Performance monitoring and behavioral adaptation during task switching: An fMRI study

Neuroscience 285 (2015) 227–235

PERFORMANCE MONITORING AND BEHAVIORAL ADAPTATION DURING TASK SWITCHING: AN FMRI STUDY J. VON DER GABLENTZ, a* C. TEMPELMANN, b T. F. MU¨NTE a AND M. HELDMANN a

Accordingly, the ability to shift flexibly between different task sets is an essential prerequisite. The term task set refers to the configuration of mental resources comprising the representation of task-relevant stimuli, task-relevant responses, and the corresponding stimulus–response mapping (Kiesel et al., 2010). At the behavioral level, a switch between different task sets results in increased reaction times and/or error rates, an effect that is known as switch costs (Jersild, 1927). These ‘‘switch-costs’’ are evident from the comparison of switch and task repeating trials (Monsell, 2003; Kiesel et al., 2010). It is assumed, that switch-costs are caused by proactive interference from previous tasks (Yeung et al., 2006) and by task-set reconfiguration processes (Monsell, 2003). Task-set reconfiguration processes are time consuming preparation processes like the backward inhibition of the previous task set, the overcoming of the now relevant task set’s inhibition (Mayr and Keele, 2000), attention shifting between stimulus attributes, and the encoding or deleting of stimulus–response associations in working memory (Monsell, 2003), that are necessary to enable appropriate behavioral adaption to a task set switch. At the neural level, imaging but also lesion studies in monkeys (Rushworth et al., 2003; Kovach et al., 2012) identified a fronto-parietal network to be relevant for task switching related processes. This network includes the dorsal ACC (anterior cingulate cortex) for conflict monitoring (Mars et al., 2005; Hyafil et al., 2009; Ide et al., 2013), the superior parietal lobule for attentional control (Braver et al., 2003), the lateral prefrontal cortex and intraparietal sulcus for implementation of task goals (Brass and von Cramon, 2004; Hyafil et al., 2009) the pre-SMA (supplementary motor area), inferior parietal lobule and middle temporal gyrus for task-set preparation (De Baene and Brass, 2013) and the pre-SMA and basal ganglia for the inhibition of previous task sets (Whitmer and Banich, 2012). Furthermore, the anterior insular cortex has been suggested to be involved in error awareness (Bush et al., 2000; Klein et al., 2007; Ullsperger et al., 2007) and is assumed to be active in case a task switch fails. There is increasing evidence for an involvement of the noradrenergic system in task switching behavior by regulating arousal and cognitive flexibility (Lapish et al., 2007; Jocham and Ullsperger, 2009). The main source of noradrenaline is the locus coeruleus (LC), a group of neurons located in the brainstem, which projects to the prefrontal cortex, where the modulating function of the noradrenergic system on behavioral flexibility and attentional shifting might manifest (Devauges and Sara,

a Department of Neurology, University of Lu¨beck, Ratzeburger Allee 160, D-23538 Lu¨beck, Germany b

Department of Neurology, Otto-von-Guericke University, Leipziger Strasse 44, D-39120 Magdeburg, Germany

Abstract—Despite significant advances, the neural correlates and neurochemical mechanisms involved in performance monitoring and behavioral adaptation are still a matter for debate. Here, we used a modified Eriksen– Flanker task in a magnetic resonance imaging (MRI) study that required the participants to derive the correct stimulus–response association based on a feedback given after each flanker stimulus. Participants had to continuously monitor and adapt their performance as the stimulus– response association switched after a jittered time interval without notice. After every switch an increase of reaction times was observed. At the neural level, the feedback indicating the need to switch was associated with activation of the precuneus, the cingulate cortex, the insula and a brainstem region tentatively identified as the locus coeruleus. This brainstem system appears to interact with this cortical network and seems to be essential for performance monitoring and behavioral adaptation. In contrast, the cerebellum crus and prefrontal areas are activated during error feedback processing. Furthermore we found activations of the hippocampus and parahippocampal gyrus bilaterally after a correct feedback in learnable stimulus–response associations. These results highlight the contribution of brainstem nuclei to performance adaptation. Ó 2014 IBRO. Published by Elsevier Ltd. All rights reserved.

Key words: locus coeruleus, ACC, hippocampus, task switching, performance monitoring, fMRI.

INTRODUCTION In order to achieve internal goals most effectively humans are required to adapt their behavior continuously to changing environmental demands (Allport et al., 1994). *Corresponding author. Tel: +49-451-31-79-31-313; fax: +49-451500-54-57. E-mail address: [email protected] (J. von der Gablentz). Abbreviations: ACC, anterior cingulate cortex; ANOVA, analysis of variance; EPI, echo planar imaging; IR-EPI, inversion recovery echo planar imaging; LC, locus coeruleus; MRI, magnetic resonance imaging; SMA, supplementary motor area. http://dx.doi.org/10.1016/j.neuroscience.2014.11.024 0306-4522/Ó 2014 IBRO. Published by Elsevier Ltd. All rights reserved. 227

