Skip to main content
NCBI home page
Social Cognitive and Affective Neuroscience logoLink to Social Cognitive and Affective Neuroscience
. 2024 Sep 20;19(1):nsae063. doi: 10.1093/scan/nsae063

Neural dynamics underlying the illusion of control during reward processing

Ya Zheng 1,2,‡,*, Canming Yang 3,‡, Huiping Jiang 4, Bo Gao 5,6
PMCID: PMC11466228  PMID: 39300953

Abstract

The illusion of control refers to a behavioral bias in which people believe they have greater control over completely stochastic events than they actually do, leading to an inflated estimate of reward probability than objective probability warrants. In this study, we examined how reward system is modulated by the illusion of control through the lens of neural dynamics. Participants in a behavioral task exhibited a classical illusion of control, assigning a higher value to the gambling wheels they picked themselves than to those given randomly. An event-related potential study of the same task revealed that this behavioral bias is associated with reduced reward anticipation, as indexed by the stimulus-preceding negativity, diminished positive prediction error signals, as reflected by the reward positivity, and enhanced motivational salience, as revealed by the P300. Our findings offer a mechanistic understanding of the illusion of control in terms of reward dynamics.

Keywords: illusion of control, reward processing, neural dynamics, ERPs

Introduction

Imaging that you decide to buy a lottery ticket. Would you use numbers that feel lucky to you (e.g. your birth date), or would you just let a computer quick pick your numbers? If you select your own numbers, you are probably falling into the trap of the illusion of control: People believe they have greater control over events than they actually do. This behavioral bias has been demonstrated by Langer with a series of experiments using games that were governed by chance and is defined as “an expectancy of a personal success probability inappropriately higher than the objective probability would warrant” (Langer 1975, p. 311). In one of these experiments, participants who chose their own lottery tickets were less willing to sell or exchange their tickets than those who were given lottery tickets at random, despite the same probability of winning the lottery for both groups. Subsequent research has documented that this cognitive bias is driven by several psychological mechanisms including affective or motivational enhancement (Leotti et al. 2010), effort investment (Patall et al. 2008), regret anticipation (Bar-Hillel and Neter 1996), and preexisting false beliefs (Klusowski et al. 2021).

Given that the illusion of control inflates estimates of reward probability, it is possible that the illusion of control modulates neural processes involved in the subjective judgement of reward events. Kool et al. (2013) examined this possibility in a previous study using functional magnetic resonance imaging (fMRI). Participants in the study exhibited a classic illusion-of-control effect by giving higher estimates of winning probability for gambles they had picked themselves (i.e. their choice was accepted) than for identical gambles that were imposed on them (i.e. their choice was vetoed). Based on the behavioral effect, the authors hypothesized that the illusion of control would reduce neural prediction error signal, that is, the difference between the predictive value of future events and their realized value (Sutton et al. 1978, Schultz et al. 1997). However, brain areas encoding reward prediction error signals showed similar activity in response to gambling outcomes, regardless of whether the gambles were selected at will or imposed randomly. Instead, greater activation was observed in the anterior cingulate cortex and precuneus following accepted choices compared to vetoed choices, suggesting that the illusion of control enhances motivational salience during reward consumption. A following fMRI study observed that compared to those who reported no illusion of control, participants experiencing an illusion of control showed increased neural responses during reward anticipation within a corticostriatal network including the ventral striatum and right inferior frontal gyrus (Lorenz et al. 2015). While these fMRI studies highlight the modulation of neural correlates underlying reward processing by the illusion of control, there are important issues that cannot be addressed because of the sluggish characteristics of the hemodynamic response. Most importantly, studies using fMRI are not suitable to pinpoint and distinguish neural responses occurring closely in time such as the anticipatory and consummatory phases during reward processing (Berridge and Robinson 2003, Meyer et al. 2021).

Here, we aimed to examine how reward system is modulated by the illusion of control through the lens of neural dynamics. We used the event-related potentials (ERPs) instead of the fMRI because ERPs have a higher temporal resolution, which enables the characterization of reward dynamics from anticipation to consumption. We hypothesized that the neural signatures associated with reward anticipation and consumption would be affected after the initial registration of the illusion of control. The stimulus-preceding negativity (SPN) is an ERP component primarily associated with reward anticipation. It is a slow, nonmotoric, negative-going wave that increases its amplitude as feedback arrives and is manifest as a right hemisphere preponderance (for reviews, see Brunia et al. 2011, Hackley et al. 2014). Previous studies have demonstrated that the SPN is sensitive to reward probability, reflecting either an approach motivation toward desirable but unlikely rewards (Fuentemilla et al. 2013, Zheng et al. 2020a) or an avoidance motivation in anticipating negative events (Umemoto and Holroyd 2017). If the illusion of control inflates the estimate of reward probability, then the anticipatory SPN should be more negative when participants experience an illusion of control.

In terms of reward consumption, the two most relevant ERP components are the reward positivity (RewP) and the subsequent P300. The RewP (also known as the feedback-related negativity) is a positive-going deflection peaking between 250 and 350 ms over frontocentral areas following positive feedback but is absent or reduced following negative feedback (Proudfit 2015). According to a computational framework, the RewP represents a prediction error signal, indicating as a teaching signal whether the delivered outcome is better (positive prediction error) or worse (negative prediction error) than expected (Holroyd and Coles 2002, Walsh and Anderson 2012). However, it is controversial whether the RewP is sensitive to signed reward prediction errors (Heydari and Holroyd 2016, Mulligan and Hajcak 2018, Zheng and Mei 2023) or unsigned salience prediction errors (Oliveira et al. 2007, Talmi et al. 2013, Hoy et al. 2021). According to the signed reward prediction error model, more-than-expected gains and less-than-expected losses are favorable outcomes and thus should elicit a more positive RewP than unfavorable outcomes (i.e. less-than-expected gains and more-than-expected losses). In contrast, the unsigned salience prediction error model predicts a more positive RewP for more-than-expected than less-than-expected outcomes, regardless of whether they are positive or negative. Under the prediction error framework, the illusion of control would make positive outcome less surprising and negative outcome more surprising, leading to a shift in prediction error. The RewP component would represent this modulation of prediction error by the illusion of control. Specifically, the reward prediction error model would predict reduced RewP responses to both gains (less-than-expected) and losses (more-than-expected) under the illusion of control. In contrast, the salience prediction error model would predict reduced RewP responses to gains but enhanced RewP responses to losses under the illusion of control. The P300 is a positive deflection peaking between 300 and 500 ms after stimulus onset over parietal areas (Sutton et al. 1978). This component is typically larger for gains than for losses and has been thought to reflect attention allocation based on motivational salience (Nieuwenhuis et al. 2005). Given the role of motivational salience in control (Kool et al. 2013, Wang et al. 2021), the P300 should be increased by the illusion of control.

