Research Article

Neuroscience

Human cerebellum and ventral tegmental area interact during extinction of learned fear

Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Germany
Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Germany
Institute for Neural Computation, Faculty of Computer Science, Ruhr University Bochum, Germany
Institute for Diagnostic and Interventional Radiology and Neuroradiology, Essen University Hospital, University of Duisburg-Essen, Germany
Institute for Diagnostic and Interventional Radiology, Neuroradiology and Nuclear Medicine, University Hospital Knappschaftskrankenhaus Bochum, Ruhr University Bochum, Germany
Department of Psychology, Philipps-University Marburg, Germany
Department of Cognitive Psychology, Institute of Cognitive Neuroscience, Ruhr University Bochum, Germany
High-Field and Hybrid MR Imaging, Essen University Hospital, University of Duisburg-Essen, Germany

Jul 13, 2026

https://doi.org/10.7554/eLife.105399.3

Open access
Copyright information

eLife Assessment

This important study provides insights into the role of the cerebellum in fear conditioning, addressing a key gap in the literature. The evidence presented in support of the conclusions is solid. This work will be of interest to both the extinction learning and cerebellar research communities.

https://doi.org/10.7554/eLife.105399.3.sa0

Significance of the findings:

Important: Findings that have theoretical or practical implications beyond a single subfield

Landmark
Fundamental
Important
Valuable
Useful

Strength of evidence:

Solid: Methods, data and analyses broadly support the claims with only minor weaknesses

Exceptional
Compelling
Convincing
Solid
Incomplete
Inadequate

During the peer-review process the editor and reviewers write an eLife Assessment that summarises the significance of the findings reported in the article (on a scale ranging from landmark to useful) and the strength of the evidence (on a scale ranging from exceptional to inadequate). Learn more about eLife Assessments

Abstract
Introduction
Results
Discussion
Methods
Appendix 1
Data availability
References
Article and author information
Metrics

Abstract

The key elements for fear extinction learning are unexpected omissions of expected aversive events, which are considered to be rewarding. Given its reception of reward information, we tested the hypothesis that the cerebellum contributes to reward-like prediction error processing driving extinction learning via its connections with the ventral tegmental area (VTA). Forty-three young and healthy participants performed a three-day fear conditioning paradigm in a 7T MR scanner. The cerebellum and VTA were active during unexpected omissions of aversive unconditioned stimuli in the initial extinction trials and in other learning phases, in line with the proposed role of prediction-error processing. Increased functional connectivity was observed between the cerebellum and VTA, indicating that they are functionally coupled during fear extinction learning. These results suggest that an interaction between the cerebellum and VTA should be incorporated into the existing model of the fear extinction network.

Introduction

Deficits in learning to extinguish previously associated threat responses to cues that no longer signal danger are thought to be one of the main causes in the development of anxiety disorders, including post-traumatic stress disorder and social anxiety disorder (Grupe and Nitschke, 2013). Exposure therapy addresses this deficit and attempts to compensate for failed or absent extinction learning from the past (Papalini et al., 2020). This is modeled effectively with extinction training in classical fear conditioning paradigms where the previously paired conditioned stimulus (CS+) is no longer followed by the aversive unconditioned stimulus (US) (Milad and Quirk, 2012; Lonsdorf et al., 2017). At the beginning of extinction training the omission of the US is unexpected and this unexpected lack of the US is considered to be rewarding (Luo et al., 2018; Salinas-Hernández et al., 2018; Kalisch et al., 2019). This reward-like prediction error is thought to drive safety learning. This newly learned safety association should then inhibit the initially learned fear association (which is never fully lost) (Craske et al., 2012).

The reward system in the brain is primarily associated with the mesolimbic dopamine system (Schultz and Dickinson, 2000), i.e., the ventral tegmental area (VTA) and the ventral striatum. In fact, recent rodent studies provide strong evidence that the VTA is central for extinction learning, and that (reward-like) prediction errors are encoded in a subset of VTA neurons (Luo et al., 2018; Salinas-Hernández et al., 2018; Kalisch et al., 2019; Salinas-Hernández et al., 2023).

Importantly, new evidence indicates that the cerebellum receives reward signals (Wagner et al., 2017; Heffley and Hull, 2019; Kostadinov et al., 2019), and has direct efferent connections with the mesolimbic dopaminergic system, in particular the VTA (Snider and Maiti, 1976; Watabe-Uchida et al., 2012; Carta et al., 2019; Pisano et al., 2021).

