Cerebral Cortex Advance Access originally published online on August 2, 2006
Cerebral Cortex 2007 17(6):1314-1322; doi:10.1093/cercor/bhl042
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Time–Frequency Intracranial Source Localization of Feedback-Related EEG Activity in Hypothesis Testing
1 Laboratoire de Psychologie Cognitive, CNRS UMR 6146, Université de Provence, Marseille, France, 2 Functional Brain Imaging Unit, Wohl Institute for Advanced Imaging, Sourasky Medical Center, Tel Aviv, Israel, 3 Department of Computer Science, University College London, London, UK, 4 Neurosciences Cognitives et Imagerie Cérébrale, LENA, CNRS UPR 640, Centre Hospitalier Pitié-Salpêtrière, Paris, France
Address correspondence to David Papo, Functional Brain Imaging Unit, Wohl Institute for Advanced Imaging, Sourasky Medical Center, 6, Weizmann Street, 64239 Tel Aviv, Israel. Email: david.papo{at}chups.jussieu.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The neural correlates of the response to performance feedback have been the object of numerous neuroimaging studies. However, the precise timing and functional meaning of the resulting activations are poorly understood. We studied the electroencephalographic response time locked to positive and negative performance feedback in a hypothesis testing paradigm. The signal was convoluted with a family of complex wavelets. Intracranial sources of activity at various narrow-band frequencies were estimated in the 100- to 400-ms time window following feedback onset. Positive and negative feedback were associated to 1) early parahippocampo-cingular sources of alpha oscillations, more posteriorly located and long lasting for negative feedback and to 2) late partially overlapping neural circuits comprising regions in prefrontal, cingular, and temporal cortices but operating at feedback-specific latencies and frequencies. The results were interpreted in the light of neurophysiological models of feedback and were used to discuss methodological issues in the study of high-level cognitive functions, including reasoning and decision making.
Key Words: dipole source localization • EEG • feedback • hypothesis testing • time–frequency analysis
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Problem solving under uncertainty or incomplete information requires that people generate hypotheses and modify them according to their adequacy to either internally or externally generated feedback. Feedback from the environment can be used to adjust behavior and guide switches between induction phases, during which new regularities or rules are generated and selected, and deduction phases, during which previous positively tested ones are applied.
The neural correlates of the response to performance feedback have been studied with neuroimaging techniques. Early studies using positron emission tomography (Elliott and others 1997
, 1998; Elliott and Dolan 1998) showed that the response to both positive and negative feedback was associated in a task-dependent manner to a ventro-cortico-limbic network including the medial caudate nucleus and the orbitofrontal cortex and that activation of these structures was attenuated in depressed patients (Elliott and others 1998). In an event-related functional magnetic resonance study (Monchi and others 2001) using the Wisconsin Card Sorting test, positive and negative feedback were both associated to increased dorsolateral prefrontal cortex activity with respect to a control feedback condition; however, a loop involving the midventrolateral prefrontal cortex, the caudate nucleus, and the mediodorsal thalamic nucleus increased its activity specifically after receiving negative feedback, whereas positive feedback activated a limbico-cortico-striatal circuit originating in the medial prefrontal cortex (Elliott and others 1997; Monchi and others 2001), a region implicated in the mediation of reward, via dopaminergic modulation (Thut and others 1997; Koepp and others 1998). These patterns of activations were interpreted in terms of working memory monitoring and response planning, respectively. More recently, (van Veen and others 2004), in an functional magnetic resonance imaging (fMRI) study of feedback in a time estimation task, correct feedback stimuli, compared with error feedback stimuli, were shown to engage a distributed network of brain areas comprising the right rostral and posterior cingulate cortices, bilateral middle and inferior prefrontal and superior temporal gyri, and the right lateral and medial occipital cortices.
Although the results from these neuroimaging studies were by and large mutually consistent, the time-varying pattern of activations could only be inferred, due to insufficient temporal resolution of the employed techniques.
Optimal temporal resolution was achieved in electroencephalographic (EEG) studies of trial-to-trial feedback in tasks including time-interval estimation (Miltner and others 1997), monetary gains and losses (Gehring and Willoughby 2002; Nieuwenhuis and others 2004), guessing (Ruchsow and others 2002), stimulus–response association learning (Müller and others 2005), and hypothesis testing (Papo and others 2003). The results of these studies showed that positive and negative feedback conditions were associated to topographically and chronometrically separable patterns of electrical brain activity. Positive trial-to-trial feedback elicited a parietocentral P300, whereas negative feedback was characterized by P200–N240 complexes with a frontal–central onset and lasting longer than in posterior regions.
Intracranial dipole source localization analyses have been carried out relative to the negative peaks of the event-related signal in the 200- to 400-ms time window. Dipole sources were found, respectively, in the anterior cingulate cortex (ACC) (Miltner and others 1997), in the medial prefrontal and posterior cingulate cortices (Müller and others 2005), and in the ACC as well as in the inferior frontal cortex (Ruchsow and others 2002). However, these studies took into account broadband averaged signals, where the contributions from oscillatory activities at different frequencies could not be separated and were dominated by low-frequency high-amplitude activity. Moreover, the event-related potential (ERP) signal typically takes into account frequencies <30 Hz. Thus, the contribution from different frequency bands to the observed activity could not be captured, although fast oscillating activity in the 
band (>30 Hz) was not taken into account. As a consequence, the dipole source localization analysis lacked resolution in the time–frequency domain.
Neurophysiological models of feedback and error-related processing integrating evidence from various research streams have also been proposed (Holroyd and Coles 2002; Papo and others 2003). In particular, Papo and others (2003) proposed that the response to feedback is associated to early activation of a hippocampo-cortical network, independent of feedback value; this mechanism would regulate feedback-specific reward-related dopaminergic and stress-driven noradrenergic corticostriatal circuits and would adaptively reset horizontal fronto-parieto-temporal interactions mediating working memory of associations between feedback and hypothesis-related material.