Previous studies have shown that these ERP components are sensitive to perceived control, which is typically exercised through choice (Leotti et al. 2010). Both anticipatory (the SPN) and consummatory (the RewP and P300) ERP components are enhanced when a high level of control is perceived during active choices compared to a low level of control during passive no-choices (Yeung et al. 2005, Masaki et al. 2010, Mühlberger et al. 2017, Chen et al. 2018, Mei et al. 2018, Yi et al. 2018, Hassall et al. 2019, Hackley et al. 2020, Zhang et al. 2021, Bikute et al. 2022). However, these studies have two main limitations. First, one prerequisite for the illusion-of-control effect is that the outcome should be independent of behavioral selection to exclude the confounding effects of action-outcome contingency (Tricomi et al. 2004). In many previous control-related studies, participants were unaware of the outcome probability, which might lead them to believe they could discern a pattern to earn rewards in the choice versus no-choice condition (e.g. Chen et al. 2018, Mei et al. 2018). Second, decision effort was not controlled between choice and no-choice conditions. Participants typically make a choice themselves in the choice condition but follow a computer’s choice in the no-choice condition. This manipulation makes the choice condition more effortful than the no-choice condition, as evidenced by significant differences in reaction times between the two conditions (e.g. Masaki et al. 2010, Zheng et al. 2020b, Zhang et al. 2021). Previous studies have demonstrated the impact of effort expenditure on the anticipatory and consummatory ERP components (Ma et al. 2014, Yi et al. 2020, Bogdanov et al. 2022, Wu and Zheng 2023), suggesting that the ERP differences between choice and the no-choice conditions might be confounded with decision effort. As a result, the electrocortical mechanisms underlying the illusion of control remain elusive.

In this study, we conducted two experiments to characterize the neural dynamics underlying the illusion of control during reward processing. In Experiment 1, we aimed to demonstrate the illusion of control using a behavioral paradigm adapted from Kool et al. (2013). Participants in this task made choices to earn monetary rewards. Their choices could either be accepted or vetoed to maintain or destroy the illusion of control, respectively. The illusion of control is operatively defined as an overestimation of reward probability compared to the objective probability. In Experiment 2, we used an electroencephalogram (EEG) version of the paradigm used in Experiment 1 to record the anticipatory and consummatory ERP signals associated with the illusion of control. Our central hypothesis is that neural signatures associated with reward prediction and prediction errors would be modulated after the registration of the illusion of control. As illustrated in Fig. 1a, the anticipatory SPN should be more negative when participants’ choices were accepted than vetoed, because the illusion of control inflates the estimate of reward probability. As a reliable neural signature for prediction errors, if the RewP is sensitive to the valence of prediction errors, we should observe a main effect of control such that both gains and losses would elicit a smaller RewP following accepted versus vetoed choices (Fig. 1b, left). Otherwise, if the RewP encodes the unsigned salience prediction error, we should expect an interaction between valence and the illusion of control. Gains would elicit a smaller RewP following accepted versus vetoed choices, while losses would elicit a larger RewP following accepted versus vetoed choices (Fig. 1b, right).

Figure 1.

Figure 1.

Open in a new tab

Hypothesized value system dependence on the illusion of control for the SPN elicited in the anticipatory phase (a) and the RewP elicited in the consummatory phase (b).

Note: The RewP pattern differs between the signed reward prediction error (RPE) model (left) and the unsigned salience prediction error (UPE) model (right). A.U. = arbitrary unit.

Experiment 1

In Experiment 1, we aimed to replicate the behavioral illusion-of-control effect in a wheel-of-fortune task adapted from Kool et al. (2013).

Materials and methods

Participants

Twenty right-handed volunteers (10 females; M = 19.85 years, SD = 2.06) participated in Experiment 1. The sample size was chosen to match that used in the previous study (Kool et al. 2013). All participants had normal or corrected-to-normal vision and were free of neurological or psychological disorders. Each participant received a fixed payment of ¥15 for task performance and signed a written informed consent before the experiment. The study was approved by the local institutional review board.

Procedure

Participants performed a wheel-of-fortune task modified from a previous study (Kool et al. 2013). In each trial (Fig. 2), participants were presented with three identical gambling wheels located at the corners of an invisible, randomly rotated equilateral triangle. Each wheel contained yellow and white areas with a black arrow. Winning and loss probabilities were indicated explicitly by the yellow and white proportions of the wheel. These proportions varied from trial to trial, reflecting winning probabilities of 25%, 50%, or 75% and corresponding loss probabilities of 75%, 50%, or 25%. However, the proportions remained the same across all three wheels in each trial. Participants were explicitly and truthfully informed that the three wheels had an equal chance of winning or losing in each trial. Despite this, they were required to select one for gambling. This manipulation ensured that no preferences were introduced based on perceived differences in outcome probabilities, thus ruling out the potential confounding effects of action-outcome contingency. The graphic of the three wheels remained on the screen until participants chose one using a mouse. The starting position of the mouse cursor was always at the center of the screen to prevent any response bias toward a particular wheel due to its proximity. Upon their response, the black circle around the chosen wheel turned blue, highlighting the choice for 1000 ms.

Figure 2.

Figure 2.

Open in a new tab

The wheel-of-fortune task.

Note: Participants selected one from three equal gambling wheels, and then their choice was either accepted or vetoed. Afterwards, the outcome of the ultimately selected wheel was resolved. In Experiment 1, participants additionally reported their confidence in winning the gamble before outcome delivery.

After a fixation for 1000 ms, the participants’ choice was either accepted or vetoed in the subsequent approval phase. If accepted, the wheel stayed highlighted while the surrounding blue circle turned red for 1500 ms; if vetoed, the blue circle reverted to black, and the computer randomly selected one of the other two wheels for participants, turning it red for 1500 ms. After another 1000 ms fixation, participants were asked to rate their confidence in how likely the selected wheel would yield a winning outcome. This was done by moving the mouse cursor along a continuous scale (a horizontal rectangle) ranging from “definite loss” to “definite win.” Participants were instructed to estimate their confidence based on their gut feelings. The black arrow of the selected wheel then rotated with its stopping position determining the outcome. However, participants were told the rotation was invisible. The outcome was presented 1000 ms later as either “+10” (indicating a gain of ¥1) or “-10” (indicating a loss of ¥1) for 1000 ms. Each trial ended with an intertrial interval of 1000 ms.

Before the experiment, participants were provided with an initial endowment of ¥15 and were informed that their final payment would include a base payment plus earnings from each trial. Unbeknownst to participants, the outcome of each trial was predetermined to follow the exact probability indicated by the gambling wheels. Therefore, all participants always received a bonus of ¥15. The task consisted of 96 trials divided into two blocks of 48 trials each, with a brief break between blocks. Prior to the main experiment, participants were given 12 practice trials to familiarize themselves with the task. After the gambling task, participants completed a 9-point Likert scale (1 = “not at all”; 9 = “very much”) to rate their degree of preference, perceived control, interest, attention, and regular pattern for both the accepted and vetoed trials.

Data analysis

Participants’ confidence scores were analyzed using a repeated measures variance analysis (ANOVA), with approval (accepted, vetoed) and winning probability (25%, 50%, 75%) as within-subjects factors. Greenhouse–Geisser epsilon correction was applied when factors had more than two levels. Post hoc comparisons were corrected via the Bonferroni procedure. Partial eta-squared (ηp2) was reported as a measure of effect size. Post-experimental rating data were analyzed via a series of paired-sample t-tests. All statistical analyses were performed in R v4.2.1.