228

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235

1990; Aston-Jones and Cohen, 2005; Bouret and Sara, 2005; Yu and Dayan, 2005; McGaughy et al., 2008; Sara and Bouret, 2012). Some studies already investigated the response of noradrenergic neurons in typical situations (Aston-Jones and Cohen, 2005; Sara, 2009; Sara and Bouret, 2012): During quiet wakefulness, LC neurons are in a tonic mode, firing at a regular slow rate. With the appearance of a behaviorally significant stimulus they shift to a phasic mode, firing short-latency bursts. This phasic mode is associated with focused attention to the stimulus and optimization of behavioral performance, like switching the stimulus–response association (Aston-Jones and Cohen, 2005). Whenever the behavioral relevance of a task wanes, LC neurons fall back to the slow firing rate of the tonic mode, which is at a cognitive level associated with a disengagement from the task. An increased availability of prefrontal noradrenaline induced by drug treatment (Arnsten, 2006a,b; Devilbiss and Berridge, 2008; Lin et al., 2009) or via the firing of the LC is reported to increase cognitive flexibility (Usher et al., 1999; Allen et al., 2005; Aston-Jones and Cohen, 2005; Bouret and Sara, 2005; Cain et al., 2011). Indeed, lesion studies in animals (McGaughy et al., 2008; Tanaka et al., 2009) as well as drug treatments (Lapiz and Morilak, 2006; Cain et al., 2011; Brown et al., 2012) revealed a positive relationship between noradrenaline availability and improved set shifting. In humans, a similar impact of drug treatments affecting the noradrenaline system on task switching behavior was described (Renner and Beversdorf, 2010; Demanet et al., 2011; Chamberlain and Robbins, 2013). The drug Modafinil elevates the synaptic noradrenaline and dopamine level and enhances the phasic response of LC to task-relevant events while tonic LC activity is decreased (Hou et al., 2005; Minzenberg et al., 2008) leading to an increase of prefrontal cortex activity and an improved task performance (Minzenberg et al., 2008). Moreover, manipulating the noradrenaline system has been shown to increase performance accuracy. For instance, Riba et al. (2005) reported that stimulating noradrenergic transmission via the application of the alpha-adrenoceptor antagonist yohimbine results in an improvement of performance accuracy in an Eriksen–Flanker task. The present study aimed to investigate the interaction of the neural systems involved in task switching and performance monitoring. Since action monitoring is known to rely primarily on projections of the mesolimbic dopaminergic system, we expected brain sites belonging to the dopaminergic (action monitoring) as well as the noradrenergic system (task switching) to be active. Participants had to perform a modified Eriksen flanker task, in which the valid stimulus–response association changed without informing the participant while functional magnetic resonance imaging (MRI) was recorded. Participants were able to detect a task switch via a symbolic feedback that was given after each response, informing the participant if the given response was correct regarding the valid task set. In case a task set switch took place, participants received the feedback of having responded incorrectly, although their response was correct in terms of the previously valid task set.

Accordingly, subjects had to be sure regarding their performance quality in order to detect a task set switch. We expected activation of the LC in response to the switch feedback in light of the literature indicating a role of the noradrenaline system in task switching. Furthermore, we expected to see an activation of the performance monitoring network including the anterior cingulate gyrus.

EXPERIMENTAL PROCEDURES The experimental procedures had been approved by the local ethics committee prior to the experiment. All experiments were carried out according to the declaration of Helsinki. Participants After obtaining informed written consent, sixteen healthy volunteers (nine men, seven women) participated in this study. All participants were right-handed and were between 20 to 26 years old (mean age: 23.06 years). All participants were paid for their participation (7 Euros/ hour). Experimental procedure For the Eriksen–Flanker task (Eriksen and Erkisen, 1974) five letter-strings consisting of ‘‘H’’ and ‘‘S’’ were used as congruent (HHHHH, SSSSS) and incongruent stimuli (HHSHH, SSHSS). These stimuli were presented in a random order with incongruent stimuli in 60% of all trials to increase the task’s difficulty. Participants were instructed to respond to the central letter by button-press with either the index or the middle finger. In contrast to other studies, participants were not informed about a fixed stimulus–response association. Instead, they had to figure out the currently valid stimulus–response association via a feedback stimulus, which was presented after each flanker stimulus. This feedback stimulus consisted of a colored square, which was green in case the response was correct and red after an incorrect button press. After a jittered interval (every 6th to 11th trial) the stimulus– response combination was switched (switch trial). The subjects were not explicitly informed about this switch but had to derive the new task set from the feedback. Thus, an index finger response to an H might have been appropriate for the trial preceding the switch, resulted in a red square feedback after the switch. Obviously, subjects were only able to interpret a red square as a signal to switch the task set, when they were sure to have responded correctly. In the following we name the first red square feedback after a switch ‘‘switch feedback’’. In contrast, a red square feedback after an incorrect button press is called ‘‘error feedback’’. Participants were instructed to respond as fast as possible. If the response time exceeded a deadline of 1 s a feedback comprising a gray square was given (Fig. 1). We choose the Flanker task because it is an easy task with only two stimulus– response associations which produces a fair amount of performance errors, which allows us to show the difference between performance and switch errors.

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235

Fig. 1. Paradigm and switch instruction. Representative examples of flanker task over time with switch instructions. A participant responds to the central ‘‘H’’ with an index finger button press and receives a red square feedback. This indicates the need for a switch of the response strategy, so the next time the subject responses with the middle finger to an ‘‘H’’ and receives a correct feedback.

Before entering the scanner, subjects were trained with 20 trials. The experiment proper consisted of three blocks of 155 trials each. Flanker stimuli were presented in white against a black background on a screen with a duration of 100 ms. The feedback square was displayed either in green, red or gray for 200 ms. A fast eventrelated design was used, jittering the presentation of all visual events between 1 and 3 repetition times. Image acquisition Data were collected using a neuro-optimized 1.5-T GE Signa Horizon LX scanner with a standard quadrature head coil. Functional images consisted of 23 axial slices, with 64  64 matrix, 200-mm  200-mm field of view, 5-mm thickness, 1-mm slice gap, and 3.125mm  3.125-mm in-plane resolution. They were obtained using a T2*-weighted echo planar imaging (EPI) sequence, with 2000-ms time repetition, 35-ms time echo, and 80° flip angle. Each structural image consisted of 60 contiguous slices, with 256  256 matrix, 200cm  200-cm field of view, and 1.5-mm thickness. Structural images were obtained using a T1-weighted 3D fast spoiled gradient echo sequence. For a more precise alignment of the functional to the anatomical data set we additionally acquired an inversion recovery echo planar imaging (IR-EPI) scan with the same geometry.