Functional magnetic resonance imaging (fMRI) findings in healthy human participants suggest that the cerebellum is involved in processing of prediction errors in fear conditioning (Ernst et al., 2019; Ploghaus et al., 1999). Ernst et al., 2019 studied acquisition of learned fear responses using a partial reinforcement rate, i.e., a subset of CS+ trials were not reinforced in acquisition training. Cerebellar activations were strongest in the unreinforced CS+ trials, when the US was expected but did not occur. An unreinforced CS+ trial can be considered as a very first extinction trial, and cerebellar activations may reflect prediction errors driving extinction learning. Likewise, recordings from the VTA in rodents showed strong activations at the time the US is expected and does not occur during initial extinction trials (Salinas-Hernández et al., 2018; Salinas-Hernández et al., 2023). The aim of the present fMRI study is, therefore, to test the hypothesis that the cerebellum contributes to fear extinction learning via its connection to the VTA.

In contrast to the previous one-day study design (Ernst et al., 2019), here, we used a three-day design with acquisition training on day 1, extinction training on day 2 and a recall test on day 3 (Figure 1A). This multi-day study design allowed enough time for consolidation of both fear and extinction learning. Cerebellar activity related to the unexpected omission of the US was assessed in four different learning phases, that is, during extinction training (in initial CS+ trials), during the recall test (in initial CS+ trials), during reacquisition (in interspersed non-reinforced CS+ trials), and during reextinction (the initial CS+ trials immediately following reinforced trials) (Figure 1A). The CSs were geometric shapes (i.e. square and diamond), and the US was an electrical stimulation (Figure 1B). Full (100%) reinforcement was used during acquisition on day 1 to maximize the size of the reward-like prediction error elicited by the initial unexpected US omissions during extinction. Learning phases differed in the underlying learning state, ranging from newly acquired fear, to spontaneous recovery after consolidated extinction (Bouton, 2002), to reactivated fear and reextinction, providing four distinct contexts in which to examine cerebellar and VTA responses to unexpected US omission (Doubliez et al., 2025).

Figure 1

Download asset Open asset

Overview of the three-day fear conditioning paradigm and experimental setup.

(A) Day 1 consisted of habituation and acquisition training, day 2 of extinction training, and day 3 of a recall test followed by reacquisition, reextinction, and the unexpected US phase. Extinction, recall, reacquisition, and reextinction are highlighted in purple, indicating phases in which unexpected US omissions occurred. Extinction, recall, and reextinction are shown with a gradient to illustrate the gradual decrease in unexpectedness of US omissions across trials, in contrast to the discrete omission events during partially reinforced reacquisition. (B) Schematic of the experimental setup during 7T MRI. Participants lay supine in the scanner and viewed visual conditioned stimuli (CS; geometric shapes) displayed on a screen at the end of the bore via a mirror system. The eyes were simultaneously imaged with an eye-tracking system. Electrodermal activity (EDA) was recorded from the left hand (blue/purple cables). The unconditioned stimulus (US; electrical stimulation) was delivered to the right hand (red cable), and responses were recorded using a button box.

We used a reinforcement-learning-based deep learning model fitted to group-averaged trial-by-trial skin conductance response (SCR) data to compute per-trial prediction errors. SCRs reflect autonomic arousal linked to stimulus salience and expectancy. Averaged SCRs were used here to capture predictive learning about the CS-US association at the population level, allowing us to derive model-based prediction error values. 7T fMRI allowed us to assess VTA activation.

Results showed activation in the cerebellum and VTA for all four phases related to unexpected US omissions, despite differences in learning contexts between phases. Results also showed significant functional interaction between the cerebellum and VTA related to prediction errors in unexpected omission trials. These findings suggest that the cerebellum contributes to extinction learning via the VTA, and thus supports safety learning inhibiting the initial fear association.

Results

We first report behavioral results demonstrating that participants acquired and partially extinguished conditioned fear in the three-day fear conditioning paradigm (Figure 1). We then present fMRI results related to the prediction, presentation, and unexpected omission of the US in the cerebellar cortex (CB), deep cerebellar nuclei (DCN), and VTA. Finally, we report analyses of functional connectivity between the cerebellum and the VTA.

Behavioral data

Behavioral responses were assessed using SCRs (Figure 2), pupil size responses (PSRs; Figure 3), and self-reported ratings. Results for all three measures are reported below, with non-parametric ANOVA-type statistics and descriptive summaries of self-reported ratings provided in the Supplementary Information (Supplementary files 1–6).

Skin conductance responses (SCRs)

During habituation (day 1), SCRs did not differ between CS+ and CS- (Figure 2, top row). SCRs for both stimuli decreased significantly across habituation trials. Non-parametric ANOVA-type statistics revealed a significant effect for Time (early vs. late; F₁=26.49; p<0.001) but not for Stimulus (CS+ vs. CS-; F₁=1.19; p=0.276) or the Stimulus × Time interaction (F₁=0.06; p=0.802). During fear acquisition training (day 1), participants underwent a fully reinforced, semi-instructed differential fear conditioning protocol. Participants quickly learned the associations, and their responses decreased as expected due to habituation effects. SCRs for the CS+ were significantly higher than for the CS-, and decreased in late trials. Non-parametric ANOVA-type statistics revealed a significant effect of Stimulus (CS+ vs. CS-; F₁=20.79; p<0.001) and Time (early vs. late; F₁=28.75; p<0.001). No significant Stimulus x Time interaction was observed (F₁=3.80; p=0.051).

