOVA with two factors: ERP (N2 and P3) and condition (congruent, incongruent, neutral and NoGo). In addition, we obtained the mean amplitude in the N2 (250–350 ms) and NoGo P3 (400–700 ms) time window for frontal, central and parietal regions. For frontal regions, channels Fz, FCz, Cz, FC1 and FC2 were pooled, for central regions channels FCz, Cz, CPz, C1 and C2 and for parietal regions channels comprised Cz, CPz, Pz, CP1 and CP2. FMRI data analysis FMRI data analysis was carried out using SPM8 software (Statistical Parametric Mapping, Welcome Department of Cognitive Neurology, University College London, London UK). Preprocessing included slice time correction, realignment to correct for movement artefacts, spatial normalization into a standard stereotactic space (2 mm3) via an EPI template (Montreal Neurological Institute [MNI] space), and spatial smoothing with a three-dimensional Gaussian filter of 8 mm fullwidth half-maximum using standard SPM8 methods. Low frequency temporal trends were minimized through high-pass filtering with a cutoff of 128 s, and intrinsic autocorrelations were modelled. For the conventional fMRI analysis, we constructed a general linear model, containing eight regressors of interest, and nine regressors of no interest. Regressors of interest were formed using onsets of the four experimental conditions (separately for correctly responded trials with the central arrow pointing right and left), and convolved with the standard hemodynamic response function. Three regressors of no interest contained the onsets of error-trials and fixation crosses during ISI and null events. Six further regressors contained the motion parameters obtained during realignment. First level results were calculated for main effects of the four conditions as well as for contrasts between NoGo and neutral in both directions, and congruent and incongruent in both directions. Second level results were calculated by means of T-tests. For the combined EEG and fMRI analysis across conditions the model contained only three regressors of interest. The first regressor was formed by all correctly responded trials, regardless of condition. The two further regressors consisted of the corresponding parametric modulators of the correctly responded trials using the trial-by-trial ERP measures of N2 and P3. Importantly, in contrast to previous work, ERPs were entered as parametric modulators in one model in the order of their temporal appearance instead of creating separate models for each parametric modulator. The orthogonalisation of successive

352

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

parametric modulators ensures that only unique variance not explained by the N2 is modelled by the subsequent P3 (Buchel et al., 1998; Wood et al., 2008). Note that this approach does not preclude sequential activation of the same brain region by the successive components as long as their fluctuations are at least partly uncorrelated. To simplify interpretation of the negative N2-values, each N2 value was multiplied by –1 prior to submission to the model. Results were calculated for effects of each parametric modulator. Three separate, additional models tested the single-trial ERP measures at central, frontal and parietal regions (instead of Cz). Second level analysis was performed by means of one-sample T-tests. Results for each parametric modulator were thresholded at puncorr. b 0.001 with a cluster-threshold of k = 20 and saved as binary (zero or one) images. The conjunction between the results for ERPs quantified at Cz and frontal, central and parietal regions was calculated by summating these binary images and applying the appropriate intensity threshold. Results Behavioural data Accuracy of responses differed between conditions (F(3,176) = 60.62, p b 0.001). With a mean of 11.56% errors in the NoGo condition, error percentage was significantly higher for NoGo than for all other conditions (p b 0.001 for all comparisons). However, differences between neutral (0.80%), congruent (0.13%) and incongruent (1.02%) conditions were not significant. This pattern held in the EEG subsample (condition: F(3,128) = 50.07, p b 0.001, NoGo: 11.90%, neutral: 0.31%, congruent: 0.17%, incongruent: 0.59%). The RT's also differed between conditions (F(2,132) = 9.03, p b 0.001). Subjects showed significantly faster RT in congruent trials (M = 517.28 ms) compared to incongruent trials (M = 574.02 ms, p b 0.001). Reaction times in the neutral condition (M = 542.74 ms) did not differ significantly from RT in the other conditions. In the EEG subsample, conditions differed as well (F(2,96) = 6.81, p = 0.002), showing the same pattern as in the full sample (congruent: M = 515.83 ms incongruent M = 574.18 ms neutral M = 538.69 ms). EEG results The EEG analysis revealed distinct components for all experimental conditions. When comparing conditions, we found the N2 component to be significantly enhanced in NoGo compared to neutral trials (t (16) = − 3.44, p = 0.003) in a time window of 268 to 296 ms after stimulus onset at electrode Cz (see Fig. 1 A, for regional waveshapes see supplementary material, Fig. 1). There were no significant differences between incongruent and congruent trials in the N2 time window. P3 amplitude was significantly higher for NoGo compared to neutral trials during a time window between 466 and 750 ms after stimulus onset (t(16) = 6.92, p b 0.001) at electrode Cz (see Fig. 1 A and supplementary material Fig. 2 A). There were no significant P3 amplitude differences between incongruent and congruent trials in the P3 time window (waveshapes in supplementary material, Fig. 2 B). The repeated measures ANOVA across the significant N2 and P3 epochs yielded a significant ERP × condition interaction (F(1.95,31.13) = 34.02, p b 0.001). Post hoc analysis indicated larger amplitudes in the N2 time window for NoGo (M = −1.73 μV) compared to all other conditions (neutral: M = −0.35 μV, p = 0.020, congruent: M = − 0.02 μV, p = 0.007, incongruent: M = − 0.05 μV, p = 0.016). The same pattern was evident in the NoGo P3 time window, with higher amplitudes in NoGo (M = 2.25 μV) compared to all other conditions (neutral: M = − 0.14 μV, p b 0.001, congruent: M = 0.07 μV, p b 0.001, incongruent: M = 0.15 μV, p = 0.001). A repeated measures ANOVA carried out for the mean single trial amplitudes selected for the combined analysis revealed a significant ERP × condition interaction (F(1.93,30.89) = 26.32, p b 0.001). Post hoc