229

Statistical analyses were realized using two steps, taking into account first the intra-individual then inter-individual variance, respectively. The first-level analysis was conducted on each participant using a general linear model. Correct responses of the participants were analyzed for the switch trial (switch), the one after switching (switch n + 1), the next three trials were summarized (switch n + (2–4)) to omit too much testing and the trial directly before switching (switch n 1) (see Fig. 2). The 5th to 10th trial exist, because of unequal numbers of trial per session, only in some of the cases. Therefore, we decided to summarize them in one regressor (switch n + (5–10)). In the switch trial a response was considered as being correct if the subject responded according to the established task set. Due to the covered switching of the task set, the correct response resulted nevertheless in a red square feedback. In order to make sure that participants were able to differentiate this switch feedback from the error feedback, only switch trials were entered into the ‘‘switch regressor’’ when all responses between the current and the previous switch were correct. In contrast, the ‘‘error regressor’’ includes all cases of error feedback. Additionally we included a ‘‘regressor of no interest’’ to account for eventrelated signal changes that were not included in the inferential testing (e.g. too-slow-response feedback). We then defined the following contrasts to test: (a) each of the three ‘‘feedback regressors’’ (switch n + (2– 4), n + (5–10), n 1) against the switch regressor (switch); (b) the error regressor against the switch regressor; (c) the correct feedback regressor switch n + (5–10) against the error regressor. After model estimation, the obtained first-level contrast images from each subject were entered into a secondlevel analysis using a one-sample t-test at each voxel. Resulting maps were corrected for multiple comparisons with the false discovery rate and considered at P < 0.05 with a cluster size of five voxels. Behavioral data analysis

Data analysis

Statistical analyses of behavioral data were performed using Aabel (Gigawiz Ltd. Co., Tulsa, OK, USA). The reaction times of correct responses were calculated for all feedback trials using a single-factor repeated measures analysis of variance (ANOVA). To assess differences between trials we used Tukey’s Honestly

Preprocessing and statistical analyses of the data were conducted on a voxel-by-voxel basis using SPM8 (http:// www.fil.ion.ucl.ac.uk/spm). The first three volumes were discarded because of equilibration effects. Functional images were first slice time corrected and motion corrected. To standardize the preprocessed EPIs we first coregistered them to the IR-EPI, coregistered in a second step the individual T1-image to the IR-EPI, segmented the T1 image, and finally applied the transformation matrix from the segmentation step to the EPIs. Standardized images were then resampled to 2  2  2-mm3 voxels and spatially smoothed using a Gaussian kernel with full width at half maximum of 8 mm. Data were also temporal filtered with a high-pass filter of 128 s.

Fig. 2. Stimulus and feedback trials. The screen shows the stimulus trials over time, whereas switch means the presentation of the stimulus when the feedback changes and switch n + 1 the stimulus presented after the switch. The stimulus presentations of the second to fourth after switching are summarized, as well as the fifth to eleventh afterward. Response of the participant and feedback are depicted with a smile for a suggested respectively an obtained correct feedback and a red cross for false ones.

230

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235

Significant Difference post hoc Test. All results in the following text are given as mean ± standard error of the mean (SEM).

RESULTS Behavioral results Fig. 3 illustrates that switch n + 1 trials were associated with the slowest reactions whereas switch and switch n 1 were fastest. The ANOVA revealed a trial main effect (F(4,60)= 6.20, p < 0.001). Post-hoc tests showed significant results for the comparison of switch n 1 versus switch n + 1 and for switch n + 1 versus switch. The comparison of the percentage errors per condition revealed no errors in the switch n 1 condition. Accordingly, the statistical comparison just comprised the conditions switch n + 1, switch n + (2–4) and switch n + (5–10). Neither the analysis of the percentage errors nor the reaction time data revealed a significant main effect (percentage errors: F(2,30) = 0.64, p > 0.5; reaction time error responses F(2,30) = 2.03, p = 0.148; for mean reaction time and error values see Fig. 3). FMRI results To identify brain areas involved in the processing of the switch signal the following contrast was analyzed: switch feedback versus correct feedbacks switch n + (2–4), switch n + (5–10) and switch n 1. A complete list of activations can be found in Table 1. Considering the activations of switch versus correct trials significant clusters were mainly found in the

precuneus, the insula, the ACC and the brainstem (Fig. 4A). The brainstem activations most likely correspond to the LC. Previous studies identified the location of the LC as anterolateral to the floor of the fourth ventricle just behind the periaqueductal gray in the upper pontine tegmentum (Aston-Jones and Cohen, 2005; Shibata et al., 2006; Keren et al., 2009; Astafiev et al., 2010). This description matches with the activation we found within the brainstem. Furthermore the MNI coordinates fit very well to those found in previous functional MRI-studies (coordinates from x = 2 to 12 or x = 1 to 10, y = 14 to 40 and z = 9 to 24) (Liddell et al., 2005; Sterpenich et al., 2006; Raizada and Poldrack, 2007; Vandewalle et al., 2007; Berman et al., 2008; Minzenberg et al., 2008; Schmidt et al., 2009, 2010; Astafiev et al., 2010; Steuwe et al., 2014). Significant activation clusters for the comparison of error feedback versus switch feedback were observed in the Crus 1 and 2 of the Cerebellum as well as in the superior and middle frontal gyrus (see Fig. 4C and Table 2). Another contrast was performed for the feedback after correct (switch n + (5–10)) versus error trials (error), when subjects received the correct or error feedback. We found a prominent activation of the hippocampal and parahippocampal region bilaterally (see Fig. 4B). The list of activations related to this contrast can be found in Table 3.