During the recall test (day 3), SCRs for CS+ were significantly higher than for CS- (Figure 2, bottom row). In late recall, no difference between SCRs for CS+ and CS- was observed. Non-parametric ANOVA-type statistics revealed a significant main effect of Stimulus (CS+ vs. CS-; F₁=4.99; p=0.026), Time (early vs. late; F₁=36.74; p<0.001), and a Stimulus × Time interaction (F₁=5.53; p=0.019). Post hoc tests revealed that the CS+ had a significantly higher SCR than the CS- in early trials (p<0.001) but not in late trials (p=1.000). SCRs for both stimuli significantly decreased in late trials compared to early trials (both: p<0.001).

During reacquisition on day 3, SCRs for the CS+ were significantly higher than for the CS-, and decreased for both stimuli in late reacquisition. Non-parametric ANOVA-type statistics revealed a significant effect of Stimulus (CS+ vs. CS-; F₁=27.20, p<0.001), Time (early vs. late; F₁=35.44, p<0.001), and a Stimulus × Time interaction (F₁=6.80, p=0.009). Post hoc tests showed that both stimuli were higher in early vs late reacquisition training (CS- early vs. CS- late; p=0.007; CS+ early vs. CS+ late; p<0.001), and a significant difference between stimulus types during early trials (p<0.001), but no significant difference in late trials (p=0.094).

During reextinction on day 3, SCRs related to the CS+ were significantly higher than for the CS-, although the difference was small. Non-parametric ANOVA-type statistics revealed a significant main effect of Stimulus (CS+ vs. CS-; F₁=4.47; p=0.035), but neither for Time (early vs. late; F₁=0.08; p=0.780) nor for the Stimulus × Time interaction (F₁=0.95; p=0.329).

In the last phase on day 3, three CS+ trials were unexpectedly reinforced (‘unexpected US phase’). SCRs for the CS+ were significantly higher than for the CS-. Non-parametric ANOVA-type statistics revealed a significant main effect of Stimulus (CS+ vs. CS-; F₁=5.43; p=0.020), but neither for Time (early vs. late; F₁=0.29; p=0.593) nor for the Stimulus × Time interaction (F₁=2.34; p=0.126).

During extinction training (day 2), SCRs were higher for the CS+ than for the CS- only in initial extinction trials. Non-parametric ANOVA-type statistics revealed a significant effect of Stimulus (CS+ vs. CS-; F₁=7.71; p=0.006) and Time (early vs. late; F₁=21.10; p<0.001), whereas the Stimulus × Time interaction was not significant (F₁=3.74; p=0.053). Exploratory post hoc comparisons indicated higher SCRs to the CS+ than to the CS- during early extinction trials (p=0.017), but no differentiation during late extinction (p=0.798). These post hoc results are reported descriptively, as the Stimulus x Time interaction did not reach statistical significance.

Figure 2 with 1 supplement see all

Download asset Open asset

Skin conductance responses (SCRs) for each trial, with CS+ (shown in orange) and CS- (shown in blue) responses paired in blocks.

Reinforcement of the CS+ by a US (CS+/US) is indicated by filled squares. The CS- is never reinforced. Bar plots on the top right show mean SCRs for early and late trials, defined as the first and second halves of each phase, respectively. Significance markers indicate post hoc comparisons between CS+ and CS- within early or late trials; markers are shown in black when the Stimulus × Time interaction was significant and in red when the interaction was not significant. On day 1, there was no differentiation between CS+ and CS- in the habituation phase, with significant differentiation emerging during acquisition training. On day 2, differentiation was apparent in early extinction trials and was no longer evident later in extinction (trend-level effect; non-significant Stimulus × Time interaction). On day 3, during the initial recall test, participants exhibited renewed CS+/CS- differentiation, consistent with spontaneous recovery. During initial reacquisition, there were again differential responses to the CS+ and CS-, which appeared reduced in reextinction and the unexpected US phase. Mean values are shown with error bars representing the standard error of the mean. Data are shown for n=43 participants. Significance levels: *p<0.05, **p<0.01, ***p<0.001. CS: conditioned stimulus; US: unconditioned stimulus; SCR: skin conductance response; μS: microsiemens.

Pupil size responses (PSRs)

During habituation on day 1, PSRs related to the CS+ and CS- did not differ significantly (non-parametric ANOVA-type statistics; all p>0.199).

During fear acquisition training on day 1, PSRs were significantly higher in CS+ trials than in CS- trials, and PSR was more pronounced in early compared to late trials (Figure 3, top row). Non-parametric ANOVA-type statistics revealed a significant main effect of Stimulus (CS+ vs. CS-; F₁=39.51; p<0.001) and Time (early vs. late; F₁=24.31; p<0.001). No significant Stimulus × Time interaction was observed (F₁=0.58; p=0.446).