analysis identified yielded no significant differences between conditions for N2. However, mean amplitudes were significantly larger for P3 in the NoGo condition (M = 2.14 μV) compared to the neutral (M = 0.09 μV, p b 0.001), congruent (M = 0.36 μV, p b 0.001) and incongruent (M = 0.44 μV, p = 0.008) conditions. A paired t-test revealed significantly longer latencies for P3 peak at Pz in incongruent (494.00 ms) compared to congruent (455.76 ms) trials (t(16)=2.43, p = 0.027). fMRI results For the conventional fMRI analysis in the full sample we investigated contrasts between the NoGo and the neutral condition as well as contrasts between incongruent and congruent trials. We found significant activation in the bilateral insula and left ACC for NoGo compared to neutral stimuli (see Fig. 2). Further activation for this contrast comprised bilateral superior frontal gyrus, supramarginal gyrus, right precuneus and right middle, medial and inferior frontal gyri. The activation in the frontal regions can be assigned to the DLPFC. For complete list of activations see Table 1. There were no significant differences between incongruent and congruent stimuli at a threshold of pFWE b 0.05. Combined EEG–fMRI analysis The parametric modulation of mean N2 amplitude at Cz across conditions revealed significant deactivation of areas comprising the right aMCC (according to the classification of the cingulate cortex suggested by Vogt et al. (1995, 2005)), bilateral occipital and superior temporal gyri (STG), left fusiform gyrus and bilateral lingual gyrus (see Fig. 3 A; for complete list of peak deactivations see Table 2). Conjunction between the N2 modulation at Cz and at all (frontal, central and parietal regions) was found in the right medial frontal gyrus, right precuneus, left fusiform gyrus, right lingual gyrus, as well as occipital areas. Conjunction between N2 modulation at Cz and at central regions comprised the same areas, plus activity in the left cuneus and right aMCC. Conjunction between the N2 at Cz and both parietal and frontal regions yielded clusters in the bilateral precentral gyrus and right superior frontal gyrus, while the conjunction with the parietal region alone further included the STG and MTG. For a complete list of conjunction clusters see Table 4. The parametric modulation of mean NoGo P3 amplitude at Cz across conditions revealed significant activation of the left thalamus, right posterior cingulate cortex (PCC) and a cluster comprising the left anterior insula and IFC (see Fig. 3 B and Table 3). Conjunction between the NoGo P3 modulation at Cz and at frontal, central and parietal regions was present in the left IFG/insula region. Conjunction with central and frontal regions alone yielded additional clusters in the left thalamus and right PCC, while conjunction with NoGo P3 at parietal regions also yielded activation in the PCC, but with a slightly different position. For a complete list of conjunction clusters see Table 4. Discussion The goal of the present study was to investigate cross-modal correlations of well-established response inhibition measures. First, the results confirmed that the paradigm targeted the expected executive functions of inhibitory control. As expected, subjects made significantly more mistakes during the NoGo compared to all other conditions, and were significantly slower on incongruent compared to congruent trials. The ERPs revealed the expected N2 and P3 enhancement during the NoGo condition. The topography of the NoGo P3 was localized less anterior than expected, and did not show a clearly more frontocentral distribution compared to Go trials (Fallgatter et al., 1997). Nevertheless, the difference maps of NoGo-neutral showed a more anterior distribution, suggesting that overlap from a shared activity with Go P3-like

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

353

Fig. 1. ERP results for NoGo and neutral stimuli. A. Stimulus-locked ERP waveshapes to neutral (black) and NoGo (red) correctly responded trials (blue bars p b 0.01). B. Corresponding maps in the time range of N2 and P3 and t-maps for NoGo vs. neutral trials.

topography in both Go and NoGo conditions masked the NoGo P3 topography in this task. The fMRI data showed activation in areas commonly associated with response inhibition, such as the insula, ACC and DLPFC, but also precuneus and IFC (Menon et al., 2001; Swick et al., 2011). ACC activation also approached significance while bilateral insula and IFC survived FWE-correction in the EEG-subsample. Thus, we confirmed that an established fMRI paradigm reliably activated key regions of inhibitory control (Swick et al., 2011), and at

the same time showed landmark ERP effects of inhibitory control and conflict monitoring. The combined analysis of EEG and fMRI data provided intriguing new insights into the specific characteristics of sequential inhibitory processes. Deactivation of a part of the cingulate gyrus which can most likely be assigned to the anterior midcingulate cortex (aMCC) was associated with increased N2 amplitude. This midcingulate region is considered part of the cingulate motor area (CMA, Ullsperger and von Cramon

354

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

Fig. 2. fMRI results of the NoGo vs. neutral contrast at pFWE b 0.05, k = 20.

DMN. In other words, higher N2 amplitude is associated with a greater deactivation of parts of the DMN, which might represent a better allocation of attentional resources to the task. It might also represent some aspects of conflict processing, which requires increased attention. In summary, we found a significantly higher N2 amplitude in NoGo trials as well as an association between an increased N2 amplitude and a deactivation of the rCMA and brain regions thought to be part of the DMN. This led us to the conclusion that an increased N2 amplitude might reflect a higher demand to allocate attentional resources to the task (DMN deactivation) and disruption of a repeated movement (rCMA deactivation). This interpretation is supported by recent findings of Eichele et al. (2008, 2010) who reported that a decrease of the N2 and less deactivation of the DMN are associated with a reduction of effort and an increase in errors. NoGo P3 amplitude variations following the N2 was independently correlated with activation in the left anterior insula and IFG, left thalamus and right posterior cingulate across conditions, meaning that higher P3 amplitude went along with a higher amount of brain activation in those areas. The anterior insula and IFC are areas that are consistently linked to response inhibition (Swick et al., 2011), while the posterior cingulate is usually activated during wakeful rest, but also during internal reflection as part of the DMN (Raichle et al., 2001), self-appraisal and memory retrieval (Ries et al., 2006). Higher P3 amplitude at Cz was found for NoGo compared to all other trials. The