DISCUSSION In the present study, we assessed the behavior and the brain activity associated with task switching in a modified Eriksen–Flanker task. Behaviorally, switch costs manifested themselves in an increase of reaction time to trials following the implicit switch signal, i.e. the switch feedback. The switch signal gave rise to activations in the brainstem (most likely the LC), the precuneus, the insula and the cingulate cortex. Activation clusters related to error feedback versus switch feedback were found in the cerebellum and the prefrontal cortex. Furthermore, we found a distinct activation in hippocampus bilaterally when comparing correct feedback with error feedback. Behavioral results The results of the reaction time analysis indicate a profound reaction time increase in the first trial following the switch signal. It is assumed, that switch-costs are caused by task-set reconfiguration processes which include attention shifting between stimulus attributes or the encoding or deleting of stimulus–response associations within working memory (Monsell, 2003). The behavioral data indicate that the experimental design provides a valid basis for the investigation of neural processes related to task switching processes relying on internal action monitoring. Brain activity during task switching

Fig. 3. Reaction times for all trials. Mean reaction time is shown for all correct trials (A) and error trials (B). The percentage of errors per condition is shown in (C).

In switch versus correct trials we found an activation of the insula, the cingulate cortex, the precuneus and a

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235 Table 1. List of brain regions showing significant changes in BOLD response for the contrast of switch feedback (switch) versus correct feedback. (a) switch versus switch n + (2–4) (p < 0.05, FDR corrected for multiple comparisons, cluster threshold five voxel) (b) switch versus switch n + (5–10) (p < 0.05, FDR corrected for multiple comparisons, cluster threshold five voxel) (c) switch versus switch n 1 (p < 0.01, FDR corrected for multiple comparisons, cluster threshold five voxel) x

y

z

Size

t-Value

Region

H

10 36 14 2 30 16 36 22 12 22 12

9369 3944 1221 75 92 46 25 188 30 26 24

10.80 6.98 6.29 5.38 4.47 4.43 4.40 4.32 3.68 3.59 3.51

Insula Precuneus Brainstem Temporal Mid Frontal Mid Frontal Sup Cerebellum Cingulate Post Cingulum Ant Vermis Frontal Mid

L R R R R R L L R R R

22 28 14 24 64 2 26 24 68 40 14 14 34 50 34 40 52 52 10 10 46 36

6 4 8 40 38 48 10 4 34 12 34 22 14 16 26 16 8 16 16 24 32 22

220

Insula Insula Supp Motor Area Cingulate Mid Precuneus Frontal Mid Frontal Inf Tri Insula Precuneus Frontal Lobe Frontal Mid Frontal Inf Oper Brainstem Frontal Mid Frontal Lobe Frontal Lobe

10 9

11.16 8.92 9.40 7.54 8.36 7.78 7.78 7.30 7.51 6.63 6.57 6.48 6.44 6.35 6.29 4.94 5.99 5.73 5.76 5.63 5.66 5.34

Frontal Sup Frontal Mid Extra-Nuclear Frontal Lobe Parietal Lobe

L L R R R R R R L R L L R L R R L L R R R L

(c) 28 2 36 48 32

24 14 48 38 24

6 50 36 42 0

185 330 63 39 26

9.47 8.77 7.01 6.17 6.14

Insula Supp Motor Area SupraMarginal SupraMarginal Insula

L L L R R

6 10 40 10

22 66 46 30

26 66 38 36

25 10 13 9

6.09 5.81 5.68 5.65

Cingulum Ant Precuneus Parietal Inf Frontal Sup Medial

L R R L

(a) 32 8 10 52 30 28 28 2 14 6 40

18 64 32 50 36 58 60 34 40 48 42

(b) 30 32 12 10 10 26 34 32 10 18 40 46 8 36 26 32 20 26 34 36 38 28

292 118 15 205 85 28 169 26 11 40 55 21

MNI coordinates, size in number of voxels, t statistic values; H, hemisphere; L, left; R, right; Temporal Mid, middle temporal gyrus; Frontal Mid, middle frontal gyrus; Frontal Sup, superior frontal gyrus; Cingulate Post, posterior Cingulate; Cingulum Ant, anterior Cingulum; Supp Motor Area, supplemental motor area; Cingulate Mid, mid-cingulate gyrus; Front Inf Tri, inferior frontal gyrus (pars triangularis); Frontal Inf Oper, inferior frontal gyrus (pars opercularis); SupraMarginal, supramarginal gyrus; Parietal Inf, inferior parietal gyrus; Frontal Sup Medial, medial superior frontal gyrus.

brainstem region tentatively identified as the LC based on previous studies (Liddell et al., 2005; Sterpenich et al., 2006; Raizada and Poldrack, 2007; Vandewalle et al.,