During extinction training on day 2, PSRs for the CS+ decreased from early to late extinction, but there were no differences between early and late trials for the CS-. Non-parametric ANOVA-type statistics revealed a significant main effect of Time (early vs. late; F₁=18.21; p<0.001) and a significant Stimulus × Time interaction (F₁=4.28; p=0.039). No significant main effect of Stimulus was observed (CS+ vs. CS-; F₁=1.12; p=0.290). Post hoc tests showed that PSRs related to the CS+ were significantly lower during late extinction compared to early (p<0.001). This result was not observed for PSRs related to the CS- (p=0.643).

During the recall test on day 3, PSRs for the CS+ were significantly higher than for the CS- during early, but not late, recall trials (Figure 3, bottom row). Non-parametric ANOVA-type statistics revealed a significant effect of Stimulus (CS+ vs. CS-; F₁=6.59; p=0.010) and Time (early vs. late; F₁=6.41; p=0.011) and a significant Stimulus x Time interaction (F₁=4.87; p=0.027). Post hoc tests showed higher PSRs for the CS+ than for the CS- during early recall (p=0.017), but not during late recall (p=0.862), and that the PSRs related to the late CS+ were significantly lower than early (p=0.003).

During reacquisition on day 3, PSRs for the CS+ decreased from early to late trials, whereas no change over time was observed for the CS-. PSRs differed between CS+ and CS- in both early and late trials. Non-parametric ANOVA-type statistics revealed significant effects of Stimulus (CS+ vs. CS-; F₁=47.87, p<0.001) and Time (early vs. late; F₁=4.54, p=0.033), as well as a significant Stimulus × Time interaction (F₁=7.70, p=0.006). Post hoc tests indicated that PSRs for the CS+ were higher early than late (p=0.004), whereas no such change was observed for the CS- (p=0.974).

During reextinction on day 3, PSRs for CS+ were significantly higher than for CS- early, but the difference became insignificant in late trials. Non-parametric ANOVA-type statistics revealed a significant main effect of Stimulus (CS+ vs. CS-; F₁=9.10; p=0.003) and a Stimulus × Time interaction (F₁=4.43; p=0.035), but not a significant main effect of Time (early vs. late; F₁=0.83, p=0.364). Post hoc tests showed a significant difference between the CS+ and CS- early but not late (early: p=0.007; late: p=0.775).

During the unexpected US phase on day 3, PSRs for the CS+ were higher than for the CS-. Non-parametric ANOVA-type analysis revealed a significant main effect of Stimulus (CS+ vs. CS-; F₁=17.35; p<0.001), but not of Time (early vs. late; F₁=0.83; p=0.363) and the Stimulus x Time interaction (F₁=0.03; p=0.872).

Figure 3 with 1 supplement see all

Download asset Open asset

Pupil size responses (PSRs) for each trial, with CS+ (shown in orange) and CS- (shown in blue) responses paired in blocks.

Reinforcement of the CS+ by a US (CS+/US) is indicated by filled squares. The CS- is never reinforced. Bar plots on the top right show mean PSRs for early and late trials, defined as the first and second halves of each phase, respectively. Significance markers indicate post hoc comparisons between CS+ and CS- within early or late trials; markers are shown in black when the Stimulus × Time interaction was significant and in red when the interaction was not significant. On day 1, there was no differentiation between CS+ and CS- in the habituation phase, with significant differentiation emerging during acquisition training. On day 2, a significant Stimulus × Time interaction was observed during extinction, reflecting a decrease from early to late extinction trials for the CS+ but not for the CS-, despite the absence of significant post hoc CS+/CS- differences. On day 3, during the initial recall test, participants exhibited renewed CS+/CS- differentiation, consistent with spontaneous recovery. During initial reacquisition, there were again differential responses to the CS+ and CS-, which appeared reduced in reextinction and the unexpected US phase. Mean values are shown with error bars representing the standard error of the mean. Data are shown for n=43 participants. Significance levels: *p<0.05, **p<0.01, ***p<0.001. CS: conditioned stimulus; US: unconditioned stimulus; PSR: pupil size response.

Self-reports

Valence, arousal, fear, and US expectancy ratings of the CS+ did not differ significantly from the CS- before acquisition training (Supplementary files 5 and 6). Following acquisition training, the CS+ was rated as more arousing, fearful, and unpleasant compared to the CS-, and participants expected the US to follow the CS+ but not the CS- (all p<0.001). Although less pronounced, the CS+/CS- differentiation persisted after extinction training and recall test (all p<0.004). However, extinction and recall CS+ ratings were significantly reduced compared to acquisition (all p≤0.001), whereas CS- ratings remained unchanged (all p>0.532), reflecting extinction learning. The difference between CS+ and CS- increased at the end of day 3, when CS+/US pairings were reintroduced.