In the present study, we sought direct evidence for the implication of key neuroanatomical regions associated by Papo and others (2003) to the production of P200–N240 and P300 complex waves. Moreover, we conjectured that, because the ERP signal was dominated by low-frequency signals, more transient functionally significant aspects of electrical activity could have been overshadowed. To explore the possibility of frequency-specific activity and to disentangle the contribution of sources of oscillatory activity at different frequencies, we analyzed the same data of the study of Papo and others.
Contrary to similar existing electrophysiological investigations 1) dipole source analysis (Scherg and Ebersole 1993) was carried out at various narrow-band frequencies rather than relative to the broadband signal; 2) dipole sources were estimated at various latencies in the 100- to 400-ms time window, rather than concentrating on a single peak in that interval; and 3) contrary to ERP studies, a large frequency spectrum, including the range (30–100 Hz), was taken into account.
We predicted that positive and negative feedback would be associated to a common early hippocampo-frontal network; that the latter would differentially modulate activity associated to working memory maintenance of task-related material in prefrontal, cingular, and parietotemporal areas; and that this differential modulation would start around 200 ms after feedback onset, once hippocampo-prefrontal mismatch-related activity is completed and would have its effect in the P300 time window. Although no specific predictions were formulated as to the specific frequency bands at which these activities would be found, we tested whether enhanced precision in the time–frequency domain could provide indications as to the status of a recorded activity at any given spatial region and time.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Subjects
Thirteen right-handed graduate and undergraduate students volunteered in the experiment. Subjects had normal or corrected to normal vision and no history of neurological or psychiatric disease. Two of them could not be kept for further analyses due to excessive rates of recording artifacts; whereas one subject was excluded because she was not naive as to the task's manipulations. The 10 remaining subjects (5 women and 5 men; mean age = 25.2 years, range 21–30 years) were all blind as to the experiment's objectives.
Stimuli and Procedure
Subjects were sat in front of a computer screen and judged, using a double-button press device, whether or not triplets of numbers were instances of a hidden rule chosen by the experimenter. At the beginning of each trial, one triplet was presented. Following a time interval varying between 800 and 1200 ms, subjects received the experimenter's feedback on the computer screen, indicating whether their performance was correct or not. Triplet, subjects' response, and feedback stayed on the screen till 1 s after feedback onset. The screen was then offset before the following trial could start. There was a 3- to 5-s intertrial interval. There were 30 blocks of 10 trials each, corresponding to 30 different hidden rules. Subjects were informed that the rule was changed at the end of each block of trials. Successive blocks were separated by time intervals of the order of 30 s/1 min. Unbeknownst to the subjects, feedbacks were controlled by the experimenter and not by subjects' performance, as there were no rules behind the triplets. Feedback frequency was manipulated, so that an equal number of blocks had 8:2, 5:5, and 2:8 positive-to-negative feedback ratios. In the (8:2) and (2:8) blocks, the last 5 trials were paired to 5 consecutive positive and negative feedback, respectively, whereas in all blocks, the first 5 trials comprised either 2 or 3 positive/negative feedback responses. The order of presentation of blocks was quasi-randomized across subjects.
EEG Recording
Brain electrical activity was recorded from a standard 62 electrode montage, with electrodes positioned according to the extended 10-20 System location, with a nasion reference. The electro-oculogram was also recorded, for blink, vertical, and horizontal eye movement correction (Gratton and others 1983). The EEG was amplified (0.05–100 Hz) and digitized at 500 Hz over a 900-ms epoch, including a 200-ms prestimulus baseline.
Time–Frequency and Dipole Source Analyses
The EEG signal was convolved over the whole 200- to 700-ms time window with a family of complex Gabor wavelets w (t, f0), Gaussian in both the time (standard deviation [SD]: 
t) and frequency domain (SD: f), with respect to the central frequency 
normalizes the total power to 1. We used a wavelet family defined by the constant ratio f0/f = 7 (the wavelet duration is given by 2t, i.e., approximately 2 cycles at f0), with f0 ranging from 1 to 100 Hz. For each frequency band, power variations across time were given by E(t,f0)=|w(t,f0)xs(t)|2. Finally, for each feedback condition, dipole source localization using the brain electric source analysis (BESA) algorithm (Version 4.2.28) was performed at several frequency ranges (6, 10, 30, and 40 Hz), on the grand average of the convolved signal, around the peaks in the 100- to 400-ms time window.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We report results regarding the dipole source estimations carried out in the present study. For the sake of clarity, we first provide a brief summary of behavioral and ERP results, a full account of which can be found in Papo and others (2003)
. Behavioral Performance and Summary of ERP Analysis
Figure 1 shows mean response times for the first and last 5 trials in each block, for each block type. A feedback frequency (3 levels: 8:2, 2:8, and 5:5) x within block position (2 levels: beginning and end) analysis of variance revealed that response times increased as a function of the frequency of negative feedback (F2,18 = 4.93, P < 0.001)), with a significant feedback frequency x block position interaction (F1,9 = 15.308, P < 0.001 between 8:2 and 2:8 feedback ratios). Recalling that feedback was pre-established and thus no correct answer existed, response times provided the only indirect behavioral criterion for correct performance. This finding is coherent with verbal debriefings carried out for each subject after the experiment, revealing that none of the subjects was aware of the experimental manipulation. Within a single trial, subjects' response precedes the experimenter's feedback. Thus, in principle, response time is independent of feedback type. The fact that response times varied according to feedback frequency indicated the meaningfulness of this experimental manipulation. Finally, as expected, response times relative to the first 5 trials in each block did not significantly differ across blocks with different positive-to-negative feedback ratios.
|
Figure 2 shows the grand averaged ERPs elicited by positive and negative feedback stimuli. Mean amplitudes for all positive and negative feedback trials, irrespective of block type, were measured relative to a 200-ms prestimulus baseline at all electrodes.
|
A spatiotemporal principal component analysis showed early left frontal and midline frontocentral and late centro-posterior feedback-related activity, with comparable morphological characteristics. Positive feedback elicited a P300, whereas negative feedback was characterized by P200–N240 complexes with a frontal–central onset and lasting longer than in posterior regions.