(2003)). Picard and Strick (1996, 2001) have suggested that the dorsal part of the rostral cingulate zone in humans (consistent with the aMCC) is corresponding to a part of the cingulate motor area in primates (CMAv) which is most likely related to attention, selection for action and motor function. They also reported that tasks, which do not clearly dissociate conflict monitoring from response selection, most likely lead to response related activations within this area. Recently, Iwata et al. (2013) were able to show differential reaction of subparts of the rostral part of the CMA (rCMA) related to either repetitive movements or increased activity with a reward-based movement switch in primates, thus suggesting a relation to motor inhibition in our task. While we expected a significant correlation of the N2 amplitude with activation in areas like the ACC, which had previously been identified as a source of N2 (Bekker et al., 2004; Nieuwenhuis et al., 2003), we did not find the ACC to be linked to N2 variability across conditions. Furthermore, increased N2 amplitude was also correlated with decreased activation in the right precuneus, bilateral STG and right medial frontal gyrus. These regions have consistently been reported as part of the default mode network (DMN), which is activated during wakeful rest and self-referential processing, and deactivated during engagement in goal-oriented activity and allocation of attentional resources (Buckner et al., 2008; Raichle et al., 2001). Therefore, we speculate that the association of a deactivation of these brain regions and N2 amplitude fluctuation could represent successful inhibition of parts of the

Table 1 Peak activations for the contrast of NoGo versus neutral trials. Results reported at p b 0.05 FWE-corrected, k = 20. Regions marked † are ROI(FWE b 0.05) significant in the EEG-subsample. Full Sample Region

Hemisphere

Label

T

p(unc)

x

y

z

Frontal

r r l r r l r l l r r r r l l l

Inferior frontal gyrus Inferior frontal gyrus Inferior frontal gyrus Medial frontal gyrus Middle frontal gyrus Middle frontal gyrus Insula Insula Anterior cingulate† Inferior parietal lobule Inferior parietal lobule Superior parietal lobule Supramarginal gyrus Supramarginal gyrus Supramarginal gyrus Lingual gyrus

7.986 7.226 7.721 7.060 7.976 7.649 11.080 7.825 7.077 7.589 6.877 7.636 8.153 7.417 6.995 8.170

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

33.66 29.7 −33.66 9.9 31.68 −27.72 33.66 −33.66 −11.88 54.45 45.54 25.74 63.36 −63.36 −57.42 −25.74

17.02 26.96 23.08 15.95 45.58 49.63 17.53 17.53 24.45 −43.46 −41.33 −61.64 −43.19 −43 −45.21 −75.73

−9.26 −4.71 −4.52 47.1 17.99 21.47 0.97 0.97 22.73 24.28 27.86 49.14 29.8 33.47 28.05 0.43

l r l l l r

Inferior frontal gyrus Insula Insula Inferior parietal lobule Supramarginal gyrus Supramarginal gyrus

9.58 12.84 9.70 8.07 9.09 9.77

0.000 0.000 0.000 0.000 0.000 0.000

−33.66 31.68 −33.66 −57.42 −61.38 63.36

25.02 17.35 17.53 −37.19 −47.06 −43.19

−4.62 −2.55 0.97 33.18 30.00 29.8

Sublobar Limbic Parietal

Occipital EEG subsample Frontal Sublobar Parietal

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

355

Fig. 3. Results for combined EEG–fMRI analysis. A. Deactivation associated with N2 amplitude (puncorr. b 0.001, k = 20). B. Activation associated with P3 amplitude (puncorr. b 0.001, k = 20).

that the activation of the posterior cingulate can be interpreted in terms of internal reflection, which subjects might have needed to remember the instructions for stopping during the NoGo condition. This would also be in line with the involvement of the PCC in memory retrieval (Ries et al., 2006). The thalamic regions found to be activated in this study were the lateral dorsal and lateral posterior nuclei. The thalamus can be closely linked to motor control and motor preparation (Plichta et al., 2013; Sommer, 2003). Therefore, we further speculate that NoGo trials required participants to recollect the instructions about when to withhold their response, and then to inhibit this motor response.

left anterior insula was significantly activated during NoGo trials, and was also shown to be the most reliably activated region during response inhibition in a recent meta-analysis (Swick et al., 2011). The results especially point out that higher NoGo P3 amplitude is linked to higher activation of the left anterior insula/IFC area. This provides further proof of the specific meaning of the central P3 component for inhibition which had previously been concluded in a number of ERP studies (Bruin et al., 2001; Dimoska et al., 2006; Smith et al., 2007). It is also in line with some of the previous source-localization and multimodal studies which found the NoGo P3 to be linked to this region (Bokura et al., 2001; Enriquez-Geppert et al., 2010; Huster et al., 2011). We speculate

Table 2 Peak deactivations for covariation of N2 and brain activity across all conditions. Results presented at p b 0.001 uncorrected, k = 20. Region

Hemisphere

Label

T

p(unc)

x

y

z

Frontal

r r l r r r l r l l r r l l r l l l r l r r r l

Medial frontal gyrus Middle frontal gyrus Subcallosal gyrus Superior frontal gyrus Superior frontal gyrus Cingulate gyrus Fusiform gyrus Middle temporal gyrus Middle temporal gyrus Middle temporal gyrus Superior temporal gyrus Superior temporal gyrus Superior temporal gyrus Postcentral gyrus Precuneus Cuneus Fusiform gyrus Fusiform gyrus Lingual gyrus Lingual gyrus Lingual gyrus Lingual gyrus Middle occipital gyrus Middle occipital gyrus