Non-parametric ANOVA-type statistics showed a significant main effect of Stimulus, Time, and a Stimulus × Time interaction (main effects and interaction; all p<0.001; Supplementary file 6). Post hoc tests revealed significant differences between stimulus type (CS+, CS-) following acquisition training, extinction training, recall test and at the end of day 3 but not after habituation.

fMRI results

Analysis of BOLD fMRI responses revealed significant activations related to the prediction, the presentation and the unexpected omission of the US (Figure 4; Figure 5; Figure 6; Figure 7). Our primary focus was on activations and connectivity related to the unexpected omission of the US, which could be detected across four distinct phases: extinction training, recall test, reacquisition, and reextinction (Figure 6; Figure 7; Figure 8). Supplementary file 7 reports VOI-level summary statistics (mean contrast estimates, 95% confidence intervals, and Cohen’s d) for a CB VOI derived from the conjunction of unexpected US omission contrasts, as well as for the DCN, and VTA across all fMRI contrasts, while Supplementary files 8–12 report cluster-level activation and connectivity results.

Activations related to US prediction (CS+ > CS- and CS+ × prediction)

In the event-based analysis, there was no significantly higher activation in the cerebellum during the CS+ compared to the CS- (CS+ > CS-) during acquisition or extinction training. In the same contrast, the lateral VTA exerted significant differential activations during acquisition (Figure 4A; threshold-free cluster enhancement (TFCE) and family-wise error (FWE) corrected p<0.05), but not during extinction training (Figure 5A). Trend-level results showed differential cerebellar activation in posterolateral regions (lobule VI, Crus I, Crus II, lobule VIIb, and VIIa; Figure 4A and D; uncorrected p<0.05) during acquisition training, which match results obtained by Ernst et al., 2019.

In the parametric modulation analysis, there were significant activations in the cerebellum (lobule and vermis I-VI), DCN, and VTA negatively associated with CS+ prediction value modulation during acquisition training (CS+ × prediction; Figure 4B), whereas there were significant positively associated activations during extinction training (Figure 5B; lobule VI and Crus I bilaterally, right lobule V and VTA). Parametric modulation results indicate that early CS+ presentations are paired with high cerebellar activations in both acquisition and extinction training, with prediction values being low during early and high during late acquisition, whereas prediction values are high during early and low during late extinction (Figure 9). In other words, the activations were negative due to high activations of early CS+ compared to late CS+ events with prediction values taking the opposite course (i.e. low during early acquisition and high during late acquisition).

Activations related to US presentation (US post CS+ > no US post CS-)

During acquisition training, significant activations related to the presentation of the US (US post CS+ > no US post CS-) were widespread and included the whole cerebellum and VTA (Figure 4C). Cerebellar activations extended to the DCN. Peak activations in the cerebellum were found ipsilateral to the US (electrical stimulation applied to right hand).

Figure 4

Download asset Open asset

Cerebellar cortex (CB), DCN, and VTA fMRI activations related to prediction and presentation of the US during acquisition training.

CB activations are shown on cerebellar flatmaps (SUIT Diedrichsen and Zotow, 2015), while VTA and DCN activations are displayed on coronal slices progressing from posterior to anterior, with MNI y-coordinates indicated at the lower left of each slice. A 3D rendering in the bottom right shows slice locations within the probabilistic DCN and VTA atlases (see Methods for atlas generation). Trend-level results are indicated by gray diagonal lines in the background of the panels. (A) The event-based CS+ > CS- prediction contrast showed trend-level activations in posterolateral cerebellar lobules and the VTA; no DCN activations were observed. (B) Inverse activations were observed in the parametric modulation analysis of CS+×prediction, with activations mainly in the anterior cerebellum, paravermis, lobule VI and Crus I, and the VTA, with trend-level activations in the DCN. (C) Presentation of the US elicited widespread activations in the CB, DCN, and VTA. Data are shown for n=43 participants. Statistical significance was defined as TFCE-FWE-corrected p<0.05; trend-level results correspond to uncorrected p<0.05. VTA: ventral tegmental area; DCN: deep cerebellar nuclei; DN: dentate nucleus; IN: interposed nucleus; FN: fastigial nucleus; CS: conditioned stimulus; US: unconditioned stimulus; MNI: Montreal Neurological Institute standard brain; SUIT: spatially unbiased atlas template of the cerebellum; TFCE t: threshold-free cluster-enhanced test statistic; FWE: family-wise error.

Figure 5

Download asset Open asset

Cerebellar cortex (CB), DCN, and VTA fMRI activations related to the prediction and presentation of the US.