231

2007; Berman et al., 2008; Minzenberg et al., 2008; Schmidt et al., 2009, 2010; Astafiev et al., 2010; Steuwe et al., 2014). These findings thus suggest that the fronto-parietal network which has been suggested to be relevant for task switching and error awareness (Rushworth et al., 2003; Kovach et al., 2012) interacts with the noradrenaline system of the LC. An involvement of the precuneus in post-error processing has been reported by different studies (Badgaiyan and Posner, 1998; Menon et al., 2001; Zhang and Li, 2012). For example, Badgaiyan and Posner (1998) found a post-error activation only when an external feedback switched from correct to incorrect suggesting that activation occurs when feedback signals the need to change the response strategy. Likewise it has been suggested that precuneus is involved in shifting of attention (Nagahama et al., 1999; Astafiev et al., 2003; Barber and Carter, 2005; Wenderoth et al., 2005; Shulman et al., 2010; Huang et al., 2012). Furthermore, precuneus has been proposed to evaluate stimulus–response associations and to bias top–down control to enhance task performance according to this association (Bunge et al., 2002; Barber and Carter, 2005; Pourtois, 2011). Taken together the activation of precuneus found in the switch trial of the current study signals the shift of attention to another stimulus–response association in order to adjust task performance. The anterior insula has been suggested to play a role in error awareness (Klein et al., 2007, 2013; Hester et al., 2009; Orr and Hester, 2012) and its activation in the switch trial which entailed error feedback is consistent with this view. Furthermore, the insula is strongly connected with the ACC which is another region we found to be activated during task switching (Craig, 2011). The cingulate cortex plays a major role in detecting discrepancies between the intended and the actual outcome of an action (Halgren et al., 2002; Holroyd et al., 2004; Debener et al., 2005; Wang et al., 2005; Ullsperger et al., 2007; Orr and Hester, 2012). Consequently, this region signals the need for behavioral adjustment (Holroyd and Coles, 2002; Ridderinkhof et al., 2004; Mars et al., 2005; Klein et al., 2007; Ide et al., 2013). Consistent with these assumptions, in the present investigation the cingulate cortex became active whenever there was a clear discrepancy between the feedback signal, indicating an outcome worse than expected, and internal action monitoring processing, representing the given response as correct action. This means the cingulate cortex was activated whenever behavioral adjustment, namely the switch of the response strategy, was necessary. An activation of the ACC can also be detected during erroneous task performance, whenever a mistake has been made even if the participant is unaware of this error (Hester et al., 2005; Klein et al., 2007). Likewise an adjustment of task performance is required in these situations to avoid mistakes in further task execution. A further important activation was found in the brainstem and was identified to be the LC. According to our results the LC has been suggested to be involved in focusing attention, cognitive flexibility and optimizing task performance (phasic mode) as well as in disengagement from a behavioral irrelevant task (tonic

232

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235

Fig. 4. Activation during task switching. Significant changes in BOLD response for the contrast of switch feedback versus correct feedback (switch n + (5–10)) (A) and for the contrast of correct feedback (switch n + (5–10)) versus error feedback (B) as well as for error versus switch feedback (C) rendered as z-scores on an brain template averaged on T1 images of all subjects. Corrected for multiple comparisons using a false discovery rate at p < 0.05 and a cluster size of five (A), respectively a false discovery rate at p < 0.001 and a cluster size of ten (B and C).

Table 2. List of brain regions showing significant changes in BOLD response for the contrast of error feedback versus switch (p < 0.001, FDR corrected for multiple comparisons) x

y 12 6 22 46 58 26 30 34 40

z 74 34 54 24 36 66 62 52 66

36 34 20 32 14 32 40 14 40

Size

t-Value

Region

H

77 72 86 31 5 5 9 8 10

9.27 8.22 7.85 7.62 7.20 6.91 6.68 6.55 6.53

Cerebellum Crus2 Frontal Sup Medial L Frontal Sup Frontal Mid Temporal Mid Cerebellum Crus1 Cerebellum Crus1 Frontal Mid Cerebellum Crus2

R

Error processing

L L L L R R R

To show the difference between negative feedback processing and switch feedback processing we computed the contrast error feedback versus switch feedback, which revealed activation of prefrontal cortex areas and of the Cerebellar Crus 1 and 2. This frontal-cerebellar activation pattern confirms previous investigations reporting this link to be important in the regulation of error processing (Schweizer et al., 2007; Tanaka et al., 2009) and cognitive control (Balsters et al., 2013). Based on the difference to the task switching contrast, in the present report, feedback based error processing seems to involve a motor component (Schweizer et al., 2007) and the reconfiguration of cerebellar related stimulus–response mapping (Balsters et al., 2013). Since the functional link between the cerebellum and prefrontal cortex was shown by resting state and task-related fMRI studies (Allen et al., 2005; Demirci et al., 2009; Lin et al., 2009; Ide and Li, 2011), we would like to argue that the non-cortical brain activations are an essential aspect in the functional differentiation of switch and performance feedback.

MNI coordinates, size in number of voxels, t statistic values; cluster threshold five voxel. H, hemisphere; L, left; R, right; Frontal Sup Medial, medial superior frontal gyrus; Frontal Sup, superior frontal gyrus; Frontal Mid, middle frontal gyrus; Temporal Mid, middle temporal gyrus.

Table 3. List of brain regions showing significant changes in BOLD response for the contrast of correct (switch n + (5–10)) versus error feedback (p < 0.001, FDR corrected for multiple comparisons) x

y 26 24

z 14 20

24 22

contains little behaviorally relevant information, because the participant in general is able to detect the error before receiving the error feedback by using the internal monitoring system (Heldmann et al., 2008).

Size

t-Value

Region

H

121 81

6.27 5.71

Hippocampus Parahippocampal

L R

MNI coordinates, size in number of voxels, t statistic values; cluster threshold 10 voxel. H, hemisphere; L, left; R, right.

mode) (Aston-Jones and Cohen, 2005; Sara, 2009; Cain et al., 2011; Sara and Bouret, 2012). In addition, this area showed no activation for true error feedback, e.g. when the participant made a mistake. In this case the feedback

Hippocampus activation When comparing correct feedback against the error feedback, we found increased activation of hippocampal and parahippocampal regions bilaterally for the correct

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235

feedback. This pattern of activations is consistent with a study linking hippocampal activation to the consolidation of stimulus response associations. In a modified Wisconsin Card Sorting Test, Graham et al. (2009) described increased hippocampal activation when participants received the positive feedback, that the applied sorting rule, and accordingly the applied task set, was correct. In their investigation, the strength of hippocampal activation was related to the temporal distance to the prior task set switch: the more often a sorting rule was successively valid, the stronger the hippocampal response to the positive feedback stimulus. Accordingly, we would like to argue, that in the present investigation the increase in hippocampal activation after correct feedback reflects the consolidation of correct stimulus–response associations.