3.861 4.020 4.271 4.624 4.061 4.399 4.108 4.273 5.154 4.625 4.859 3.776 4.209 4.061 4.314 4.338 4.522 3.854 4.988 4.550 4.548 4.524 5.365 5.206

0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000 0.000

13.86 55.44 −1.98 13.86 5.94 9.9 −47.52 53.46 −35.64 −43.56 57.42 51.48 −63.36 −49.5 13.86 −15.84 −21.78 −27.72 27.72 −11.88 5.94 23.76 25.74 −45.54

50.65 1.93 3.29 58.22 −0.56 17.43 −51.22 −65.51 −78.71 −75.11 −39.77 −36.08 −45.95 −11.08 −73.35 −78.52 −68.32 −50.96 −78.01 −68.24 −87.53 −87.69 −87.28 −74.13

2.99 38.59 −11.94 −1.07 66.35 37.82 −14.26 10.65 18.87 12.97 20.41 16.54 13.35 50.3 47.89 22.35 −6.67 −9.22 −6.19 −5 −2.35 −5.7 2.69 −6.38

Limbic Temporal

Parietal Occipital

356

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

Table 3 Peak activations for covariation of P3 and brain activity across all conditions. Results reported at p b 0.001 uncorrected, k = 20. Region

Hemisphere

Label

T

p(unc)

x

y

z

Frontal

l l r l

Inferior frontal gyrus/insula Precentral gyrus Cingulate gyrus Thalamus

4.609 4.823 5.252 5.447

0.000 0.000 0.000 0.000

−33.66 −41.58 5.94 −13.86

21.77 18 −15.97 −16.43

8.12 10 30.28 −21.09

Limbic Sub-cortical

In summary, increased central NoGo P3 amplitude is connected to increased activation in areas that are associated with response inhibition, internal reflection and memory recollection, linking the NoGo P3 not only to response inhibition as previously described, but also to reflection about the premises for the reaction. The conjunction analysis between ERPs at Cz and over frontal, central and parietal regions seem to justify the use of Cz as a suitable common electrode providing robust signals for the single-trial analysis of both conflict- or inhibition-related ERP components. Our approach captured both spatially specific and unspecific aspects of the electric activity. While for both ERP components most of the major clusters remained significant across all regional averages, some varied with scalp region. Regarding the NoGo P3, the left IFG/insula region remained significant for measures obtained at frontal, central and parietal regions, while the thalamus activation was only linked to parietal and central regions. The PCC remained significant within all 3 regions, however with a slightly shifted cluster for frontal regions, resulting in no overlap between all 3 regions. Overall, the results for central and parietal regions did not differ, but frontal regions yielded slightly different results. This hints at a more widespread topography of NoGo P3 activity underlying the correlation with the IFG/insula region and PCC, while the correlation with the thalamus seems to be more specifically linked to centroparietal activity. Regarding the N2 the results also showed considerable overlap in clusters like the right precuneus, right medial frontal gyrus, right lingual gyrus as well as occipital areas for frontal, central and parietal regions. While clusters in the bilateral precentral gyrus were found for frontal and parietal, but not for central regions, the right aMCC and left cuneus clusters were found only in the conjunction with central and parietal regions. This hints towards a specificity of topographical distribution for these clusters. In conclusion, single-trial measures obtained at a single electrode (in this case Cz) seem to be able to represent both spatially specific as well as more general aspects of these ERPs. While our findings demonstrate how modelling sequential ERP amplitudes at the same electrode can disentangle their spatially distinct fMRI correlates, topographic analysis and fMRI constrained source modelling clearly offers alternative approaches and additional insights. The sequential modelling of the two ERPs in this study also suggested a temporal order of cognitive processes during inhibitory tasks, where we speculate that the allocation of attentional resources takes part within the N2 time window, thereby shortly before retrieving information about task instructions and the actual inhibition within the P3 time window. Separable aspects of motor control seem to play a major part in both time windows. While the presented results provide intriguing new insights into the dynamics of response inhibition, some limitations need to be addressed. One limitation of this study is the small sample size of only 23 healthy subjects, and a slightly smaller sample with sufficient EEG data quality for ERP and combined analyses. While the sample size is comparable or larger than in other simultaneous studies, these findings should nevertheless be replicated within larger samples, to provide higher power for the relatively small effects in the combined analysis. The small sample size potentially also accounts for the null effects regarding the congruency (incongruent vs. congruent) effects in the fMRI and EEG