CB activations are shown on cerebellar flatmaps (SUIT Diedrichsen and Zotow, 2015), while VTA and DCN activations are displayed on coronal slices progressing from posterior to anterior with MNI y-coordinates indicated on the lower left for each slice. A 3D rendering in the bottom right shows the slice locations within the probabilistic DCN and VTA atlases (see methods for atlas generation). Trend-level results are indicated by gray diagonal lines in the background of the panels. (A) The event-based CS+ > CS- contrast showed trend-level activations in lobules VIIb, VIIIa, and VIIIb and the central VTA, with no DCN activations. (B) The parametric modulation contrast CS+ × prediction showed activations in lobule VI and Crus I, the vermis, and the VTA, as well as trend-level activations in the DCN. (C) Trend-level activations related to the omission of the US were observed in the CB, DCN, and VTA. Data are shown for n=43 participants. Statistical significance was defined as TFCE-FWE-corrected p<0.05; trend-level results correspond to uncorrected p<0.05. VTA: ventral tegmental area; DCN: deep cerebellar nuclei; DN: dentate nucleus; IN: interposed nucleus; FN: fastigial nucleus; CS: conditioned stimulus; US: unconditioned stimulus; MNI: Montreal Neurological Institute standard brain; SUIT: spatially unbiased atlas template of the cerebellum; TFCE t: threshold-free cluster-enhanced test statistic; FWE: family-wise error.

Activations related to the unexpected omission of the US (first 3 no US post CS+ > no US post CS- and no US post CS+ × prediction error)

Considering the whole extinction training phase, trend-level results showed activations in posterolateral lobules and the VTA related to the omission of the US (no US post CS+ > no US post CS-; Figure 5C). When focusing specifically on the early trials in extinction training to quantify activations related to the unexpected omission of the US, we found significant activations in both event-based and parametric modulation analyses.

In the event-based analysis, the CB, DCN, and VTA showed significant activations related to unexpected omissions of the US in initial trials for all four phases (first 3 no US post CS+ > no US post CS-; extinction training, recall test, reacquisition, and reextinction; Figure 6). The location of activations was consistent across phases, with highest activations in left posterolateral lobule VI and Crus I. Activations in the left posterolateral CB were paired with high activations in the left dentate nucleus. The VTA was bilaterally activated.

Figure 6

Download asset Open asset

Cerebellar cortex (CB), DCN, and VTA fMRI activations to the unexpected omission of the US (event-based analysis).

CB activations are shown on cerebellar flatmaps (SUIT Diedrichsen and Zotow, 2015), while VTA and DCN activations are displayed in coronal slices progressing from posterior to anterior with MNI y-coordinates indicated on the lower left for each slice. A 3D rendering in the bottom right shows the slice locations within the probabilistic DCN and VTA atlas (see Methods for atlas generation). Trend-level results are indicated by gray diagonal lines in the background of the panels (**A–D**). Event-based contrasts of the first three unexpected omissions showed activations in the CB, DCN, and VTA for all four phases. Activations in the DCN during reextinction were trend-level (D). Activations were most consistent in left lobule VI and Crus I, with DCN activation mainly in the left dentate, while VTA activations were bilateral. Data are shown for n=43 participants. Statistical significance was defined as TFCE-FWE-corrected p<0.05; trend-level results correspond to uncorrected p<0.05. VTA: ventral tegmental area; DCN: deep cerebellar nuclei; DN: dentate nucleus; IN: interposed nucleus; FN: fastigial nucleus; CS: conditioned stimulus; US: unconditioned stimulus; MNI: Montreal Neurological Institute standard brain; SUIT: spatially unbiased atlas template of the cerebellum; TFCE t: threshold-free cluster-enhanced test statistic; FWE: family-wise error.

In the parametric modulation analysis, significant activations were found in the cerebellum for unexpected omissions during extinction training and the recall test (Figure 7A and B), and showed trend-level activations during reacquisition and reextinction (Figure 7C and D). The highest activations were again found in left lobule VI and Crus I. In the DCN, significant activations were found during extinction training with peak activations in the left dentate, with trend-level results in the recall test. In the VTA, results were significant during extinction training (Figure 7A), with trend-level activations during the recall test, reacquisition, and reextinction.

Figure 7

Download asset Open asset

Cerebellar cortex (CB), DCN, and VTA fMRI activations to the unexpected omission of the US.