Results and discussion

In Experiment 1, our objective was to confirm that the current task can yield a reliable illusion-of-control effect. In particular, we predicted that participants would report higher confidence when their choices were accepted than vetoed. As shown in Fig. 3a, confidence ratings increased monotonically as a function of winning probability, F(2, 38) = 47.93, P < .001, ηp2 = 0.72. Participants had greater confidence when their choices were accepted over those that were vetoed, F(1, 19) = 8.45, P = .009, ηp2 = 0.31. Moreover, the approval effect was significant when the winning probability was 75% (P < .001), but not at 50% (P = .057) and 25% (P = .088), as reflected by a significant interaction between probability and approval, F(2, 38) = 8.41, P = .002, ηp2 = 0.31. The confidence data demonstrated that our behavioral task is robust enough to produce an illusion-of-control effect, a false belief that keeping the initial choice increases the likelihood of desired outcomes in a scenario determined by chance, especially when the winning probability is high.

Figure 3.

Figure 3.

Open in a new tab

Raincloud plots of confidence scores in response to accepted and vetoed trials as a function of winning probability for Experiment 1 (a) and Experiment 2 (b).

Note: The density plots depict the distributions, the boxplots represent the median and the 1st and 3rd quartiles, and the colored circles and dots indicate the mean for each participant and across participants, respectively. Error bars represent the within-subject standard error of the mean.

Table 1 shows the post-experimental rating data for the accepted and vetoed conditions. Results showed that participants exhibited a higher preference for the accepted condition than the vetoed condition (P < .001) and were more interested in the accepted condition than the vetoed condition (P < .001), indicating that our experimental manipulation changed the affective/motivational salience of the outcomes. However, no significant differences were found between the two conditions in terms of perceived control (P > .90), attention (P = .222), or pattern degree (P = .149), suggesting that participants’ control belief and overall task engagement were comparable across the conditions.

Table 1.

Post-experimental rating data of preference, control, interest, attention, and pattern (M ± SD).

Experiment 1 Experiment 2
Accepted Vetoed P Cohen’s d Accepted Vetoed P Cohen’s d
Preference 6.60 ± 2.11 3.40 ± 1.50 <.001 0.98 6.65 ± 1.33 4.10 ± 1.37 <.001 1.07
Control 4.65 ± 2.16 4.65 ± 1.95 >.90 <0.01 4.95 ± 2.04 4.47 ± 2.01 .024 0.37
Interest 7.25 ± 1.41 4.05 ± 2.33 <.001 1.09 6.33 ± 1.62 5.00 ± 1.91 .003 0.51
Attention 6.55 ± 1.67 5.70 ± 2.51 .222 0.28 6.30 ± 1.71 5.63 ± 1.81 .053 0.32
Pattern 4.70 ± 3.03 3.80 ± 2.71 .149 0.34 4.65 ± 2.53 4.70 ± 2.30 .826 0.04
Open in a new tab

Statistically significantly P values (<.05, two-sided) are shown in bold.

Experiment 2

In Experiment 2, we used the EEG technique to measure the anticipatory and consummatory reward activity associated with the illusion of control. Given the illusion-of-control effect observed at the winning probability of 75% in Experiment 1, we hypothesized that the ERP correlates underlying the illusion of control would be more pronounced when reward probability was high (i.e. 75%).

Materials and methods

Participants

We chose a larger sample size for the EEG experiment. Forty-one right-handed volunteers who did not participate in Experiment 1 were recruited for Experiment 2. Data from one participant were excluded due to an insufficient number of artifact-free trials (less than 50% of trials) for the ERP analysis. The final sample thus included 40 participants (20 females) with a mean age of 19.42 years (SD = 1.43). All participants had a normal or corrected-to-normal vision and reported no history of neurological or psychological disorders. Each received ¥10 for participation and a fixed bonus of ¥40 based on task performance. All participants signed a written informed consent before the experiment. This study was approved by the local institutional review board.

Procedure

This experimental task (Fig. 2) closely followed the design of Experiment 1, except that we recorded EEG signals and changed the task parameters as follows. First, we removed the continuous scale on each trial. Instead, participants were asked to estimate their overall confidence levels for the six trial types (i.e. the combinations of accepted/vetoed and 25%/50%/75% winning probability trials) via a 9-point Likert scale (1 = “not at all”; 9 = “very much”) after the EEG experiment. We switched from the continuous scale to the 9-point Likert scale to test the robustness of the behavioral illusion-of-control effect. Second, we added an anticipatory period with a fixation for 2500 ms after the approval phase to differentiate reward anticipation from reward consumption. Third, the experimental task consisted of 360 trials divided into six blocks of 60 trials, with a short break between blocks. Fourth, participants were provided with an initial endowment of ¥40 before the formal experiment.

Recording and analysis

Continuous EEG signals were recorded using an elastic cap embedded with 64 Ag/AgCl electrodes based on the extended International 10/20 system. The signals were referenced online to the left mastoid electrode and rereferenced offline to the average of the activity at the left and right mastoids. Horizontal eye movements were monitored via a pair of electrodes placed on the left and right outer canthi. Blinks and vertical eye movements were detected using another pair of electrodes placed above and below the left eye. The EEG signals were amplified and digitalized via a Neuroscan SynAmps2 amplifier with a low pass of 100 Hz in DC acquisition mode at a sampling of 500 Hz. Electrode-to-skin impedances were kept under 5 KΩ throughout the experiment.

The EEG data were analyzed offline on MATLAB (v2019b; Math-Works, USA) using the EEGLAB toolbox (v13.1.1; Delorme and Makeig 2004). For the SPN analysis, the raw data were filtered with a band pass of 0.01 and 30 Hz (roll-off 6 dB/octave) and then epoched from −4000 to 2000 ms relative to feedback onset with the activity from −2400 to −2200 ms serving as the baseline. For the RewP and P300 analyses, the raw EEG data were filtered with a band pass of 0.1 and 30 Hz (roll-off 6 dB/octave), and epochs were defined as 2000 ms prior to and 2000 ms relative to feedback onset with a baseline of −200 to 0 ms. All epoched data were screened for artifacts using the Fully Automated Statistical Thresholding algorithm with default parameters (Nolan et al. 2010). Bad electrodes were interpolated from neighboring electrodes using the EEGLAB spherical interpolation algorithm. The data were then entered into an infomax independent component analysis (ICA; Delorme and Makeig 2004) to remove ocular artifacts (i.e. eye blinks and movements). To remove other nonspecific artifacts, the ICA-corrected data were applied with a semi-automated procedure to the time windows of interest (−2500 to 500 ms for the SPN and −500 to 1000 ms for the RewP and P300) on all electrodes to remove epochs with a voltage difference  > 50 μV between sample points, a voltage difference  > 200 μV within a trial, or a maximum voltage difference  < 0.5 μV within 100 ms intervals, as well as based on visual inspection. Because trial number was relatively small for the RewP and P300 analyses for each probability condition (see Table S1 in Supplemental Materials for the trial count information), we extracted the single-trial instead of condition-averaged ERP data for statistical analyses (Heise et al. 2022).

Single-trial ERP amplitude was measured using a region-of-interest approach. Measurement parameters were determined based on the grand-averaged ERP waveforms and topographic maps (Fig. 4) collapsed across all conditions and all participants (Luck and Gaspelin 2017). Specifically, the SPN was measured as the mean activity from −500 ms to 0 ms before feedback onset over left and right frontotemporal areas (F7, FT7, F8, FT8); the RewP from 260 to 360 ms post feedback onset over frontocentral areas (Fz, FC1, FCz, FC2); the P300 from 350 to 500 ms over centroparietal areas (CPz, P1, Pz, P2). These single-trial data of interest were exported into R v4.2.1 for further analyses.