amplitude. However, previously described latency effects could be replicated even in this small sample. It can also be viewed as a limitation that data quality, especially in the EEG, is diminished by the simultaneous measurement and cannot be entirely “cleaned” of artefacts. However, the current study is the first one to address inhibitory control in a simultaneous EEG–fMRI study while treating the successive N2 and P3 components separately but in one model, focusing on their independent (orthogonalized) sequential modulation. By choosing this type of analysis we were able to show that within an inhibitory task these two components were linked to activation in different brain regions providing further insight into the neural generators of N2 and P3. Therefore we are convinced that the benefits of the simultaneous measurement here outbalance the rather small losses concerning data quality. Our approach to not separate between experimental conditions follows some previous studies (Mulert et al., 2008; Warbrick et al., 2009). This approach does not imply that a given ERP component in general reflects only the same processes across all condition, but that the amplitude variations between conditions capture a modulation of the same process similar to the standard assumption in EEG literature of comparing averaged ERPs across conditions. Therefore, we were interested in fluctuations between task conditions, also incorporating the large variation in N2 amplitude between Go and NoGo trials. However, as a follow up in future studies it would be interesting to further investigate the smaller fluctuations of ERP amplitudes within the different task conditions separately, comparable for example to Debener et al. (2005) and Benar et al. (2007). Additionally, it would be interesting to further investigate the regional specificity of ERP measures obtained from different topographical regions on the scalp through source analysis, and by using alternative asymmetric or symmetric approaches to EEG–fMRI fusion (Huster et al., 2012). This might also help to further disentangle subfunctions of the ERPs. Conclusions The present study is the first to combine EEG and fMRI data from both major conflict- or inhibition-related ERP components (NoGo N2 and NoGo P3). This provides further insight into the neural generators of these ERPs during a combined Flanker and NoGo task, namely that increased N2 amplitude is associated with decreased activation of parts of the default mode and motor networks, while increased P3 amplitude is associated with increased activation in inhibitory networks, but also in areas associated with memory recollection and internal reflection. To our knowledge, this study is also the first to use two parametric modulators of temporally successive ERP activations in one model, thereby accounting for the variance of the first modulator in the results for the second modulator and adding a temporal aspect to the analysis. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.neuroimage.2014.01.023. Acknowledgments This research was supported by the Deutsche Forschungsgemeinschaft (SFB 636). The authors thank Dagmar Gass, Heike Schmidt, Vera Zamoscik, Christine Niemeyer, and Katharina Heubach for their assistance

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

357

Table 4 Peak areas of clusters for conjunction of analysis at Cz and pooled channels at frontal, parietal and central regions. Region

Label

x

y

z

Medial frontal gyrus Precuneus Fusiform gyrus Inferior occipital gyrus Inferior occipital gyrus Lingual gyrus Lingual gyrus Middle occipital gyrus Middle occipital gyrus

15.84 13.86 −23.76 23.76 23.76 23.76 9.9 −45.54 −41.58

58.1273 −75.4799 −66.5493 −93.5914 −91.234 −78.1747 −89.6323 −72.362 −82.9479

−2.9088 44.3078 −10.1266 −7.0912 1.2013 −9.5448 −5.6073 −9.8357 11.5201

r r r l l r r r r l l

Medial frontal gyrus Cingulate gyrus Precuneus Cuneus Fusiform gyrus Inferior occipital gyrus Inferior occipital gyrus Lingual gyrus Lingual gyrus Middle occipital gyrus Middle occipital gyrus

13.86 18.7739 13.86 −17.82 −23.76 23.76 23.76 27.72 9.9 −45.54 −41.58

56.1057 19.2774 −75.4799 −80.6425 −66.5493 −93.5914 −91.234 −78.1747 −89.6323 −72.362 −82.9479

−4.4897 35.8814 44.3078 18.7739 −10.1266 −7.0912 1.2013 −9.5448 −5.6073 −9.8357 11.5201

r r l r l r r r r r l l

Medial frontal gyrus Precentral gyrus Precentral gyrus Superior frontal gyrus Fusiform gyrus Precuneus Inferior occipital gyrus Inferior occipital gyrus Lingual gyrus Lingual gyrus Middle occipital gyrus Declive

15.84 53.46 −29.7 5.94 −47.52 13.86 23.76 23.76 23.76 9.9 −41.58 −25.74

58.1273 −0.098342 −21.8779 −0.65649 −59.0509 −75.4799 −93.5914 −91.234 −78.1747 −89.6323 −83.0399 −64.6957

−2.9088 36.851 67.4177 64.5134 −15.5481 44.3078 −7.0912 1.2013 −9.5448 −5.6073 9.6824 −11.9014

r r r l l r r r l r r r l l r l l r l r

Frontal Medial frontal gyrus Precentral gyrus Precentral gyrus Precentral gyrus Superior frontal gyrus Cingulate gyrus Middle temporal gyrus Superior temporal gyrus Superior temporal gyrus Midbrain Precuneus Cuneus Fusiform gyrus Lingual gyrus Middle occipital gyrus Middle occipital gyrus Middle occipital gyrus Middle occipital gyrus Culmen

1.98 13.86 55.44 −43.56 −29.7 5.94 18.7739 61.38 −61.38 61.38 9.9 13.86 −17.82 −23.76 27.72 −45.54 −25.74 51.48 −41.58 17.82

1.2659 56.1057 −0.1903 −13.2016 −21.8779 −0.65649 19.2774 −27.4619 −47.9796 −41.891 −17.858 −75.4799 −80.6425 −66.5493 −78.1747 −72.362 −94.9412 −67.6312 −82.9479 −41.1929

−13.5202 −4.4897 35.0133 46.7182 67.4177 64.5134 35.8814 −5.3542 11.6125 16.8347 −7.5169 44.3078 18.7739 −10.1266 −9.5448 −9.8357 4.751 7.069 11.5201 −8.0313