CB activations are shown on cerebellar flatmaps (SUIT Diedrichsen and Zotow, 2015), while VTA and DCN activations are displayed in coronal slices progressing from posterior to anterior with MNI y-coordinates indicated on the lower left for each slice. A 3D rendering in the bottom right shows the slice locations within the probabilistic DCN and VTA atlas (see Methods for atlas generation). Trend-level results are indicated by gray diagonal lines in the background of the panels (**A–D**). Parametric modulation contrasts related to prediction errors from unexpected US omissions showed activations in all four phases in the cerebellum, DCN and VTA. Activations during the recall test in the DCN and VTA were trend-level (B). Activations in reacquisition and reextinction were trend-level (**C, D**). Activations were most consistently found in lobule VI and Crus I. In addition, activations were found in the vermis, Crus II and lobule VIIb. Data are shown for n=43 participants. Statistical significance was defined as TFCE-FWE-corrected p<0.05; trend-level results correspond to uncorrected p<0.05. VTA: ventral tegmental area; DCN: deep cerebellar nuclei; DN: dentate nucleus; IN: interposed nucleus; FN: fastigial nucleus; CS: conditioned stimulus; US: unconditioned stimulus; MNI: Montreal Neurological Institute standard brain; SUIT: spatially unbiased atlas template of the cerebellum; TFCE t: threshold-free cluster-enhanced test statistic; FWE: family-wise error.

PPI: functional connectivity during unexpected US omissions between the cerebellum and VTA

Next, we aimed to assess functional interactions between the cerebellum and VTA using psychophysiological interaction (PPI), using the VTA as a seed region. In the parametric modulation analysis, we observed trend-level positive functional connectivity in left lobule VI and Crus I for extinction training, recall test, reacquisition, and reextinction (Figure 8A). When combining the contrasts by summing, the combination of all four phases showed similar positive functional connectivity between the CB and the VTA in lobule VI and Crus I, but only on a trend-level (Figure 8B). No activations were seen in the DCN.

To assess whether similar connectivity patterns were observed when using an alternative seed, we performed an additional PPI analysis using a cerebellar region of interest defined from the conjunction analysis of unexpected US omissions (see volumes of interest (VOI) definition in Methods and Figure 8—figure supplement 2) as the seed. In this analysis, connectivity effects in the VTA were very limited and restricted to a small number of voxels at an uncorrected threshold (Figure 8—figure supplement 1).

Figure 8 with 2 supplements see all

Download asset Open asset

Functional connectivity analysis using PPI and DCM.

(**A, B**) Trend-level activations in the cerebellum found by PPI analysis using the VTA as a seed region, shown on cerebellar flatmaps (Diedrichsen and Zotow, 2015). To highlight activations, a zoomed cutout of lobule VI and Crus I is also shown. Trend-level results are indicated by gray diagonal lines in the background of the panels. (A) Trend-level connectivity with VTA is found in lobule VI and Crus I for extinction training, the recall test, reacquisition, and reextinction using PPI of the unexpected US omission contrasts from the parametric modulation analysis with a VTA seed. (B) Summary of trend-level PPI results by summing all contrasts shown in panel A. (**C–E**) DCM analysis showed significant modulation of cerebellar cortex (CB) to VTA connections with prediction errors during unexpected omission events. (C) DCM model is shown, modulations act on both the CB/DCN to VTA as well as the VTA to CB/DCN connections. Events, i.e., US post CS+, no US post CS+, no US post CS-, CS+, and CS- for each phase, are provided as inputs to both nodes. Colors of bars are determined by the posterior probability calculated in the PEB analysis (very strong: p>0.99; strong: 0.95<p<0.99; moderate: 0.75<p<0.95; weak: p<0.75). Very strong and strong results are considered significant, moderate results are considered trend-level. (D) Significant modulations during extinction training, the recall test, and reacquisition. The strongest result was found during extinction training in the connection from the cerebellum to VTA. (E) Trend-level DCM results showed connectivity from the DCN to VTA in extinction, and from the VTA to DCN in reextinction. Data are shown for n=43 participants. Trend-level PPI results correspond to uncorrected p<0.05. PPI: Psychophysiological interaction; DCM: Dynamic causal modeling; PEB: Parametric Empirical Bayes; PE: Prediction error; P: Posterior probability; VTA: Ventral tegmental area; CB: Cerebellar region in lobule VI and Crus I consistently active during unexpected US omissions; US: unconditioned stimulus; L: left; R: right; SUIT: spatially unbiased atlas template of the cerebellum; t: test statistic.

DCM: functional connectivity during unexpected US omissions between the cerebellum and VTA

In the dynamic causal modeling (DCM) parametric modulation analysis, unexpected omissions significantly positively modulated both the connection from CB to VTA and VTA to CB during extinction training (Figure 8D; very strong evidence, posterior probability p>0.99, strong evidence: 0.95<p<0.99). In the recall test, the VTA to CB connection was positively modulated. In reacquisition, the CB to VTA connection was positively modulated. Results were less strong for the other unexpected omission modulations, with trend-level moderate evidence for the CB to VTA connection in the recall test and the VTA to CB connections in reacquisition, and weak evidence for both CB to VTA and VTA to CB connections in reextinction (Figure 8D; moderate 0.75<p<0.95; weak p<0.75). In the model incorporating the DCN and VTA, there was a trend-level modulation from the DCN to the VTA during extinction training, as well as a trend-level modulation from the VTA to the DCN during reextinction (Figure 8E; moderate significance, 0.75<p<0.95).