Figure 4.

Figure 4.

Open in a new tab

Topographic distribution maps for the SPN (−500–0 ms), the RewP (260–360 ms), and the P300 (350–500 ms) as a function of approval and winning probability.

Note: The maps for the RewP represent the difference waves between gain and loss feedback. The black dots present electrodes used for quantification.

All estimates and statistics on ERP data were obtained by fitting mixed-effects single-trial regression models with varying intercepts and slopes (unstructured covariance matrix) using the lme4 package (v1.1.28; Bates et al. 2015b). Fixed effects of the model for SPN data included approval (accepted, vetoed), winning probability (25%, 50%, 75%), and their interactions, whereas those for RewP and P300 data included approval, probability, valence (gain, loss), and their interactions. The predictors of approval and valence were contrast coded (−0.5 for vetoed and +0.5 for accepted; −0.5 for loss and +0.5 for gain), and the predictor of winning probability was orthogonally polynomial coded to examine the linear and quadratic trends. For the RewP analyses, we further controlled for variation in the preceding P2 amplitude of 170–210 ms (z-scored within participants) to account for the component overlap between the RewP and the P2 (Holroyd et al. 2003, San Martín et al. 2013). Random effects were determined using singular value decomposition to prevent the model from being overparameterized and degenerated relative to the information in data (Bates et al. 2015a, Matuschek et al. 2017). Satterthwaite approximation was used to estimate degrees of freedom and obtain two-tailed P values, as implemented in lmerTest (v3.1.3; Kuznetsova et al. 2017). Follow-up pairwise tests of significant interactions were conducted on estimated marginal means using the emmeans package (v1.7.1.1; Lenth 2022). For all models, plots of residuals against fitted values and Q-Q plots revealed no obvious deviations from normality and homoscedasticity. Variance inflation factors were less than 1.50, revealing no problems with multicollinearity.

Results and discussion

Behavioral and rating data

In a repeated-measure ANOVA model including approval and probability factors, we observed that participants were more confident in obtaining gains as the winning probability increased, F(2, 78) = 16.11, P < .001, ηp2 = 0.29, and when their choices were accepted than when they were vetoed, F(1, 39) = 7.39, P = .010, ηp2 = 0.16. Importantly, these main effects were modulated by a significant interaction between probability and approval, F(2, 78) = 12.66, P < .001, ηp2 = 0.25. Post hoc comparisons (Fig. 3b) showed that confidence ratings were significantly higher for the accepted than the vetoed condition when the winning probability was 75% (P < .001). No significant differences of confidence ratings were found when the winning probability was 25% (P = .078) or 50% (P = .264). These results demonstrate an illusion of control, especially when there is a high probability of winning, mirroring the pattern observed at the trial level in Experiment 1.

Similar to Experiment 1 (Table 1), participants in Experiment 2 exhibited a clear preference for the accepted condition (P < .001) and were more interested in the accepted versus vetoed trials (P = .003). Unlike Experiment 1, participants felt more control when their choices were accepted than when vetoed (P = .024), which may be attributed to a larger sample size and thus greater power in Experiment 2. Similar to Experiment 1, no significant condition differences were found for attention (P = .053) and pattern degree (P = .826).

ERP data

We observed a pronounced SPN (Fig. 5a; split-half reliability r = 0.82) during the anticipatory phase, as well as a RewP (Fig. 5b; split-half reliability r = 0.92) and a P300 (Fig. 5c; split-half reliability r = 0.95) during the consummatory phase. Importantly, these ERP components are obviously modulated by approval and winning probability (see Tables S2 and S3 in Supplemental Materials for full coefficient estimates for each ERP model).

Figure 5.

Figure 5.

Open in a new tab

Grand-averaged ERP waveforms for each condition during the anticipatory stage over left and right frontotemporal areas (a) and during the consummatory phase over frontocentral areas (b) and centroparietal areas (c).

Note: Shaded areas represent time windows used for quantification. The SPN data were applied with a low-pass cutoff at 7 Hz for visualization.

During the anticipatory phase (Fig. 5a), single-trial SPN amplitude was significantly modulated by approval (B = 0.54, P = .013) such that smaller SPN amplitude was observed when participants’ choices were accepted than when it was vetoed. The SPN amplitude was increased linearly as a function of winning probability (B = −0.48, P = .010). Winning probability did not moderate the overall effect of approval on the amplitude of the SPN, as revealed by the nonsignificant interaction between approval and probability (linear: B = −0.28, P = .451; quadratic: B = 0.10, P = .794).

During the consummatory phase (Fig. 5b), we found that gains compared with losses led to larger RewP amplitudes (B = 1.00, P < .001). We observed a significant Approval × Probability × Valence interaction for RewP data (linear: B = −1.11, P = .031). Follow-up pairwise comparisons indicated that when choices were vetoed, the effect of reward on the RewP was significant when the winning probability was 75% (B = 2.03, Z = 4.47, P < .003) but not when it was 25% (B = 0.76, Z = 1.67, P = .095) or 50% (B = 0.63, Z = 1.54, P = .124). In contrast, when choices were accepted, the effect of reward on the RewP was significant when winning probability was 50% (B = 1.25, Z = 3.07, P = .002) but not when it was 25% (B = 0.72, Z = 1.59, P = .113) or 75% (B = 0.59, Z = 1.31, P = .192). Moreover, gains from accepted choices elicited a smaller RewP than gains from vetoed choices when the winning probability was 75% (B = −0.65, Z = −2.30, P = .021). No other significant control-related effects were found (Ps > .108). Given that the illusion of control was observed only for the winning probability of 75%, we focused our RewP analyses on this condition. We found a significant interaction between approval and valence (B = −1.49, P = .009). Post hoc comparisons (Fig. 6a) revealed that accepted choices elicited a less positive RewP than vetoed choices for gain feedback (B = −0.65, Z = −2.26, P = .024). In contrast, accepted choices elicited a more positive RewP than vetoed choices for loss feedback at a trend level (B = 0.84, Z = 1.70, P = .089). Thus, echoing the behavioral illusion of control, we found a diminished positive prediction error signal resulting from the illusion of control during the RewP period.

Figure 6.

Figure 6.

Open in a new tab

Predicted amplitude values of the RewP (a) and P300 (b) in response to gain and loss feedback obtained following accepted and vetoed choices during the condition with a high (75%) probability of winning.

Note: Participants during this condition showed an illusion of control. Error bars represent standard errors of the predicted values.

With respect to the P300 (Fig. 5c), we observed a significant interaction between valence and probability (linear: B = −0.81, P = .018) such that the P300 was larger for gains than for losses when the winning probability was 25% (B = 0.98, Z = 2.49, P = .013) but not when it was 50% (B = 0.64, Z = 1.80, P = .072) or 75% (B = −0.16, Z = −0.42, P = .677). Similar to the RewP, we observed a significant three-way interaction among approval, probability, and valence (quadratic: B = −1.66, P = .007). Follow-up pairwise comparisons indicated that when choices were accepted, the amplitude of the P300 was larger for gains than for losses under the winning probabilities of 25% (B = 1.19, Z = 2.28, P = .023) or 50% (B = 1.41, Z = 3.05, P = .002), but not under the winning probability of 75% (B = −0.87, Z = −1.67, P = .096). In contrast, when choices were vetoed, P300 amplitude was comparable between gains and losses for each of the three winning probabilities (25%: B = 0.77, Z = 1.48, P = .139; 50%: B = −0.13, Z = −0.29, P = .772; 75%: B = 0.54, Z = 1.03, P = .302). Moreover, gains from accepted choices elicited a larger P300 than gains from vetoed choices when the winning probability was 50% (B = 0.86, Z = 2.04, P = .041). No other significant control-related effects were found (Ps > .079). When focusing on the 75% probability trials, we obtained a significant interaction between approval and valence (B = −1.42, P = .039). However, post hoc comparisons (Fig. 6b) revealed comparable P300 amplitudes between accepted and vetoed choice trials regardless of whether the feedback was positive (B = −0.62, Z = −1.62, P = .105) or negative (B = 0.80, Z = 1.30, P = .192).