Dipole Source Estimation
Dipole source estimations started with one dipole, and further dipoles were added until residual variance was not improved any longer. For all estimations, convergence was fast and robust with respect to initial dipole localization. In some instances, the goodness of fit was improved with symmetry constraints; however, in all these cases dipole localization was comparable with that without symmetry constraints. There were overall more source estimations for positive feedback, as greater signal dispersion across the scalp for negative feedback sometimes hampered reliable dipole source localization, especially at higher frequencies. Figure 3 shows an example of dipole source analysis.
|
Table 1 shows the results of all the dipole source localization carried out around 6, 10, 30, and 40 Hz for all electrodes and for both positive and negative feedback. Figure 4 shows the time-varying sequence of the various dipoles estimated at different frequencies.
|
|
The earliest peaks of activity were found in the
band. This activity was prominent at occipitoparietal areas where it peaked earlier for positive (130 ms) than for negative feedback (200 ms) (P < 0.05). This result was significant at the individual level for all but one subject. Dipole analysis carried out around these peaks showed that positive and negative feedback were associated to the early activity of 2 bilateral parahippocampal and midline cingular dipole sources, Brodmann areas (BAs) 32 and 23/31 for positive and negative feedback, respectively. In the 200- to 400-ms time window, the signal associated to positive feedback was modeled by dipoles in the premotor (BA 6, 30 Hz, 310 ms), posterior cingular (BA 31, 40 Hz, 310 ms), and frontal areas (inferior and median frontal gyri, 40 Hz, 310 ms) and in the anterior cingulate (BA 24, 6 Hz, 330 ms; BA 32, 367 ms, 10 Hz) and superior temporal gyri bilaterally (367 ms, 10 Hz). For negative feedback, a dipole was estimated in the superior temporal gyrus, bilaterally (30 Hz, 210 ms), although activity at P300 latencies was modeled by bilateral inferior and middle prefrontal activities (BAs 9 and 44, 6 Hz, 280 ms).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The EEG signal associated to positive and negative feedback in a hypothesis testing paradigm was examined in the time–frequency domain. Both feedback conditions were associated to sustained early
-band activity, originating from a largely overlapping parahippocampo-cingular source. At P300 latencies, positive and negative feedback were associated to spatially partially overlapping brain circuits comprising inferior frontal, cingulate, and superior temporal regions, which nonetheless operated at different frequencies and with a different temporal pattern. We discuss the meaning of activity at these frequencies, and the implications for the neurophysiology of feedback. Some methodological remarks concerning the neuroimaging of various high-level cognitive functions are also proposed.
Early Hippocampo-Cingular -Band Activity
From an electrophysiological viewpoint, task-induced power increases originating from anterior cerebral and medial temporal brain regions have already been reported (Ba
ar and others 2001; Connemann and others 2005). Moreover, it has been suggested that when a stimulus produces oscillatory activity at frequencies around 10 Hz in various brain regions, this activity fulfills a signal communication function between these regions (Baar and others 1997).
Peak latencies of the early -band activity in each feedback condition can be interpreted in terms of early visual processing and hippocampo-medial prefrontal resonance. Parahippocampal cortex and ACC can respond within 100-ms stimulation to visual and auditory stimulations (Wilson and others 1983; Thorpe and others 1996). The resonant process would produce 
40-ms-long hypothesis testing cycles (Grossberg 1984), a multiple of conduction latencies in the hippocampo-prefrontal region which are estimated at 15.6 ± 3.6 (Thierry and others 2000). With negative feedback, mismatch-induced resonance would be associated to longer lasting resonance, whereas a feedback signal to the temporal cortex would entail a further 60- to 100-ms delay (Rolls 2000). Our results thus suggest the presence of additional hypothesis testing cycles within a circuit comprising the hippocampus and the cingulum. Moreover, the -band peak latency for negative feedback coincides with the P200–N240 revealed by ERP analysis (Papo and others 2003), which was proposed to reflect activation of an orienting system, merging with persistent reset-related inhibition (Grossberg 1984). This would be coherent with a parahippocampal-cingular role in sensory gating (Baar and others 1997), orienting (Tesche and Karhu 2000), as well as in the processing of stimulus emotional content (Adolphs 2002) and saliency (Phillips and others 2003).
In one possible interpretation of the hippocampo-cingular coactivation, these 2 regions, which have reciprocal connections (Öngür and Price 2000), may be part of a network integrating feedback valence to recall, working memory, selective attention, and action monitoring operations (Petrides and Pandya 2002). The visual feedback signal would separately reach the parahippocampal and ACC regions (Thorpe and others 1996). The parahippocampal region may encode rule-reinforcement associations, whereas the ACC may represent a functional link between medial temporal and parieto-occipital regions (Ruff and others 2003).
These results can also be interpreted in the light of a model of feedback in which hippocampo-cingular activity acts as a flip-flop modulating a reward-related network (Papo and others 2003). On one hand, reward-related dopaminergic modulation in the hippocampal region can alter medial prefrontal activity, via a depression of the subicular output (Behr and others 2000). The medial prefrontal region selectively responds to reward (O'Doherty and others 2001), amplifying its meaning via corticostriatal circuits, as the hippocampal stress response is held in check (Kim and others 2001). With positive feedback, the early subsiding of 10-Hz cycles of afferent hippocampo-cingular input would enable frontal regions to respond to reward-related dopaminergic reinforcement (Floresco and Grace 2003). On the other hand, the hippocampal formation predicts ACC activity and amplifies aversive events triggering adaptive behavioral responses (Ploghaus and others 2001). Moreover, the hippocampus prevents excessive stress-related inhibition and terminate the error-related processing following negative feedback (Grossberg 1984). The ACC, in turn, facilitates fast detection and evaluation of motivationally relevant environmental stimuli by maintaining sustained tonic activity (Gusnard and others 2001). Early -band activity associated to positive feedback may then be explained in terms of stress-related maintenance of ACC activity. Response evaluation processes associated to negative feedback may be associated to hippocampo-posterior cingular activity, mediating rule-feedback association, and to concurrent task-related attenuation of ACC activity (Simpson and others 2001), allowing the affective response to guide decision making (Dikman and Allen 2000).