P3 Cz + central + frontal + parietal Frontal l

Inferior frontal gyrus/insula

−43.56

23.5268

4.3496

P3 Cz + central Frontal Limbic Subcortical

l r l

Inferior frontal gyrus/insula Cingulate gyrus Thalamus

−39.6 3.96 −15.84

19.6516 −19.9339 −20.4857

4.5435 28.6321 17.6059

P3 Cz + frontal Frontal Limbic

l r

Inferior frontal gyrus/insula Cingulate gyrus

−43.56 5.94

23.5268 −15.9668

4.3496 30.2758

P3 Cz + parietal Frontal Limbic Subcortical

l r l

Inferior frontal gyrus/insula Cingulate gyrus Thalamus

−39.6 3.96 −15.84

19.6516 −19.9339 −20.4857

4.5435 28.6321 17.6059

N2 Cz + central + frontal + Frontal Parietal Occipital

N2 Cz + central Frontal Limbic Parietal Occipital

N2 Cz + frontal Frontal

Temporal Parietal Occipital

Cerebellum N2 Cz + parietal Frontal

Limbic Temporal

Midbrain Parietal Occipital

Cerebellum

Hemisphere parietal r r l r r r r l l

358

S. Baumeister et al. / NeuroImage 94 (2014) 349–359

with data collection as well as Michael Hoppstädter and Christian Bäuchle for the insightful discussions regarding parametric modulation. Conflict of interest Prof. Banaschewski served in an advisory or consultancy role for Hexal Pharma, Lilly, Medice, Novartis, Otsuka, Oxford outcomes, PCM scientific, Shire and Viforpharma. He received conference attendance support and conference support or received speaker’s fee by Lilly, Medice, Novartis and Shire. He is/has been involved in clinical trials conducted by Lilly, Shire & Viforpharma. Prof. Meyer-Lindenberg receives consultant fees and travel expenses from AstraZeneca, Hoffmann-La Roche, Lundbeck Foundation, speaker’s fees from Pfizer Pharma, Lilly Deutschland, Glaxo SmithKline, Janssen Cilag, Bristol-Myers Squibb, Lundbeck and AstraZeneca. Prof. Holtmann served in an advisory or consultancy role for Lilly, Novartis, Shire and Bristol-Myers Squibb, and received conference attendance support or was paid for public speaking by AstraZeneca, Bristol-Myers Squibb, Janssen-Cilag, Lilly, Medice, Neuroconn, Novartis and Shire. Dr. Hohmann received a speaker’s fee from Janssen-Cilag. All other authors declare that they have no biomedical financial interest or potential conflicts of interest. The present work is unrelated to the above grants and relationships.

References Albrecht, B., Brandeis, D., Uebel, H., Valko, L., Heinrich, H., Drechsler, R., Heise, A., Muller, U.C., Steinhausen, H.C., Rothenberger, A., Banaschewski, T., 2012. Familiality of neural preparation and response control in childhood attention deficit-hyperactivity disorder. Psychol. Med. 1–15. Aron, A.R., 2007. The neural basis of inhibition in cognitive control. Neuroscientist 13, 214–228. Aron, A.R., Robbins, T.W., Poldrack, R.A., 2004. Inhibition and the right inferior frontal cortex. Trends Cogn. Sci. 8, 170–177. Bekker, E.M., Kenemans, J.L., Verbaten, M.N., 2004. Electrophysiological correlates of attention, inhibition, sensitivity and bias in a continuous performance task. Clin. Neurophysiol. 115, 2001–2013. Benar, C.G., Schon, D., Grimault, S., Nazarian, B., Burle, B., Roth, M., Badier, J.M., Marquis, P., Liegeois-Chauvel, C., Anton, J.L., 2007. Single-trial analysis of oddball event-related potentials in simultaneous EEG–fMRI. Hum. Brain Mapp. 28, 602–613. Blasi, G., Goldberg, T.E., Weickert, T., Das, S., Kohn, P., Zoltick, B., Bertolino, A., Callicott, J.H., Weinberger, D.R., Mattay, V.S., 2006. Brain regions underlying response inhibition and interference monitoring and suppression. Eur. J. Neurosci. 23, 1658–1664. Bokura, H., Yamaguchi, S., Kobayashi, S., 2001. Electrophysiological correlates for response inhibition in a Go/NoGo task. Clin. Neurophysiol. 112, 2224–2232. Brandeis, D., Banaschewski, T., Baving, L., Georgiewa, P., Blanz, B., Warnke, A., Steinhausen, H.C., Rothenberger, A., Scheuerpflug, P., 2002. Multicenter P300 brain mapping of impaired attention to cues in hyperkinetic children. J. Am. Acad. Child Adolesc. Psychiatry 41, 990–998. Bruin, K.J., Wijers, A.A., van Staveren, A.S., 2001. Response priming in a go/nogo task: do we have to explain the go/nogo N2 effect in terms of response activation instead of inhibition? Clin. Neurophysiol. 112, 1660–1671. Buchel, C., Holmes, A.P., Rees, G., Friston, K.J., 1998. Characterizing stimulus–response functions using nonlinear regressors in parametric fMRI experiments. Neuroimage 8, 140–148. Buckner, R.L., Andrews-Hanna, J.R., Schacter, D.L., 2008. The brain's default network: anatomy, function, and relevance to disease. Ann. N. Y. Acad. Sci. 1124, 1–38. Bunge, S.A., Dudukovic, N.M., Thomason, M.E., Vaidya, C.J., Gabrieli, J.D., 2002. Immature frontal lobe contributions to cognitive control in children: evidence from fMRI. Neuron 33, 301–311. Debener, S., Ullsperger, M., Siegel, M., Fiehler, K., von Cramon, D.Y., Engel, A.K., 2005. Trial-by-trial coupling of concurrent electroencephalogram and functional magnetic resonance imaging identifies the dynamics of performance monitoring. J. Neurosci. 25, 11730–11737. Dhar, M., Been, P.H., Minderaa, R.B., Althaus, M., 2010. Information processing differences and similarities in adults with dyslexia and adults with Attention Deficit Hyperactivity Disorder during a Continuous Performance Test: a study of cortical potentials. Neuropsychologia 48, 3045–3056. Dimoska, A., Johnstone, S.J., Barry, R.J., 2006. The auditory-evoked N2 and P3 components in the stop-signal task: indices of inhibition, response-conflict or error-detection? Brain Cogn. 62, 98–112. Doehnert, M., Brandeis, D., Imhof, K., Drechsler, R., Steinhausen, H.C., 2010. Mapping attention-deficit/hyperactivity disorder from childhood to adolescence—no neurophysiologic evidence for a developmental lag of attention but some for inhibition. Biol. Psychiatry 67, 608–616.