Together, our ERP results indicate that accepting versus vetoing a choice increases neural activity during the anticipatory phase. In the consummatory phase, accepting the choice of a gambling wheel with a high probability (75%) of winning decreases the RewP in response to gains. On the other hand, accepting the choice of a gambling wheel with a low (25%) or medium (50%) probability of winning increases the P300 amplitude differences between gains and losses.

General discussion

In this study, we investigated the neural dynamics underlying the illusion of control, which is operatively defined as an inflated estimate of reward probability than objective probability warrants. At the behavioral level, we replicated the classic illusion-of-control effect. Participants reported more confidence in achieving favorable outcomes when they chose a gambling wheel themselves than when they were assigned with a wheel randomly, all other things being equal. Importantly, this illusion-of-control effect was captured by reward-relevant ERP signatures during both the anticipatory phase (indexed by the SPN) and the consummatory phase (indexed by the RewP and the P300).

Our behavioral results are consistent with the preexisting false belief framework (Klusowski et al. 2021). According to this framework, participants may choose options in line with their belief that certain gambling wheels are more likely to win than others. Compared to the classical Langer experiment wherein choices are self-determined versus externally imposed (Langer 1975), our task represents a more extreme version of the illusion of control. In the externally imposed condition, the chosen option was vetoed and replaced. Thus, it is possible that the preexisting false belief was suppressed when the preferred option was vetoed. In support, our post-experimental rating data revealed that participants preferred and were more interested in accepted versus vetoed trials. However, we observed no reliable differences in perceived control between the accepted and vetoed conditions in both experiments (Cohen’s d: < 0.1 for Experiment 1 and 0.37 for Experiment 2), denying the possibility that the illusion of control engenders a higher level of control belief. This could be because participants were explicitly and truthfully informed that all three gambling wheels had an equal chance of winning or losing. Our design has also controlled for the confounding influence of effort investment in the illusion of control. In previous studies, decision effort is usually higher during self-determined decisions than externally imposed decisions, as supported by the large differences in reaction times between the two conditions (e.g. Masaki et al. 2010, Zheng et al. 2020b, Zhang et al. 2021). In our study, however, participants were required to make a decision and were unaware of whether their choices were ultimately accepted or vetoed.

In contrast to the previous study observing an illusion of control irrespective of outcome probability (Kool et al. 2013), we found that the illusion of control was influenced by winning probability. This illusion was observed when the winning probability was high (i.e. a 75% chance of winning), regardless of whether rating data were collected using a continuous scale during the task (Experiment 1) or a 9-point Likert scale after the task (Experiment 2). This discrepancy between our study and the Kool et al. study could be due to subtle methodological differences. In the Kool et al. study, outcome probability was not only explicitly communicated by the size of the win sector but also reflected a compressed probability space to collect more data for low probability outcomes. This made participants’ ratings rely on both description and experiential sampling. In our study, however, the proportions of wins and losses followed exactly the probabilities indicated by the win sector size, resulting in a description-based rating. Indeed, previous work on decision-making has highlighted a robust description-experience distinction in risky choices (Hertwig and Erev 2009).

During the anticipatory phase, the SPN was more negative when anticipating outcomes with higher reward probabilities in both accepted and vetoed conditions. If the illusion of control inflates reward probabilities, we would expect more negative SPN amplitudes in the accepted condition than the vetoed condition. However, the opposite result was obtained. Our finding that trials with accepted choices have smaller SPN values than those with vetoed choices would not be predicted by the illusion of control. Indeed, the SPN result contrasts with our behavioral evidence of the illusion of control, where participants reported greater confidence in achieving rewards when their choices were accepted rather than vetoed. A possible explanation for this unexpected SPN finding is that higher confidence in winning the gambles results in a decrease in the informative value of stimulus feedback, thereby decreasing the amplitude of the SPN following accepted compared to vetoed choices. This explanation aligns with previous research showing that the SPN is sensitive to changes in the information provided by feedback stimuli (Moris et al. 2013, Hirao et al. 2017). Alternatively, participants might have more negative affective-motivational responses when their choices were vetoed, as the SPN has been associated with greater avoidance of negative outcomes including probable nongains (Umemoto and Holroyd 2017), aversive noises (Kotani et al. 2001), and painful stimuli (Seidel et al. 2015). However, this explanation contrasts with our finding that the SPN carefully tracks explicit reward probabilities. Together, our SPN results suggest that the illusion of control may not work on subjective representations of reward probabilities during anticipation.

As expected, the RewP had a larger amplitude for gains compared to losses (Proudfit 2015). Moreover, this valence effect was enhanced for the vetoed condition than for the accepted condition only when the probability of winning was high (i.e. 75%). The RewP has long been established as a neural signature for prediction errors, reflecting the difference between the predicted value and the received value (Holroyd and Coles 2002, Walsh and Anderson 2012). Our RewP findings can be attributed to prediction errors induced by the illusion of control. Under the illusion of control, estimates of reward probabilities tend to be higher. As a result, prediction errors should be lower for positive outcomes (i.e. less-than-expected) and higher for negative outcomes (i.e. more-than-expected). Supporting this view, the RewP was reduced for gain feedback after accepted choices compared to gain feedback after vetoed choices. For loss feedback, the RewP was enhanced for the accepted condition than the vetoed condition at a trend level, which is consistent with the unsigned salience prediction error model of the RewP (Oliveira et al. 2007, Talmi et al. 2013, Hoy et al. 2021). This finding aligns with a recent view that the RewP tracks prediction error specifically on positive outcomes (Holroyd et al. 2008, Becker et al. 2014, Proudfit 2015). Interestingly, the RewP finding was present only when the probability of winning was high, echoing our behavioral finding that the illusion of control was observed for the high winning probability. Supporting the reinforcement learning theory (Holroyd and Coles 2002), previous studies have established that the RewP varies in proportion to the size of prediction errors, which is determined by feedback properties including outcome valence, magnitude, and probability (Nieuwenhuis et al. 2004, Walsh and Anderson 2012, Sambrook and Goslin 2015). Our study extends these results to demonstrate the sensitivity of the RewP to prediction errors by distorting outcome probability estimates through the illusion of control.