Our results may also have some implications for the characterization of the response to feedback in various pathologies. The parahippocampal gyrus has been shown to be deactivated during symptom provocation in subjects suffering from anxiety (Bremner and others 1999), panic attacks (Malizia and others 1998), and posttraumatic stress disorder (Nutt and Malizia 2004). Moreover, it was proposed that the ERP P200–N240 wave complex is controlled by serotonergic and noradrenergic neuromodulations in the hippocampus (Papo and others 2003), which indirectly regulate anxiety by acting on hippocampo-prefrontal -aminobutyric acidergic activity (Malizia and others 1998; Fingelkurts and others 2004). Our results suggest a relationship between anxiety and feedback-related hippocampo-cingular -band activity duration. Abnormal response to negative feedback that is thought to specifically characterize unipolar depression (Beck 1967; Beats and others 1996; Elliott and others 1998) may also result from abnormal hippocampo-prefrontal resonance (Rocher and others 2004).
Response in the N200–P300 Range
Negative Feedback
The early 30-Hz peak for negative feedback (210 ms) was associated to the activity of dipoles in the superior and median temporal gyri. The simultaneity with -band activity of hippocampo-cingular origin suggests that this activity may be related to an evaluative function under uncertainty and to the inhibition of invalidated rules (Paulus and others 2001). It is interesting to compare this early high-frequency superior temporal dipole associated to negative feedback with the later one, associated to positive feedback, located in a similar region but generated by low-frequency activity (367 ms, 10 Hz). Whereas in the latter case, the temporal activation may reflect a reinforcement of rule-related activity in the Papez circuit, in the former, it may reflect an opposite inhibitory function.
Dipole sources at frequencies could not reliably be estimated at N200–P300 latencies. However, -band activity was characterized by widespread synchronization across the scalp surface, circumscribing a right centro-parietal desynchronization region with a topographic location comparable with the desynchronization region in positive feedback. This centro-parietal scalp region may be implicated in rule-specific information processing (Dehaene and others 1999), whereas the overall pattern may be interpreted in terms of focal activation within and diffuse activity suppression without the focal desynchronization region (Pfurtscheller 2003). Persistent centroparietal synchronization together with strong 30-Hz frontal synchronization for negative but not positive feedback may reflect an inhibitory function (Neuper and Pfurtscheller 2001) and indicate active suppression of invalidated rules by the attentional system (Jensen and others 2002).
For negative feedback, 
-band oscillations immediately prior to P300 onset (280 ms) were explained in terms of the activity of 2 prefrontal dipoles (median and inferior frontal gyri). Dipoles with similar locations have been found in association with negative feedback in a guessing paradigm (Ruchsow and others 2002) and were interpreted as reflecting error processing (Carter and others 1998) or novel rule generation once the current rule was falsified (Wolford and others 2000). Moreover, based on the simultaneous activity of a dipole in the ACC, Ruchsow and others (2002) concluded that the underlying negative deflection in the ERP reflected a comparator function between expected and actual results. Given its latency, this activity is unlikely to reflect the generation of novel hypotheses, whereas error detection should occur at an earlier stage and in circuits outside the ACC (Holroyd and Coles 2002; Stemmer and others 2003). Interestingly, the inferior frontal gyrus was implicated at comparable latencies in positive feedback, but at frequencies. It is then reasonable to assume that this inferior region had a different, or even opposite, role for the 2 feedback conditions.
Positive Feedback
At P300 latencies, positive feedback was characterized by prominent -band oscillations. Separable dipole sources could be estimated for 30 and 40 Hz, respectively, located in the premotor and in the inferior frontal and posterior cingulate cortices. The dissociation of intracranial sources was equally visible on the scalp surface in the form of a topographic dissociation between 30 and 40 Hz activity, the former being essentially frontocentral, the latter prefrontal and frontopolar. These coexisting dipoles and patterns of activity may represent a correlate in the frequency domain of a complex prefrontocingular interplay. The frontocentral regions may depend on anteriorly located frontal ones for information necessary to interpret the current value of feedback, as it serves an alerting function mobilizing the affective systems; the midline frontocentral regions may either directly or indirectly feed the frontopolar one with the information necessary to promote appropriate compensatory behavior (Tucker and others 1999; Gehring and Knight 2000). The consequences of these activities on further rule-related activity are suggested by the implication of the posterior cingulate cortex, an area cytoarchitectonically different from the anterior portion of the cingulate region (Bush and others 2000) thought to fulfill evaluative functions (Vogt and others 1992), particularly of the valence of emotionally unpleasant stimuli (Maddock 1999).
Positive feedback was also associated to -band activity originating from a brain region in the affective portion of the ACC (BA 24) (Drevets and Raichle 1998; Bush and others 2000), ventral to a region associated to monitoring functions (Botvinick and others 1999). A similar localization was reported for -band activity associated to the execution of a working memory task (Gevins and others 1997). This finding was taken as evidence for the implication of the Papez circuit, which connects, via re-entrant loops, the hippocampal formation to cingulate cortex. -Band oscillations in this circuit are thought to be instrumental in the creation of long-range corticocortical synchronization (Buzsaki 1996; von Stein and Sarnthein 2000; Vertes and others 2001) and could play a crucial role in the synchronization of emotional evaluations (Lewis 2005).