Donkers, F.C., van Boxtel, G.J., 2004. The N2 in go/no-go tasks reflects conflict monitoring not response inhibition. Brain Cogn. 56, 165–176. Egner, T., Hirsch, J., 2005. Cognitive control mechanisms resolve conflict through cortical amplification of task-relevant information. Nat. Neurosci. 8, 1784–1790. Eichele, T., Specht, K., Moosmann, M., Jongsma, M.L., Quiroga, R.Q., Nordby, H., Hugdahl, K., 2005. Assessing the spatiotemporal evolution of neuronal activation with singletrial event-related potentials and functional MRI. Proc. Natl. Acad. Sci. U. S. A. 102, 17798–17803. Eichele, T., Debener, S., Calhoun, V.D., Specht, K., Engel, A.K., Hugdahl, K., von Cramon, D.Y., Ullsperger, M., 2008. Prediction of human errors by maladaptive changes in eventrelated brain networks. Proc. Natl. Acad. Sci. U. S. A. 105, 6173–6178. Eichele, H., Juvodden, H.T., Ullsperger, M., Eichele, T., 2010. Mal-adaptation of eventrelated EEG responses preceding performance errors. Front. Hum. Neurosci. 4. Enriquez-Geppert, S., Konrad, C., Pantev, C., Huster, R.J., 2010. Conflict and inhibition differentially affect the N200/P300 complex in a combined go/nogo and stop-signal task. Neuroimage 51, 877–887. Eriksen, B.A., Eriksen, C.W., 1974. Effects of noise letters upon identification of a target letter in a nonsearch task. Percept. Psychophys. 16, 143–149. Falkenstein, M., Hoormann, J., Hohnsbein, J., 1999. ERP components in Go/Nogo tasks and their relation to inhibition. Acta Psychol. (Amst) 101, 267–291. Fallgatter, A.J., Brandeis, D., Strik, W.K., 1997. A robust assessment of the NoGoanteriorisation of P300 microstates in a cued Continuous Performance Test. Brain Topogr. 9, 295–302. Fallgatter, A.J., Ehlis, A.C., Seifert, J., Strik, W.K., Scheuerpflug, P., Zillessen, K.E., Herrmann, M.J., Warnke, A., 2004. Altered response control and anterior cingulate function in attention-deficit/hyperactivity disorder boys. Clin. Neurophysiol. 115, 973–981. Fallgatter, A.J., Ehlis, A.C., Rosler, M., Strik, W.K., Blocher, D., Herrmann, M.J., 2005. Diminished prefrontal brain function in adults with psychopathology in childhood related to attention deficit hyperactivity disorder. Psychiatry Res. 138, 157–169. Folstein, J.R., Van Petten, C., 2008. Influence of cognitive control and mismatch on the N2 component of the ERP: a review. Psychophysiology 45, 152–170. Gehring, W.J., Gratton, G., Coles, M.G., Donchin, E., 1992. Probability effects on stimulus evaluation and response processes. J. Exp. Psychol. Hum. Percept. Perform. 18, 198–216. Halder, P., Brem, S., Bucher, K., Boujraf, S., Summers, P., Dietrich, T., Kollias, S., Martin, E., Brandeis, D., 2007. Electrophysiological and hemodynamic evidence for late maturation of hand power grip and force control under visual feedback. Hum. Brain Mapp. 28, 69–84. Huster, R.J., Eichele, T., Enriquez-Geppert, S., Wollbrink, A., Kugel, H., Konrad, C., Pantev, C., 2011. Multimodal imaging of functional networks and event-related potentials in performance monitoring. Neuroimage 56, 1588–1597. Huster, R.J., Debener, S., Eichele, T., Herrmann, C.S., 2012. Methods for simultaneous EEG–fMRI: an introductory review. J. Neurosci. 32, 6053–6060. Huster, R.J., Enriquez-Geppert, S., Lavallee, C.F., Falkenstein, M., Herrmann, C.S., 2013. Electroencephalography of response inhibition tasks: functional networks and cognitive contributions. Int. J. Psychophysiol. 87, 217–233. Iwata, J.I., Shima, K., Tanji, J., Mushiake, H., 2013. Neurons in the cingulate motor area signal context-based and outcome-based volitional selection of action. Exp Brain Res. 229, 138–149. Klein, P.A., Petitjean, C., Olivier, E., Duque, J., 2013. Top-down suppression of incompatible motor activations during response selection under conflict. Neuroimage 86, 138–149. Kopp, B., Mattler, U., Goertz, R., Rist, F., 1996. N2, P3 and the lateralized readiness potential in a nogo task involving selective response priming. Electroencephalogr. Clin. Neurophysiol. 99, 19–27. Menon, V., Adleman, N.E., White, C.D., Glover, G.H., Reiss, A.L., 2001. Error-related brain activation during a Go/NoGo response inhibition task. Hum. Brain Mapp. 12, 131–143. Meyer-Lindenberg, A., Buckholtz, J.W., Kolachana, B., Hariri, A.R., Pezawas, L., Blasi, G., Wabnitz, A., Honea, R., Verchinski, B., Callicott, J.H., Egan, M., Mattay, V., Weinberger, D.R., 2006. Neural mechanisms of genetic risk for impulsivity and violence in humans. Proc Natl Acad Sci U S A 103, 6269–6274. Mulert, C., Seifert, C., Leicht, G., Kirsch, V., Ertl, M., Karch, S., Moosmann, M., Lutz, J., Moller, H.J., Hegerl, U., Pogarell, O., Jager, L., 2008. Single-trial coupling of EEG and fMRI reveals the involvement of early anterior cingulate cortex activation in effortful decision making. Neuroimage 42, 158–168. Mullinger, K.J., Yan, W.X., Bowtell, R., 2011. Reducing the gradient artefact in simultaneous EEG-fMRI by adjusting the subject's axial position. Neuroimage 54, 1942–1950. Nee, D.E., Wager, T.D., Jonides, J., 2007. Interference resolution: insights from a metaanalysis of neuroimaging tasks. Cogn. Affect Behav. Neurosci. 7, 1–17. Nieuwenhuis, S., Yeung, N., van den Wildenberg, W., Ridderinkhof, K.R., 2003. Electrophysiological correlates of anterior cingulate function in a go/no-go task: effects of response conflict and trial type frequency. Cogn. Affect Behav. Neurosci. 3, 17–26. Pascual-Marqui, R., Sekihara, K., Brandeis, D., Michel, C.M., 2009. Imaging the electric neuronal generators of EEG/MEG. In: Michel, C.M., Koenig, T., Brandeis, D., Gianotti, L.R.R., Wackermann, J. (Eds.), Electrical Neuroimaging. Cambridge University Press, Cambridge. Pfefferbaum, A., Ford, J.M., Weller, B.J., Kopell, B.S., 1985. ERPs to response production and inhibition. Electroencephalogr. Clin. Neurophysiol. 60, 423–434. Picard, N., Strick, P.L., 1996. Motor areas of the medial wall: a review of their location and functional activation. Cereb. Cortex 6, 342–353. Picard, N., Strick, P.L., 2001. Imaging the premotor areas. Curr. Opin. Neurobiol. 11, 663–672. Plichta, M.M., Wolf, I., Hohmann, S., Baumeister, S., Boecker, R., Schwarz, A.J., Zangl, M., Mier, D., Diener, C., Meyer, P., Holz, N., Ruf, M., Gerchen, M.F., Bernal-Casas, D.,