Taken together, our behavioral data indicate a successful manipulation of the illusion of control, showing an inflated estimate of reward probability than objective probability warrants. However, our ERP data only provide partial support for this behavioral effect. Specifically, our RewP results seem to reflect inflated reward probabilities during the illusion of control when feedback was a gain. However, this is refuted by our SPN results with its amplitudes being less negative under the illusion of control. These findings suggest that the behavioral illusion of control maybe not actually work on the subjective representation of reward probabilities, echoing a previous fMRI study that used a similar task (Kool et al. 2013). However, it should be pointed out that this possibility awaits more evidence. First, the illusion of control can take various forms (Nordgren et al. 2007, Filippin and Crosetto 2016). For instance, participants can be allowed to choose freely (control over choice) or to exert an illusory influence over the outcome (control over outcome). A previous study found that these two types of control affect risk perception differently such that control over choice increases perceived risk while control over outcome decreases perceived risk (Nordgren et al. 2007), suggesting that they rely on at least partially distinct cognitive mechanisms. In the current task, the illusion of control is more related to choices because participants are informed that all wheels had an equal chance of winning or losing, whether their choices were accepted or vetoed. Second, our task is based on description sampling instead of experiential sampling. Given the known difference in decision-making based on description versus experience (Hertwig and Erev 2009), our results might differ in tasks that rely more on experiential sampling such as reinforcement learning tasks (Cockburn et al. 2014).

Finally, although the P300 was not the focus of this research, it relates in interesting ways to the illusion of control. We observed a larger P300 for gains after accepted choices than gain after vetoed choices in uncertain scenarios with the medium winning probability. The control effect for gains, instead of losses, suggests that task relevance, or gain feedback as the natural target (Duncan-Johnson and Donchin 1977), may be enhanced during the illusion of control. Given that the P300 is associated with attention allocation based on motivational salience (Nieuwenhuis et al. 2005), it is possible that accepting choices enhance motivational salience during outcome evaluation (Wang et al. 2021). This explanation aligns with the view that the freedom to choose carries intrinsic value (Leotti et al. 2010), inflates the subjective value of rewards (Wang and Delgado 2019), and covaries with positive affect (Stolz et al. 2020) but extends it to demonstrate the enhanced affective and motivation salience by control beliefs in scenarios with the medium winning probability. Using a similar paradigm (Jiwa et al. 2021), Jiwa and colleagues found that participants were more inclined to place higher bids to learn the outcome of the roulette wheel that was chosen themselves than that was assigned randomly, especially when the outcome was uncertain (i.e. the 40% and 60% chances of winning compared to 20% and 80%). These results indicate that choosing a lottery increases the subjective value of resolving uncertainty about the outcome. Echoing and extending this study, our P300 findings suggest that resolving a roulette wheel with high uncertainty (i.e. a 50% chance of winning) gives rise to greater motivational salience when participants have an illusory control over the wheel.

In conclusion, this study found that participants assigned a higher value to the gambling wheels they selected themselves than to those assigned randomly, indicating a behavioral bias associated with the illusion of control. This bias is associated with reduced reward anticipation as indexed by the SPN, diminished positive prediction error signals as reflected by the RewP, and enhanced motivational salience as revealed by the P300. Considering that an illusory control can distort people’s judgements and is thus contributed to persistent gambling and financial loss (Yarritu et al. 2015, Clark and Wohl 2022), an interesting avenue for future research is to determine its clinical relevance by highlighting the anticipatory and consummatory reward processing in the illusion of control.

Supplementary Material

nsae063_Supp
nsae063_supp.zip (12.4KB, zip)

Contributor Information

Ya Zheng, Department of Psychology, Guangzhou University, Guangzhou 510006, China; Center for Reward and Social Cognition, School of Education, Guangzhou University, Guangzhou 510006, China.

Canming Yang, Department of Psychology, Dalian Medical University, Dalian 116044, China.

Huiping Jiang, Faculty of Psychology, Beijing Key Laboratory of Applied Experimental Psychology, National Demonstration Center for Experimental Psychology Education, Beijing Normal University, Beijing 100875, China.

Bo Gao, Center for Reward and Social Cognition, School of Education, Guangzhou University, Guangzhou 510006, China; Department of Psychology, Dalian Medical University, Dalian 116044, China.

Supplementary data

Supplementary data is available at SCAN online.

Conflict of interest

None declared.

Funding

This work was supported by the Research Fund from Guangzhou University (YJ2023039), the Guangdong Philosophy and Social Science Planning Project (GD24CXL06), and the National Natural Science Foundation of China (31971027).

Data availability

Data and code that support the findings of this study are available on Open Science Framework at https://osf.io/394gc/.