These results are coherent with an interpretation in terms of reward-related cingular loop (O'Doherty and others 2001). Activity within this loop would reinforce the memory trace associated to the positively tested hypothesis. This interpretation is coherent with an encoding function associated to both the P300 and activity in the frequency range (Klimesch 1999; Jensen and Tesche 2002). It is reinforced by the simultaneous presence of -band (8–12 Hz) desynchronization and -band (4–7 Hz) synchronization, which are thought to reflect information transfer between memory systems (Sauseng and others 2002).
A hippocampal implication at P300 latencies could have been expected at frequencies (Kahana and others 1999; Tesche and Karhu 2000; Hasselmo and others 2002), reflecting the operating mode of hippocampo-cortical feedback circuits associated to stimulus context evaluation (Buzsaki 1996; Denham and Borisyuk 2000). Our results show that the early hippocampal oscillatory activity may be short lived, coherently with a marginal role of the hippocampus in the production of the P300 (Polich and Squire 1993; Nishitani and others 1999).
At frequencies, positive feedback was associated both to the activity of dipoles in the ACC and superior temporal gyrus and to diffuse desynchronization, reaching its maximum in a centroparietal scalp region. In the context of executive functioning and feedback-related activity, the ACC has been associated to strategic and action selection processes involved in conflict reduction between competing alternatives (Botvinick and others 1999). However, the present context involved neither a strategic choice nor action selection. In alternative, anterior cingular activation may indicate that the activity is associated to response checking (Luu and others 2000). Various ERP studies of performance feedback (e.g., Miltner and others 1997; Gehring and Willoughby 2002; Suchan and others 2003) have estimated dipole sources in the ACC in association with negative deflections of the ERP response around 250 ms after feedback onset. These results were corroborated by an fMRI study of both performance and error-related feedback (Holroyd and others 2004); this study highlighted the implication of an ACC area very close to the dipole that was active at P300 latencies and at frequencies for positive feedback. In these studies, ACC activity was generally interpreted in terms of error evaluation (Gemba and others 1986; Luu and others 2000; Stemmer and others 2003) and emotional and cognitive integration (Badgaiyan and Posner 1998; Bush and others 2000; Gehring and Fencsik 2001) processes. Taken together, this -band cingular activity may be interpreted in terms of rule-related short-term memory activity (Klimesch and others 1999).
The ACC implication in the mediation of various types of feedback has been a matter of debate in the literature (Holroyd and others 2004; van Veen and others 2004). In some fMRI studies of feedback (van Veen and others 2004), differences between positive and negative feedback were associated to posterior but not anterior cingulate activation. Interestingly, activations found in van Veen's study, including posterior cingulate, middle and inferior prefrontal, and superior temporal areas, closely mirror the dipole sources estimated in the present study for positive feedback. The present study suggests that feedback may indeed be associated to activity in different portions of both anterior and posterior cingulate cortices. However, local subcomponents of this pattern of activity, which is intermittent and characterized by feedback-specific latencies and frequencies, may be difficult to tease apart when time–frequency resolution is insufficient.
Methodological Remarks
A few methodological remarks arise from this study.
- Our study indicates that, to be fully specified, spatial information concerning task-specific activations should be complemented by temporal information at time scales much smaller than those of the phenomenon under examination, as well as by information concerning the operating modes of the observed activity, for example, its characteristic frequencies.
- Although not allowing to precise whether a given region receives a net excitatory or inhibitory input, dipole analysis in narrow frequency bands provides indirect indications as to the status of a recorded activity in a given region at a given time.
- The substantial overlap of the activation pattern in the present study with activations highlighted in neuroimaging studies of various high-level cognitive functions, including decision making (Ernst and others 2002; Paulus and others 2002) and reasoning (Elliott and Dolan 1998; Parsons and Osherson 2001; Luo and others 2003), involving the use of external feedback in the evaluation of performance, suggests that poor temporal resolution amplifies the impact of impulsive responses such as those associated to feedback.
- Insufficient temporal resolution may impede the detection of the early short-lived parahippocampo-cingular sources of feedback-related activity, which may therefore be difficult to capture through standard functional neuroimaging techniques. Although the hippocampal activation may simply vanish in contrasts between task conditions, the anterior cingulate one may be confounded with later activity in neighboring portions of the same region.
- The discrepancy between the occipitoparietal early scalp topography and parahippocampo-cingular intracranial localization suggests that interpreting scalp surface in terms of scalp localization may sometimes be misleading and that more anteriorly located early sources are also possible.
Caveats and Conclusion
It must be stressed that, despite the robustness of our results, particularly for early dipole sources, the results from dipole source localization methods always need to be taken with some caution. Moreover, given the small sample and the EEG apparatus used in the present study, our results ought to be considered as preliminary and await corroboration from studies including larger cohorts and using denser electrode montages.
The electrophysiological response to positive and negative performance feedback in a hypothesis testing task was associated to 1) early activity of largely overlapping hippocampo-cingular sources operating at frequencies around 10 Hz and 2) later activity of a complex prefronto-cingulo-temporal network, active at frequencies and latencies modulated by feedback type. We suggest that the former activity represents a mechanism that may act by switching on and off a corticocortical fronto-posterior working memory–related activity, whereas the latter reflects differential condition-specific consequences of feedback.
|
Acknowledgments |
|---|
DP is supported by an annual postdoctoral grant of the Fondation pour la Recherche Médicale, Paris, France. Conflict of Interest: None declared.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Adolphs R. Neural systems for recognizing emotion. Curr Opin Neurobiol (2002) 12:169–177.[CrossRef][Web of Science][Medline]
Badgaiyan RD, Posner MI. Mapping the cingulate cortex in response selection and monitoring. Neuroimage (1998) 7:255–260.[CrossRef][Web of Science][Medline]
Baar E, Baar-Ero
lu C, Karaka S, Schürmann M. Gamma, alpha, delta, and theta oscillations govern cognitive processes. Int J Psychophysiol (2001) 39:241–248.[CrossRef][Web of Science][Medline]
Baar E, Schürmann M, Baar-Erolu C, Karaka S. Alpha oscillations in brain functioning: an integrative approach. Int J Psychophysiol (1997) 26:5–29.[CrossRef][Web of Science][Medline]
Beats BC, Sahakian BJ, Levy R. Cognitive performance in tests sensitive to frontal lobe dysfunction in elderly depressed. Psychol Med (1996) 26:591–603.[Web of Science][Medline]
Beck AT. Depression: clinical, experimental, and theoretical aspects (1967) New York: Harper and Row.