CONCLUSION The current investigation suggests that task switching and subsequent behavioral adjustment is dependent on the interaction between the brainstem (most likely the noradrenergic LC) and a cortical network comprising the cingulate cortex, the precuneus and the insula in. The brainstem activation might indicate the recruitment of attentional resources triggered by the switch signal. In contrast, during performance errors we found activation of the Cerebellar Crus and the prefrontal cortex, which shows the difference between error processing and task switching. Furthermore it indicates an involvement of the cerebellum in cognitive control. In addition, a strong activation of the hippocampus was found after positive feedback in learnable stimulus– response associations, signaling a particularly strong activation of the declarative memory in these situations.

CONFLICT OF INTERESTS The authors declare that they have no conflict of interest.

REFERENCES Allen G, McColl R, Barnard H, Ringe WK, Fleckenstein J, Cullum CM (2005) Magnetic resonance imaging of cerebellar-prefrontal and cerebellar-parietal functional connectivity. Neuroimage 28:39–48. Allport DA, Styles EA, Hsieh S (1994) Shifting intentional set: exploring the dynamic control of tasks. In: Umilta C, Moscovitch M, editors. Attention and performance, vol. XV. Hillsdale: Erlbaum. p. 421–452. Arnsten AF (2006a) Fundamentals of attention-deficit/hyperactivity disorder: circuits and pathways. J Clin Psychiatry 67(suppl. 8):7–12. Arnsten AF (2006b) Stimulants: therapeutic actions in ADHD. Neuropsychopharmacology 31:2376–2383. Astafiev SV, Shulman GL, Stanley CM, Snyder AZ, Van Essen DC, Corbetta M (2003) Functional organization of human intraparietal and frontal cortex for attending, looking, and pointing. J Neurosci 23:4689–4699. Astafiev SV, Snyder AZ, Shulman GL, Corbetta M (2010) Comment on ‘‘Modafinil shifts human locus coeruleus to low-tonic, highphasic activity during functional MRI’’ and ‘‘homeostatic sleep pressure and responses to sustained attention in the suprachiasmatic area’’. Science 328:309. author reply 309.

233

Aston-Jones G, Cohen JD (2005) An integrative theory of locus coeruleus-norepinephrine function: adaptive gain and optimal performance. Annu Rev Neurosci 28:403–450. Badgaiyan RD, Posner MI (1998) Mapping the cingulate cortex in response selection and monitoring. Neuroimage 7:255–260. Balsters JH, Whelan CD, Robertson IH, Ramnani N (2013) Cerebellum and cognition: evidence for the encoding of higher order rules. Cereb Cortex 23:1433–1443. Barber AD, Carter CS (2005) Cognitive control involved in overcoming prepotent response tendencies and switching between tasks. Cereb Cortex 15:899–912. Berman SM, Naliboff BD, Suyenobu B, Labus JS, Stains J, Ohning G, Kilpatrick L, Bueller JA, Ruby K, Jarcho J, Mayer EA (2008) Reduced brainstem inhibition during anticipated pelvic visceral pain correlates with enhanced brain response to the visceral stimulus in women with irritable bowel syndrome. J Neurosci 28:349–359. Bouret S, Sara SJ (2005) Network reset: a simplified overarching theory of locus coeruleus noradrenaline function. Trends Neurosci 28:574–582. Brass M, von Cramon DY (2004) Selection for cognitive control: a functional magnetic resonance imaging study on the selection of task-relevant information. J Neurosci 24:8847–8852. Braver TS, Reynolds JR, Donaldson DI (2003) Neural mechanisms of transient and sustained cognitive control during task switching. Neuron 39:713–726. Brown 2nd DC, Co MS, Wolff RC, Atzori M (2012) Alpha-Adrenergic receptors in auditory cue detection: alpha2 receptor blockade suppresses false alarm responding in the rat. Neuropharmacology 62:2178–2183. Bunge SA, Hazeltine E, Scanlon MD, Rosen AC, Gabrieli JD (2002) Dissociable contributions of prefrontal and parietal cortices to response selection. Neuroimage 17:1562–1571. Bush G, Luu P, Posner MI (2000) Cognitive and emotional influences in anterior cingulate cortex. Trends Cogn Sci 4:215–222. Cain RE, Wasserman MC, Waterhouse BD, McGaughy JA (2011) Atomoxetine facilitates attentional set shifting in adolescent rats. Dev Cogn Neurosci 1:552–559. Chamberlain SR, Robbins TW (2013) Noradrenergic modulation of cognition: therapeutic implications. J Psychopharmacol 27:694–718. Craig AD (2011) Significance of the insula for the evolution of human awareness of feelings from the body. Ann N Y Acad Sci 1225:72–82. De Baene W, Brass M (2013) Switch probability context (in)sensitivity within the cognitive control network. Neuroimage 77:207–214. Debener S, Ullsperger M, Siegel M, Fiehler K, von Cramon DY, Engel AK (2005) Trial-by-trial coupling of concurrent electroencephalogram and functional magnetic resonance imaging identifies the dynamics of performance monitoring. J Neurosci 25:11730–11737. Demanet J, Liefooghe B, Verbruggen F (2011) Valence, arousal, and cognitive control: a voluntary task-switching study. Front Psychol 2:336. Demirci O, Stevens MC, Andreasen NC, Michael A, Liu J, White T, Pearlson GD, Clark VP, Calhoun VD (2009) Investigation of relationships between fMRI brain networks in the spectral domain using ICA and Granger causality reveals distinct differences between schizophrenia patients and healthy controls. Neuroimage 46:419–431. Devauges V, Sara SJ (1990) Activation of the noradrenergic system facilitates an attentional shift in the rat. Behav Brain Res 39:19–28. Devilbiss DM, Berridge CW (2008) Cognition-enhancing doses of methylphenidate preferentially increase prefrontal cortex neuronal responsiveness. Biol Psychiatry 64:626–635. Eriksen BA, Erkisen CW (1974) Effects of noise letters upon the identification of target letters in a non-search task. Percept Psychophys 16:143–149.