References

  1. Bar-Hillel M, Neter E. Why are people reluctant to exchange lottery tickets? J Pers Soc Psychol 1996;70:17–27. doi: 10.1037/0022-3514.70.1.17 [DOI] [Google Scholar]
  2. Bates D, Kliegl R, Vasishth S. et al. Parsimonious mixed models. 2015a. arXiv preprint arXiv:1506.04967. [Google Scholar]
  3. Bates D, Mächler M, Bolker B. et al. Fitting linear mixed-effects models using lme4ʹ. J Stat Soft 2015b;67:1–48. doi: 10.18637/jss.v067.i01 [DOI] [Google Scholar]
  4. Becker MP, Nitsch AM, Miltner WH. et al. A single-trial estimation of the feedback-related negativity and its relation to BOLD responses in a time-estimation task. J Neurosci 2014;34:3005–12. doi: 10.1523/JNEUROSCI.3684-13.2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Berridge KC, Robinson TE. Parsing reward. Trends Neurosci 2003;26:507–13. doi: 10.1016/S0166-2236(03)00233-9 [DOI] [PubMed] [Google Scholar]
  6. Bikute K, Di Bernardi Luft C, Beyer F. The value of an action: impact of motor behaviour on outcome processing and stimulus preference. Eur J Neurosci 2022;56:5823–35. doi: 10.1111/ejn.15826 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bogdanov M, Renault H, LoParco S. et al. Cognitive effort exertion enhances electrophysiological responses to rewarding outcomes. Cereb Cortex 2022;32:4255–70. doi: 10.1093/cercor/bhab480 [DOI] [PubMed] [Google Scholar]
  8. Brunia CH, Hackley SA, van Boxtel GJ. et al. Waiting to perceive: reward or punishment? Clin Neurophysiol 2011;122:858–68. doi: 10.1016/j.clinph.2010.12.039 [DOI] [PubMed] [Google Scholar]
  9. Chen W, Li Q, Mei S. et al. Diminished choice effect on anticipating improbable rewards. Neuropsychologia 2018;111:45–50. doi: 10.1016/j.neuropsychologia.2018.01.015 [DOI] [PubMed] [Google Scholar]
  10. Clark L, Wohl MJA. Langer’s illusion of control and the cognitive model of disordered gambling. Addiction 2022;117:1146–51. doi: 10.1111/add.15649 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cockburn J, Collins AG, Frank MJ. A reinforcement learning mechanism responsible for the valuation of free choice. Neuron 2014;83:551–57. doi: 10.1016/j.neuron.2014.06.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods 2004;134:9–21. doi: 10.1016/j.jneumeth.2003.10.009 [DOI] [PubMed] [Google Scholar]
  13. Duncan-Johnson CC, Donchin E. On quantifying surprise: the variation of event-related potentials with subjective probability. Psychophysiology 1977;14:456–67. doi: 10.1111/j.1469-8986.1977.tb01312.x [DOI] [PubMed] [Google Scholar]
  14. Filippin A, Crosetto P. Click‘n’Roll: no evidence of illusion of control. De Economist 2016;164:281–95. doi: 10.1007/s10645-016-9282-3 [DOI] [Google Scholar]
  15. Fuentemilla L, Cucurell D, Marco-Pallarés J. et al. Electrophysiological correlates of anticipating improbable but desired events. NeuroImage 2013;78:135–44. doi: 10.1016/j.neuroimage.2013.03.062 [DOI] [PubMed] [Google Scholar]
  16. Hackley SA, Hirao T, Onoda K. et al. Anterior insula activity and the effect of agency on the stimulus-preceding negativity. Psychophysiology 2020;57:e13519. doi: 10.1111/psyp.13519 [DOI] [PubMed] [Google Scholar]
  17. Hackley SA, Valle-Inclán F, Masaki H et al. Stimulus-Preceding Negativity (SPN) and attention to rewards. In: Mangun GR (ed.), Cognitive Electrophysiology of Attention. San Diego: Academic Press, 2014, 216–25. [Google Scholar]
  18. Hassall CD, Hajcak G, Krigolson OE. The importance of agency in human reward processing. Cogn Affect Behav Neurosci 2019;19:1458–66. doi: 10.3758/s13415-019-00730-2 [DOI] [PubMed] [Google Scholar]
  19. Heise MJ, Mon SK, Bowman LC. Utility of linear mixed effects models for event-related potential research with infants and children. Dev Cogn Neurosci 2022;54:101070. doi: 10.1016/j.dcn.2022.101070 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hertwig R, Erev I. The description-experience gap in risky choice. Trends Cognit Sci 2009;13:517–23. doi: 10.1016/j.tics.2009.09.004 [DOI] [PubMed] [Google Scholar]
  21. Heydari S, Holroyd CB. Reward positivity: reward prediction error or salience prediction error? Psychophysiology 2016;53:1185–92. doi: 10.1111/psyp.12673 [DOI] [PubMed] [Google Scholar]
  22. Hirao T, Murphy T, Masaki H. Brain activities associated with learning of the Monty Hall Dilemma task. Psychophysiology 2017;54:1359–69. doi: 10.1111/psyp.12883 [DOI] [PubMed] [Google Scholar]
  23. Holroyd CB, Coles MGH. The neural basis of human error processing: reinforcement learning, dopamine, and the error-related negativity. Psychol Rev 2002;109:679–709. doi: 10.1037/0033-295X.109.4.679 [DOI] [PubMed] [Google Scholar]
  24. Holroyd CB, Nieuwenhuis S, Yeung N. et al. Errors in reward prediction are reflected in the event-related brain potential. NeuroReport 2003;14:2481–84. doi: 10.1097/00001756-200312190-00037 [DOI] [PubMed] [Google Scholar]
  25. Holroyd CB, Pakzad-Vaezi KL, Krigolson OE. The feedback correct-related positivity: sensitivity of the event-related brain potential to unexpected positive feedback. Psychophysiology 2008;45:688–97. doi: 10.1111/j.1469-8986.2008.00668.x [DOI] [PubMed] [Google Scholar]
  26. Hoy CW, Steiner SC, Knight RT. Single-trial modeling separates multiple overlapping prediction errors during reward processing in human EEG. Commun Biol 2021;4:910. doi: 10.1038/s42003-021-02426-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jiwa M, Cooper PS, Chong TT. et al. Choosing increases the value of non-instrumental information. Sci Rep 2021;11:8780. doi: 10.1038/s41598-021-88031-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Klusowski J, Small DA, Simmons JP. Does choice cause an illusion of control? Psychol Sci 2021;32:159–72. doi: 10.1177/0956797620958009 [DOI] [PubMed] [Google Scholar]
  29. Kool W, Getz S, Botvinick M. Neural representation of reward probability: evidence from the illusion of control. J Cognitive Neurosci 2013;25:852–61. doi: 10.1162/jocn_a_00369 [DOI] [PubMed] [Google Scholar]
  30. Kotani Y, Hiraku S, Suda K. et al. Effect of positive and negative emotion on stimulus-preceding negativity prior to feedback stimuli. Psychophysiology 2001;38:873–78. doi: 10.1111/1469-8986.3860873 [DOI] [PubMed] [Google Scholar]
  31. Kuznetsova A, Brockhoff PB, Christensen RHB. lmerTest package: tests in linear mixed effects models. J Stat Soft 2017;82:1–26. doi: 10.18637/jss.v082.i13 [DOI] [Google Scholar]
  32. Langer EJ. The illusion of control. J Pers Soc Psychol 1975;32:311–28. doi: 10.1037/0022-3514.32.2.311 [DOI] [Google Scholar]
  33. Lenth RV. Emmeans: estimated marginal means, aka least-squares means. 2022.
  34. Leotti LA, Iyengar SS, Ochsner KN. Born to choose: the origins and value of the need for control. Trends Cognit Sci 2010;14:457–63. doi: 10.1016/j.tics.2010.08.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lorenz RC, Gleich T, Kühn S. et al. Subjective illusion of control modulates striatal reward anticipation in adolescence. NeuroImage 2015;117:250–57. doi: 10.1016/j.neuroimage.2015.05.024 [DOI] [PubMed] [Google Scholar]
  36. Luck SJ, Gaspelin N. How to get statistically significant effects in any ERP experiment (and why you shouldn’t). Psychophysiology 2017;54:146–57. doi: 10.1111/psyp.12639 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Ma Q, Meng L, Wang L. et al. I endeavor to make it: effort increases valuation of subsequent monetary reward. Behav Brain Res 2014;261:1–7. doi: 10.1016/j.bbr.2013.11.045 [DOI] [PubMed] [Google Scholar]
  38. Masaki H, Yamazaki K, Hackley SA. Stimulus-preceding negativity is modulated by action-outcome contingency. NeuroReport 2010;21:277–81. doi: 10.1097/WNR.0b013e3283360bc3 [DOI] [PubMed] [Google Scholar]
  39. Matuschek H, Kliegl R, Vasishth S. et al. Balancing type I error and power in linear mixed models. J Memory Lang 2017;94:305–15. doi: 10.1016/j.jml.2017.01.001 [DOI] [Google Scholar]