Behr J, Glovell T, Schmitz D, Heinemann U. Dopamine depresses excitatory synaptic transmission onto rat subicular neurons via presynaptic D1-like dopamine receptors. J Neurophysiol (2000) 84:112–119.
Botvinick M, Nystrom LE, Fissell K, Carter CS, Cohen JD. Conflict monitoring versus selection-for-action in anterior cingulate cortex. Nature (1999) 402:179–181.[CrossRef][Medline]
Bremner JD, Narayan M, Staib LH, Southwick SM, McGlashan T, Charney DS. Neural correlates of memories of childhood sexual abuse in women with and without posttraumatic stress disorder. Am J Psychiatry (1999) 156:1787–1795.
Bush G, Luu P, Posner MI. Cognitive and emotional influences in anterior cingulate cortex. Trends Cogn Sci (2000) 4:215–222.[CrossRef][Web of Science][Medline]
Buzsaki G. The hippocampal-neocortical dialogue. Cereb Cortex (1996) 6:81–92.
Carter CS, Braver TS, Barch DM, Botvinick M, Noll D, Cohen JD. Anterior cingulate cortex, error detection and online monitoring of performance. Science (1998) 280:747–749.
Connemann BJ, Mann K, Lange-Asschenfeldt C, Ruchsow M, Schreckenberger M, Bartenstein P, Grunder G. Anterior limbic alpha-like activity: a low resolution electromagnetic tomography study with lorazepam challenge. Clin Neurophysiol (2005) 116:886–894.[CrossRef][Web of Science][Medline]
Dehaene S, Spelke E, Pinel P, Stanescu R, Tsvikin S. Sources of mathematical thinking: behavioural and brain imaging evidence. Science (1999) 284:970–974.
Denham MJ, Borisyuk RM. A model of theta rhythm production in the septal-hippocampal system and its modulation by ascending brain stem pathways. Hippocampus (2000) 10:698–716.[CrossRef][Web of Science][Medline]
Dikman ZV, Allen JJB. Error monitoring during reward and avoidance learning in high- and low-socialized individuals. Psychophysiology (2000) 37:43–54.[CrossRef][Web of Science][Medline]
Drevets WC, Raichle ME. Reciprocal suppression of regional cerebral blood flow during emotional versus higher cognitive processes: implications for interactions between emotion and cognition. Cogn Emotion (1998) 12:353–385.[CrossRef]
Elliott R, Dolan RJ. Activation of different anterior cingulate foci in association with hypothesis testing and response selection. Neuroimage (1998) 8:17–29.[CrossRef][Web of Science][Medline]
Elliott R, Frith CD, Dolan RJ. Differential neural response to positive and negative feedback in planning and guessing tasks. Neuropsychologia (1997) 35:1395–1404.[CrossRef][Web of Science][Medline]
Elliott R, Sahakian BJ, Paykel ES, Dolan RJ. Abnormal neural response to feedback on planning and guessing tasks in patients with unipolar depression. Psychol Med (1998) 28:559–571.[CrossRef][Web of Science][Medline]
Ernst M, Bolla K, Mouratidis M, Contoreggi C, Matochik JA, Kurian V, Cadet J-L, Kimes AS, London ED. Decision-making in a risk-taking task: a PET study. Neuropsychopharmacology (2002) 26:682–691.[CrossRef][Web of Science][Medline]
Fingelkurts AA, Fingelkurts AA, Kivisaari R, Pekkonen E, Ilmoniemi RJ, Kähkönen S. The interplay of lorazepam-induced brain oscillations: microstructural electromagnetic study. Clin Neurophysiol (2004) 115:674–690.[CrossRef][Web of Science][Medline]
Floresco SB, Grace AA. Gating of hippocampal-evoked activity in prefrontal cortical neurons by inputs from the mediodorsal thalamus and ventral tegmental area. J Neurosci (2003) 23:3930–3943.
Gehring WJ, Fencsik DE. Functions of the medial frontal cortex in the processing of conflict and errors. J Neurosci (2001) 21:9430–9437.
Gehring WJ, Knight RT. Prefrontal-cingulate interactions in action monitoring. Nat Neurosci (2000) 3:516–520.[CrossRef][Web of Science][Medline]
Gehring WJ, Willoughby AR. The medial frontal cortex and the rapid processing of monetary gains and losses. Science (2002) 295:2279–2282.
Gemba H, Sasaki K, Brooks VB. Error potentials in limbic cortex (anterior cingulate area-24) of monkeys during motor learning. Neurosci Lett (1986) 70:223–227.[CrossRef][Web of Science][Medline]
Gevins A, Smith ME, McEvoy L, Yu D. High-resolution EEG mapping of cortical activation related to working memory: effects of task difficulty, type of processing, and practice. Cereb Cortex (1997) 7:374–385.