234

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235

Graham S, Phua E, Soon CS, Oh T, Au C, Shuter B, Wang SC, Yeh IB (2009) Role of medial cortical, hippocampal and striatal interactions during cognitive set-shifting. Neuroimage 45:1359–1367. Halgren E, Boujon C, Clarke J, Wang C, Chauvel P (2002) Rapid distributed fronto-parieto-occipital processing stages during working memory in humans. Cereb Cortex 12:710–728. Heldmann M, Russeler J, Munte TF (2008) Internal and external information in error processing. BMC Neurosci 9:33. Hester R, Foxe JJ, Molholm S, Shpaner M, Garavan H (2005) Neural mechanisms involved in error processing: a comparison of errors made with and without awareness. Neuroimage 27:602–608. Hester R, Nestor L, Garavan H (2009) Impaired error awareness and anterior cingulate cortex hypoactivity in chronic cannabis users. Neuropsychopharmacology 34:2450–2458. Holroyd CB, Coles MG (2002) The neural basis of human error processing: reinforcement learning, dopamine, and the errorrelated negativity. Psychol Rev 109:679–709. Holroyd CB, Nieuwenhuis S, Yeung N, Nystrom L, Mars RB, Coles MG, Cohen JD (2004) Dorsal anterior cingulate cortex shows fMRI response to internal and external error signals. Nat Neurosci 7:497–498. Hou RH, Freeman C, Langley RW, Szabadi E, Bradshaw CM (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. Huang S, Belliveau JW, Tengshe C, Ahveninen J (2012) Brain networks of novelty-driven involuntary and cued voluntary auditory attention shifting. PLoS One 7:e44062. Hyafil A, Summerfield C, Koechlin E (2009) Two mechanisms for task switching in the prefrontal cortex. J Neurosci 29:5135–5142. Ide JS, Li CS (2011) A cerebellar thalamic cortical circuit for errorrelated cognitive control. Neuroimage 54:455–464. Ide JS, Shenoy P, Yu AJ, Li CS (2013) Bayesian prediction and evaluation in the anterior cingulate cortex. J Neurosci 33:2039–2047. Jersild AT (1927) Mental set and shift. Arch Psychol 14(89):81. Jocham G, Ullsperger M (2009) Neuropharmacology of performance monitoring. Neurosci Biobehav Rev 33:48–60. Keren NI, Lozar CT, Harris KC, Morgan PS, Eckert MA (2009) In vivo mapping of the human locus coeruleus. Neuroimage 47:1261–1267. Kiesel A, Steinhauser M, Wendt M, Falkenstein M, Jost K, Philipp AM, Koch I (2010) Control and interference in task switching – a review. Psychol Bull 136:849–874. Klein TA, Endrass T, Kathmann N, Neumann J, von Cramon DY, Ullsperger M (2007) Neural correlates of error awareness. Neuroimage 34:1774–1781. Klein TA, Ullsperger M, Danielmeier C (2013) Error awareness and the insula: links to neurological and psychiatric diseases. Front Hum Neurosci 7:14. Kovach CK, Daw ND, Rudrauf D, Tranel D, O’Doherty JP, Adolphs R (2012) Anterior prefrontal cortex contributes to action selection through tracking of recent reward trends. J Neurosci 32:8434–8442. Lapish CC, Kroener S, Durstewitz D, Lavin A, Seamans JK (2007) The ability of the mesocortical dopamine system to operate in distinct temporal modes. Psychopharmacology (Berl) 191:609–625. Lapiz MD, Morilak DA (2006) Noradrenergic modulation of cognitive function in rat medial prefrontal cortex as measured by attentional set shifting capability. Neuroscience 137:1039–1049. Liddell BJ, Brown KJ, Kemp AH, Barton MJ, Das P, Peduto A, Gordon E, Williams LM (2005) A direct brainstem-amygdala-cortical ‘alarm’ system for subliminal signals of fear. Neuroimage 24:235–243. Lin FH, Agnew JA, Belliveau JW, Zeffiro TA (2009) Functional and effective connectivity of visuomotor control systems demonstrated using generalized partial least squares and structural equation modeling. Hum Brain Mapp 30:2232–2251.