  40. Mei S, Yi W, Zhou S. et al. Contextual valence modulates the effect of choice on incentive processing. Soc Cognit Affective Neurosci 2018;13:1249–58. doi: 10.1093/scan/nsy098 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Meyer GM, Marco-Pallares J, Sescousse G. et al. Electrophysiological underpinnings of reward processing: are we exploiting the full potential of EEG? NeuroImage 2021;242:118478. doi: 10.1016/j.neuroimage.2021.118478 [DOI] [PubMed] [Google Scholar]
  42. Moris J, Luque D, Rodriguez-Fornells A. Learning-induced modulations of the stimulus-preceding negativity. Psychophysiology 2013;50:931–39. doi: 10.1111/psyp.12073 [DOI] [PubMed] [Google Scholar]
  43. Mühlberger C, Angus DJ, Jonas E. et al. Perceived control increases the reward positivity and stimulus preceding negativity. Psychophysiology 2017;54:310–22. doi: 10.1111/psyp.12786 [DOI] [PubMed] [Google Scholar]
  44. Mulligan EM, Hajcak G. The electrocortical response to rewarding and aversive feedback: the reward positivity does not reflect salience in simple gambling tasks. Int J Psychophysiol 2018;132:262–67. doi: 10.1016/j.ijpsycho.2017.11.015 [DOI] [PubMed] [Google Scholar]
  45. Nieuwenhuis S, Aston-Jones G, Cohen JD. Decision making, the P3, and the locus coeruleus-norepinephrine system. Psychol Bull 2005;131:510–32. doi: 10.1037/0033-2909.131.4.510 [DOI] [PubMed] [Google Scholar]
  46. Nieuwenhuis S, Holroyd CB, Mol N. et al. Reinforcement-related brain potentials from medial frontal cortex: origins and functional significance. Neurosci Biobehav Rev 2004;28:441–48. doi: 10.1016/j.neubiorev.2004.05.003 [DOI] [PubMed] [Google Scholar]
  47. Nolan H, Whelan R, Reilly RB. FASTER: fully automated statistical thresholding for EEG artifact rejection. J Neurosci Methods 2010;192:152–62. doi: 10.1016/j.jneumeth.2010.07.015 [DOI] [PubMed] [Google Scholar]
  48. Nordgren LF, van der Pligt J, van Harreveld F. Unpacking perceived control in risk perception: the mediating role of anticipated regret. J Behav Decis Making 2007;20:533–44. doi: 10.1002/bdm.565 [DOI] [Google Scholar]
  49. Oliveira FTP, Mcdonald JJ, David G. Performance monitoring in the anterior cingulate is not all error related: expectancy deviation and the representation of action-outcome associations. J Cognitive Neurosci 2007;19:1994–2004. doi: 10.1162/jocn.2007.19.12.1994 [DOI] [PubMed] [Google Scholar]
  50. Patall EA, Cooper H, Robinson JC. The effects of choice on intrinsic motivation and related outcomes: a meta-analysis of research findings. Psychol Bull 2008;134:270–300. doi: 10.1037/0033-2909.134.2.270 [DOI] [PubMed] [Google Scholar]
  51. Proudfit GH. The reward positivity: from basic research on reward to a biomarker for depression. Psychophysiology 2015;52:449–59. doi: 10.1111/psyp.12370 [DOI] [PubMed] [Google Scholar]
  52. Sambrook TD, Goslin J. A neural reward prediction error revealed by a meta-analysis of ERPs using great grand averages. Psychol Bull 2015;141:213–35. doi: 10.1037/bul0000006 [DOI] [PubMed] [Google Scholar]
  53. San Martín R, Appelbaum LG, Pearson JM. et al. Rapid brain responses independently predict gain maximization and loss minimization during economic decision making. J Neurosci 2013;33:7011–19. doi: 10.1523/JNEUROSCI.4242-12.2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Schultz W, Dayan P, Montague PR. A neural substrate of prediction and reward. Science 1997;275:1593–99. doi: 10.1126/science.275.5306.1593 [DOI] [PubMed] [Google Scholar]
  55. Seidel EM, Pfabigan DM, Hahn A. et al. Uncertainty during pain anticipation: the adaptive value of preparatory processes. Human Brain Mapp 2015;36:744–55. doi: 10.1002/hbm.22661 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Stolz DS, Muller-Pinzler L, Krach S. et al. Internal control beliefs shape positive affect and associated neural dynamics during outcome valuation. Nat Commun 2020;11:1230. doi: 10.1038/s41467-020-14800-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Sutton S, Tueting P, Hammer M et al. Evoked potentials and feedback. In: Otto DA (ed.), Multidisciplinary Perspectives in Event-Related Potential Research. Washington, DC: U.S. Government Printing Office, 1978, 184–88. [Google Scholar]
  58. Talmi D, Atkinson R, El-Deredy W. The feedback-related negativity signals salience prediction errors, not reward prediction errors. J Neurosci 2013;33:8264–69. doi: 10.1523/JNEUROSCI.5695-12.2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Tricomi EM, Delgado MR, Fiez JA. Modulation of caudate activity by action contingency. Neuron 2004;41:281–92. doi: 10.1016/S0896-6273(03)00848-1 [DOI] [PubMed] [Google Scholar]
  60. Umemoto A, Holroyd CB. Neural mechanisms of reward processing associated with depression-related personality traits. Clin Neurophysiol 2017;128:1184–96. doi: 10.1016/j.clinph.2017.03.049 [DOI] [PubMed] [Google Scholar]
  61. Walsh MM, Anderson JR. Learning from experience: event-related potential correlates of reward processing, neural adaptation, and behavioral choice. Neurosci Biobehav Rev 2012;36:1870–84. doi: 10.1016/j.neubiorev.2012.05.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Wang KS, Delgado MR. Corticostriatal circuits encode the subjective value of perceived control. Cereb Cortex 2019;29:5049–60. doi: 10.1093/cercor/bhz045 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Wang KS, Yang -Y-Y, Delgado MR. How perception of control shapes decision making. Curr Opin Behav Sci 2021;41:85–91. doi: 10.1016/j.cobeha.2021.04.003 [DOI] [Google Scholar]
  64. Wu M, Zheng Y. Physical effort paradox during reward evaluation and links to perceived control. Cereb Cortex 2023;33:9343–53. doi: 10.1093/cercor/bhad207 [DOI] [PubMed] [Google Scholar]
  65. Yarritu I, Matute H, Luque D. The dark side of cognitive illusions: when an illusory belief interferes with the acquisition of evidence-based knowledge. Br J Psychol 2015;106:597–608. doi: 10.1111/bjop.12119 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Yeung N, Holroyd CB, Cohen JD. ERP correlates of feedback and reward processing in the presence and absence of response choice. Cereb Cortex 2005;15:535–44. doi: 10.1093/cercor/bhh153 [DOI] [PubMed] [Google Scholar]
  67. Yi W, Mei S, Li Q. et al. How choice influences risk processing: an ERP study. Biol Psychol 2018;138:223–30. doi: 10.1016/j.biopsycho.2018.08.011 [DOI] [PubMed] [Google Scholar]
  68. Yi W, Mei S, Zhang M. et al. Decomposing the effort paradox in reward processing: time matters. Neuropsychologia 2020;137:107311. doi: 10.1016/j.neuropsychologia.2019.107311 [DOI] [PubMed] [Google Scholar]
  69. Zhang L, Qi G, Long C. The choice levels modulate outcome processing during outcome independent of behavior selection: evidence from event-related potentials. Int J Psychophysiol 2021;169:44–54. doi: 10.1016/j.ijpsycho.2021.08.007 [DOI] [PubMed] [Google Scholar]
  70. Zheng Y, An T, Li Q. et al. Distinct electrophysiological correlates between expected reward and risk processing. Psychophysiology 2020a;57:e13638. doi: 10.1111/psyp.13638 [DOI] [PubMed] [Google Scholar]
  71. Zheng Y, Mei S. Neural dissociation between reward and salience prediction errors through the lens of optimistic bias. Human Brain Mapp 2023;44:4545–60. doi: 10.1002/hbm.26398 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Zheng Y, Wang M, Zhou S. et al. Functional heterogeneity of perceived control in feedback processing. Soc Cognit Affective Neurosci 2020b;15:329–36. doi: 10.1093/scan/nsaa028 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

nsae063_Supp
nsae063_supp.zip (12.4KB, zip)

Data Availability Statement

Data and code that support the findings of this study are available on Open Science Framework at https://osf.io/394gc/.

Articles from Social Cognitive and Affective Neuroscience are provided here courtesy of Oxford University Press

RESOURCES