S. Baumeister et al. / NeuroImage 94 (2014) 349–359 Kolev, V., Yordanova, J., Flor, H., Laucht, M., Banaschewski, T., Kirsch, P., MeyerLindenberg, A., Brandeis, D., 2013. Simultaneous EEG and fMRI reveals a causally connected subcortical–cortical network during reward anticipation. J. Neurosci. 33, 14526–14533. Raichle, M.E., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L., 2001. A default mode of brain function. Proc. Natl. Acad. Sci. U. S. A. 98, 676–682. Randall, W.M., Smith, J.L., 2011. Conflict and inhibition in the cued-Go/NoGo task. Clin. Neurophysiol. 122, 2400–2407. Ridderinkhof, K.R., van der Molen, M.W., 1995. When global information and local information collide: a brain potential analysis of the locus of interference effects. Biol. Psychol. 41, 29–53. Ries, M.L., Schmitz, T.W., Kawahara, T.N., Torgerson, B.M., Trivedi, M.A., Johnson, S.C., 2006. Task-dependent posterior cingulate activation in mild cognitive impairment. Neuroimage 29, 485–492. Smith, J.L., Johnstone, S.J., Barry, R.J., 2006. Effects of pre-stimulus processing on subsequent events in a warned Go/NoGo paradigm: response preparation, execution and inhibition. Int. J. Psychophysiol. 61, 121–133. Smith, J.L., Johnstone, S.J., Barry, R.J., 2007. Response priming in the Go/NoGo task: the N2 reflects neither inhibition nor conflict. Clin. Neurophysiol. 118, 343–355. Sommer, M.A., 2003. The role of the thalamus in motor control. Curr. Opin. Neurobiol. 13, 663–670. Strik, W.K., Fallgatter, A.J., Brandeis, D., Pascual-Marqui, R.D., 1998. Three-dimensional tomography of event-related potentials during response inhibition: evidence for phasic frontal lobe activation. Electroencephalogr. Clin. Neurophysiol. 108, 406–413.

359

Swick, D., Ashley, V., Turken, U., 2011. Are the neural correlates of stopping and not going identical? Quantitative meta-analysis of two response inhibition tasks. Neuroimage 56, 1655–1665. Ullsperger, M., von Cramon, D.Y., 2003. Error monitoring using external feedback: specific roles of the habenular complex, the reward system, and the cingulate motor area revealed by functional magnetic resonance imaging. J. Neurosci. 23, 4308–4314. Valko, L., Doehnert, M., Müller, U.C., Schneider, G., Albrecht, B., Drechsler, R., Maechler, M., Steinhausen, H.C., Brandeis, D., 2009. Differences in neurophysiological markers of inhibitory and temporal processing deficits in children and adults with ADHD. J. Psychophysiol. 23, 235–246. van Veen, V., Carter, C.S., 2002. The anterior cingulate as a conflict monitor: fMRI and ERP studies. Physiol. Behav. 77, 477–482. Vitacco, D., Brandeis, D., Pascual-Marqui, R., Martin, E., 2002. Correspondence of eventrelated potential tomography and functional magnetic resonance imaging during language processing. Hum. Brain Mapp. 17, 4–12. Vogt, B.A., Nimchinsky, E.A., Vogt, L.J., Hof, P.R., 1995. Human cingulate cortex: surface features, flat maps, and cytoarchitecture. J. Comp. Neurol. 359, 490–506. Vogt, B.A., Vogt, L., Farber, N.B., Bush, G., 2005. Architecture and neurocytology of monkey cingulate gyrus. J. Comp. Neurol. 485, 218–239. Warbrick, T., Mobascher, A., Brinkmeyer, J., Musso, F., Richter, N., Stoecker, T., Fink, G.R., Shah, N.J., Winterer, G., 2009. Single-trial P3 amplitude and latency informed eventrelated fMRI models yield different BOLD response patterns to a target detection task. Neuroimage 47, 1532–1544. Wood, G., Nuerk, H.C., Sturm, D., Willmes, K., 2008. Using parametric regressors to disentangle properties of multi-feature processes. Behav. Brain Funct. 4, 38.