Gratton G, Coles MG, Donchin E. A new method for off-line removal of ocular artifact. Electroencephalogr Clin Neurophysiol (1983) 55:468–484.[CrossRef][Web of Science][Medline]
Grossberg S. Some psychophysiological and pharmacological correlates of a developmental, cognitive, and motivational theory. In: Karrer R, Cohen J, Tueting P, editors. Brain and information: event-related potentials. Ann N Y Acad Sci (1984) 425:58–151.[Web of Science][Medline]
Gusnard DA, Akbudak E, Shulman GL, Raichle ME. Medial prefrontal cortex and self-referential mental activity: relation to a default mode of brain function. Proc Natl Acad Sci USA (2001) 98:4259–4264.
Hasselmo ME, Bodelón C, Wyble BP. A proposed function for hippocampal theta rhythm: separate phases of encoding and retrieval enhance reversal of prior learning. Neural Comput (2002) 14:793–817.[CrossRef][Web of Science][Medline]
Holroyd C, Coles MGH. The neural basis of human error processing: reinforcement learning, dopamine, and the error-related negativity. Psychol Rev (2002) 109:679–709.[CrossRef][Web of Science][Medline]
Holroyd C, Larsen JT, Cohen JD. Context dependence of the event-related brain potential associated with reward and punishment. Psychophysiology (2004) 41:245–253.[CrossRef][Web of Science][Medline]
Jensen O, Gelfand J, Kounios J, Lisman JE. Oscillations in the alpha band (9-12 Hz) increase with memory load during retention in a short-term memory task. Cereb Cortex (2002) 12:877–882.
Jensen O, Tesche C. Frontal theta activity in humans increases with memory load in a working memory task. European J Neurosci (2002) 15:1395–1400.[CrossRef][Web of Science][Medline]
Kahana MJ, Sekuler R, Caplan JB, Kirschen M, Madsen JR. Human theta oscillations exhibit task dependence during virtual maze navigation. Nature (1999) 399:781–784.[CrossRef][Medline]
Kim J, Lee HJ, Han J-S, Packard MG. Amygdala is critical for stress-induced modulation of hippocampal long-term potentiation and learning. J Neurosci (2001) 21:5222–5228.
Klimesch W. EEG alpha and theta oscillations reflect cognitive memory performance: a review and analysis. Brain Res Rev (1999) 29:169–195.[CrossRef][Medline]
Klimesch W, Doppelmayr M, Schwaiger J, Auinger P, Winkler T. ‘Paradoxical’ alpha synchronization in a memory task. Cogn Res (1999) 7:493–501.[CrossRef]
Koepp MJ, Gunn RN, Lawrence AD, Cuningham VJ, Dagher A, Jones T, Brooks DJ, Bench CJ, Grasby PM. Evidence for striatal dopamine release during a video game. Nature (1998) 393:266–268.[CrossRef][Medline]
Lewis MD. Bridging emotion theory and neurobiology through dynamic systems modeling. Behav Brain Sci (2005) 28:169–194.[CrossRef][Web of Science][Medline]
Luo Q, Perry C, Peng D, Jin Z, Xu D, Ding G, Xu S. The neural substrate of analogical reasoning. Cogn Brain Res (2003) 17:527–534.[CrossRef][Medline]
Luu P, Flaisch T, Tucker DM. Medial frontal cortex in action monitoring. J Neurosci (2000) 20:464–469.
Maddock RJ. Retrosplenial cortex and emotion. New insights from functional imaging of the human brain. Trends Neurosci (1999) 22:310–316.[CrossRef][Web of Science][Medline]
Malizia AL, Cunnigham VJ, Bell CJ, Liddle PF, Jones T, Nutt DJ. Decreased GABAA-benzodiazepine receptor binding in panic disorder. Arch Gen Psychiatry (1998) 55:715–720.
Miltner WHR, Braun CH, Coles MGH. Event-related brain potentials following incorrect feedback in a time estimation task: evidence for a "generic" neural system for error correction. J Cogn Neurosci (1997) 9:788–798.[Web of Science]
Monchi Q, Petrides M, Petre V, Worsley K, Dagher A. Wisconsin Card Sorting revisited: distinct neural circuits participating in different stages of the task identified by event-related functional magnetic resonance imaging. J Neurosci (2001) 21:7733–7741.
Müller SV, Möller J, Rodriguez-Fornells A, Münte TF. Brain potentials related to self-generated and external information used for performance monitoring. Clin Neurophysiol (2005) 116:63–74.[CrossRef][Web of Science][Medline]
Neuper C, Pfurtscheller G. Event-related dynamics of cortical rhythms: frequency-specific features and functional correlates. Int J Psychophysiol (2001) 43:41–58.[CrossRef][Web of Science][Medline]
Nieuwenhuis S, Yeung N, Holroyd CB, Schurger A, Cohen J. Sensitivity of electrophysiological activity from medial frontal cortex to utilitarian and performance feedback. Cereb Cortex (2004) 14:741–747.
Nishitani N, Ikeda A, Nagamine T, Honda M, Mikuni N, Taki W, Kimura J, Shibasaki H. The role of the hippocampus in auditory processing studied by event-related electric potentials and magnetic fields in epilepsy patients before and after temporal lobectomy. Brain (1999) 122:687–707.
Nutt DJ, Malizia AL. Structural and functional brain changes in posttraumatic stress disorder. J Clin Psychiatry (2004) 65:11–17.
O'Doherty J, Kringelbach ML, Rolls ET, Hornak J, Andrews C. Abstract reward and punishment representations in the human orbitofrontal cortex. Nat Neurosci (2001) 4:95–102.[CrossRef][Web of Science][Medline]
Öngür D, Price JL. The organization of networks within the orbital and medial prefrontal cortex of rats, monkeys and humans. Cereb Cortex (2000) 10:206–219.
Papo D, Baudonnière P-M, Hugueville L, Caverni J-P. Feedback in hypothesis testing: an ERP study. J Cogn Neurosci (2003) 15:508–522.[CrossRef][Web of Science][Medline]
Parsons LM, Osherson D. New evidence fora distinct right and left brain systems for deductive versus probabilistic reasoning. Cereb Cortex (2001) 11:954–965.