Mars RB, Coles MG, Grol MJ, Holroyd CB, Nieuwenhuis S, Hulstijn W, Toni I (2005) Neural dynamics of error processing in medial frontal cortex. Neuroimage 28:1007–1013. Mayr U, Keele SW (2000) Changing internal constraints on action: the role of backward inhibition. J Exp Psychol Gen 129:4–26. McGaughy J, Ross RS, Eichenbaum H (2008) Noradrenergic, but not cholinergic, deafferentation of prefrontal cortex impairs attentional set-shifting. Neuroscience 153:63–71. Menon V, Adleman NE, White CD, Glover GH, Reiss AL (2001) Errorrelated brain activation during a Go/NoGo response inhibition task. Hum Brain Mapp 12:131–143. Minzenberg MJ, Watrous AJ, Yoon JH, Ursu S, Carter CS (2008) Modafinil shifts human locus coeruleus to low-tonic, high-phasic activity during functional MRI. Science 322:1700–1702. Monsell S (2003) Task switching. Trends Cogn Sci 7:134–140. Nagahama Y, Okada T, Katsumi Y, Hayashi T, Yamauchi H, Sawamoto N, Toma K, Nakamura K, Hanakawa T, Konishi J, Fukuyama H, Shibasaki H (1999) Transient neural activity in the medial superior frontal gyrus and precuneus time locked with attention shift between object features. Neuroimage 10:193–199. Orr C, Hester R (2012) Error-related anterior cingulate cortex activity and the prediction of conscious error awareness. Front Hum Neurosci 6:177. Pourtois G (2011) Early error detection predicted by reduced preresponse control process: an ERP topographic mapping study. Brain Topogr 23:403–422. Raizada RD, Poldrack RA (2007) Challenge-driven attention: interacting frontal and brainstem systems. Front Hum Neurosci 1:3. Renner KH, Beversdorf DQ (2010) Effects of naturalistic stressors on cognitive flexibility and working memory task performance. Neurocase 16:293–300. Riba J, Rodriguez-Fornells A, Morte A, Munte TF, Barbanoj MJ (2005) Noradrenergic stimulation enhances human action monitoring. J Neurosci 25:4370–4374. Ridderinkhof KR, Ullsperger M, Crone EA, Nieuwenhuis S (2004) The role of the medial frontal cortex in cognitive control. Science 306:443–447. Rushworth MF, Hadland KA, Gaffan D, Passingham RE (2003) The effect of cingulate cortex lesions on task switching and working memory. J Cogn Neurosci 15:338–353. Sara SJ (2009) The locus coeruleus and noradrenergic modulation of cognition. Nat Rev Neurosci 10:211–223. Sara SJ, Bouret S (2012) Orienting and reorienting: the locus coeruleus mediates cognition through arousal. Neuron 76:130–141. Schmidt C, Collette F, Leclercq Y, Sterpenich V, Vandewalle G, Berthomier P, Berthomier C, Phillips C, Tinguely G, Darsaud A, Gais S, Schabus M, Desseilles M, Dang-Vu TT, Salmon E, Balteau E, Degueldre C, Luxen A, Maquet P, Cajochen C, Peigneux P (2009) Homeostatic sleep pressure and responses to sustained attention in the suprachiasmatic area. Science 324:516–519. Schmidt C, Peigneux P, Maquet P, Phillips C (2010) Response to comment on ‘‘Homeostatic sleep pressure and responses to sustained attention in the suprachiasmatic area’’. Science 328:309. Schweizer TA, Oriet C, Meiran N, Alexander MP, Cusimano M, Stuss DT (2007) The cerebellum mediates conflict resolution. J Cogn Neurosci 19:1974–1982. Shibata E, Sasaki M, Tohyama K, Kanbara Y, Otsuka K, Ehara S, Sakai A (2006) Age-related changes in locus ceruleus on neuromelanin magnetic resonance imaging at 3 Tesla. Magn Reson Med Sci 5:197–200. Shulman GL, Pope DL, Astafiev SV, McAvoy MP, Snyder AZ, Corbetta M (2010) Right hemisphere dominance during spatial selective attention and target detection occurs outside the dorsal frontoparietal network. J Neurosci 30:3640–3651. Sterpenich V, D’Argembeau A, Desseilles M, Balteau E, Albouy G, Vandewalle G, Degueldre C, Luxen A, Collette F, Maquet P

J. von der Gablentz et al. / Neuroscience 285 (2015) 227–235 (2006) The locus ceruleus is involved in the successful retrieval of emotional memories in humans. J Neurosci 26:7416–7423. Steuwe C, Daniels JK, Frewen PA, Densmore M, Pannasch S, Beblo T, Reiss J, Lanius RA (2014) Effect of direct eye contact in PTSD related to interpersonal trauma: an fMRI study of activation of an innate alarm system. Soc Cogn Affect Neurosci 9(1):88–97. Tanaka Y, Fujimura N, Tsuji T, Maruishi M, Muranaka H, Kasai T (2009) Functional interactions between the cerebellum and the premotor cortex for error correction during the slow rate force production task: an fMRI study. Exp Brain Res 193:143–150. Ullsperger M, Nittono H, von Cramon DY (2007) When goals are missed: dealing with self-generated and externally induced failure. Neuroimage 35:1356–1364. Usher M, Cohen JD, Servan-Schreiber D, Rajkowski J, Aston-Jones G (1999) The role of locus coeruleus in the regulation of cognitive performance. Science 283:549–554. Vandewalle G, Schmidt C, Albouy G, Sterpenich V, Darsaud A, Rauchs G, Berken PY, Balteau E, Degueldre C, Luxen A, Maquet P, Dijk DJ (2007) Brain responses to violet, blue, and green

235

monochromatic light exposures in humans: prominent role of blue light and the brainstem. PLoS One 2:e1247. Wang C, Ulbert I, Schomer DL, Marinkovic K, Halgren E (2005) Responses of human anterior cingulate cortex microdomains to error detection, conflict monitoring, stimulus–response mapping, familiarity, and orienting. J Neurosci 25:604–613. Wenderoth N, Debaere F, Sunaert S, Swinnen SP (2005) The role of anterior cingulate cortex and precuneus in the coordination of motor behaviour. Eur J Neurosci 22:235–246. Whitmer AJ, Banich MT (2012) Brain activity related to the ability to inhibit previous task sets: an fMRI study. Cogn Affect Behav Neurosci 12:661–670. Yeung N, Nystrom LE, Aronson JA, Cohen JD (2006) Between-task competition and cognitive control in task switching. J Neurosci 26:1429–1438. Yu AJ, Dayan P (2005) Uncertainty, neuromodulation, and attention. Neuron 46:681–692. Zhang S, Li CS (2012) Functional networks for cognitive control in a stop signal task: independent component analysis. Hum Brain Mapp 33:89–104.

(Accepted 13 November 2014) (Available online 20 November 2014)