Paulus MP, Hozack N, Frank L, Brown GG. Error rate and outcome predictability affect neural activation in prefrontal cortex and anterior cingulate during decision-making. Neuroimage (2002) 15:836–846.[CrossRef][Web of Science][Medline]
Paulus MP, Hozack N, Zauscher BE, McDowell JE, Frank L, Brown GG, Braff DL. Prefrontal, parietal, and temporal cortex networks underlie decision-making in the presence of uncertainty. Neuroimage (2001) 13:91–100.[Web of Science][Medline]
Petrides M, Pandya DN. Association pathways of the prefrontal cortex and functional observations. In: Principals of frontal lobe function—Stuss DT, Knight RT, eds. (2002) New York: Oxford University Press. 31–50.
Pfurtscheller G. Induced oscillations in the alpha band: functional meaning. Epilepsia (2003) 44(suppl):2–8.
Phillips ML, Drevets WC, Rauch SL, Lane R. Neurobiology of emotion perception I: the neural basis of normal emotion perception. Biol Psychiatry (2003) 54:504–514.[CrossRef][Web of Science][Medline]
Ploghaus A, Narain C, Beckmann CF, Clare S, Bantick S, Wise R, Matthews PM, Rawlins JNP, Tracey I. Exacerbation of pain by anxiety is associated with activity in a hippocampal network. J Neurosci (2001) 21:9896–9903.
Polich J, Squire LR. P300 from amnesic patients with bilateral hippocampal lesions. Electroencephalogr Clin Neurophysiol (1993) 86:408–417.[CrossRef][Web of Science][Medline]
Rocher C, Spedding M, Munoz C, Jay TM. Acute stress-induced changes in hippocampal/prefrontal circuits in rats: effects of antidepressants. Cereb Cortex (2004) 14:224–229.
Rolls ET. Hippocampo-cortical and cortico-cortical backprojections. Hippocampus (2000) 10:380–388.[CrossRef][Web of Science][Medline]
Ruchsow M, Grothe J, Spitzer M, Kiefer M. Human anterior cingulate cortex is activated by negative feedback: evidence from event-related potentials in a guessing task. Neurosci Lett (2002) 325:203–206.[CrossRef][Web of Science][Medline]
Ruff CC, Knauff M, Fangmeier T, Spreer J. Reasoning and working memory: common and distinct neuronal processes. Neuropsychologia (2003) 41:1241–1253.[CrossRef][Web of Science][Medline]
Sauseng P, Klimesch W, Gruber W, Doppelmayr M, Stadler W, Schabus M. The interplay between theta and alpha oscillations in the human electroencephalogram reflects the transfer of information between memory systems. Neurosci Lett (2002) 324:121–124.[CrossRef][Web of Science][Medline]
Scherg M, Ebersole JS. Models of brain sources. Brain Topogr (1993) 5:419–423.[CrossRef][Medline]
Simpson JR Jr, Drevets WC, Snyder AZ, Gusnard DA, Raichle ME. Emotion-induced changes in human medial prefrontal cortex: II. During anticipatory anxiety. Proc Natl Acad Sci USA (2001) 98:688–693.
Stemmer B, Segalowitz SJ, Witzke W, Schönle PW. Error detection in patients with lesions to the medial prefrontal cortex: an ERP study. Neuropsychologia (2003) 42:118–130.[CrossRef][Web of Science]
Suchan B, Zoppelt D, Daum I. Frontocentral negativity in electroencephalogram reflects motor response evaluation in humans on correct trials. Neurosci Lett (2003) 350:101–104.[CrossRef][Web of Science][Medline]
Tesche CD, Karhu J. Theta oscillations index human hippocampal activation during a working memory task. Proc Natl Acad Sci USA (2000) 97:919–924.
Thierry A-M, Gioanni Y, Dégénéatis E, Glowinski J. Hippocampo-prefrontal cortex pathway: anatomical and electrophysiological characteristics. Hippocampus (2000) 10:411–419.[CrossRef][Web of Science][Medline]
Thorpe S, Fize D, Marlot C. Speed of processing in the human visual system. Nature (1996) 381:520–522.[CrossRef][Medline]
Thut G, Schultz W, Roelcke U, Nienhusmeier M, Missimer J, Maguire RP, Leenders KL. Activation of the human brain by monetary reward. Neuroreport (1997) 8:1225–1228.[Web of Science][Medline]
Tucker DM, Hartry-Speiser A, McDougal L, Lure P, deGrandpre D. Mood and spatial memory: Emotion and the right hemisphere contribution to spatial cognition. Biol Psychiatry (1999) 50:103–125.
van Veen V, Holroyd CB, Cohen JD, Stenger VA, Carter CS. Errors without conflict: implications for performance monitoring theories of anterior cingulate cortex. Brain Cogn (2004) 56:267–276.[CrossRef][Web of Science][Medline]
Vertes RP, Albo Z, Viana di Prisco G. Theta-rhythmically firing neurons in the anterior thalamus: implications for mnemonic functions of Papez's circuit. Neuroscience (2001) 104:619–625.[CrossRef][Web of Science][Medline]
Vogt BA, Finch DM, Olson CR. Functional heterogeneity in cingulate cortex: the anterior executive and posterior evaluative regions. Cereb Cortex (1992) 2:435–443.
von Stein A, Sarnthein J. Different frequencies for different scales of cortical integration: from local gamma to long range alpha/theta synchronization. Int J Psychophysiol (2000) 38:301–313.[CrossRef][Web of Science][Medline]
Wolford G, Miller MB, Gazzaniga M. The left hemisphere's role in hypothesis testing (2000) 1–4: J Neurosci 20: RC64.
Wilson CL, Babb TL, Halgren E, Crandall PH. Visual receptive fields and response properties of neurons in human temporal lobe and visual pathways. Brain (1983) 106:473–502.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||