Discussion

The cerebellum and VTA were active during unexpected omissions of aversive unconditioned stimuli in the initial extinction training trials and in other learning phases. Notably, this activation was accompanied by increased functional connectivity between the cerebellum and VTA. Given that the unexpected omission of an aversive stimulus can be considered rewarding, our data suggest that interactions between the cerebellum and the VTA contribute to reward-like prediction error processing during fear extinction learning in humans. Such interactions may represent one mechanism by which the cerebellum contributes to non-motor functions, in this case, the control of emotions. Our findings build on previous research in rodents showing that the cerebellum modulates addictive, social and depression-like behavior via direct projection to VTA (Carta et al., 2019; Baek et al., 2022).

Cerebellar activations were most prominent in trials with maximum prediction errors, i.e., during the initial extinction training phase, but were also seen in early recall, reacquisition, and early reextinction. The activations in lobule VI and Crus I are consistent with previous findings by Ernst et al., 2019 during unexpected omissions in a partially reinforced acquisition phase.

Additionally, other studies have reported activations in these regions during the prediction of the US in extinction training (Linnman et al., 2012; Kattoor et al., 2014; Faul et al., 2020; Labrenz et al., 2022) and during unexpected omissions of aversive stimuli (Ploghaus et al., 2000; Yágüez et al., 2005; Batsikadze et al., 2022; Doubliez et al., 2023). More broadly, activations in lobule VI and Crus I have been observed during US prediction (Fullana et al., 2016), particularly in the early phases of acquisition training. Conversely, unexpected rewards lead to activation in Crus I (Ramnani et al., 2004), and a meta-analysis implicated left lobule VI and Crus I in both the anticipation and processing of rewarding outcomes (Kruithof et al., 2023). Anticipation-related activations were more hemispheric, whereas rewarding outcome activations were more paravermal.

Activated areas in more lateral parts of lobule VI and Crus I overlap with multi-demand regions (Nettekoven et al., 2024), whereas activations neighboring the vermis overlap with emotional regions (Guell et al., 2018). The main components of the multi-demand (or executive) neural network are working memory, attention, and inhibition (Camilleri et al., 2018). Thus, attention-related processes involved in fear conditioning may also contribute to cerebellar activations. Attention-related processes have been related to the salience of the stimulus (Rescorla and Wagner, 1972), but are also thought to change when a cue is followed by an unexpected outcome with more attention being paid in the next trial (Pearce and Hall, 1980; Koenig et al., 2017). Attention-related processes may, therefore, contribute to activations in the more lateral cerebellar parts of lobule VI/Crus I. However, unexpected omission of the US should lead to increased attention to the CS in the next trial and, therefore, increased activation towards the CS (which was not the case) and not at the time the US was presented and did not occur.

Our findings of increased activation in the VTA related to the unexpected omission of the aversive US align well with rodent data (Salinas-Hernández et al., 2018; Kalisch et al., 2019). In rodents, it has been shown that dopamine neurons in the VTA are active in initial extinction trials, and that lack of this dopamine signal prevents extinction learning (Luo et al., 2018; Salinas-Hernández et al., 2018; Cai et al., 2020). In a more recent study, Salinas-Hernández et al., 2023 observed that these dopamine neurons project to extinction-related neurons of the nucleus accumbens, and receive input from the dorsal raphe. Our data show that the VTA is also involved in extinction learning in humans, and provide the first evidence that the VTA is functionally coupled with the cerebellum during extinction learning. Esser et al., 2021 found that the administration of L-Dopa in healthy human participants enhanced activation of the nucleus accumbens which was associated with VTA activity at the time of unexpected omission of the US in early extinction learning, which is in good agreement with our findings.

The unexpected omission of the aversive US is thought to be rewarding and fMRI signals in the cerebellum and VTA may be related to reward-like prediction errors. Whereas the VTA has long been known to process rewards, and fMRI data in humans indicate that the VTA is active related to the unexpected presentation of rewards (D’Ardenne et al., 2008; Schott et al., 2008), the cerebellum has only recently been found to process reward signals in rodents (Wagner et al., 2017; Kostadinov et al., 2019; Kostadinov and Häusser, 2022). fMRI signals are known to be driven by synaptic input (Logothetis and Wandell, 2004; Logothetis et al., 2001; Logothetis, 2008). In the cerebellum, this is mainly related to mossy fiber input to the CB (Shahshahani et al., 2024). There is evidence, however, that granule cell activity also modulates vasodilatation and likely contributes to fMRI BOLD signals in the CB (Gagliano et al., 2022; Mapelli et al., 2017). Likewise, recent evidence suggests that activation of dopamine neurons in the VTA leads to increased BOLD signals using DREADD-fMRI in rodents (Zheng et al., 2023). Human fMRI-PET studies have shown that VTA BOLD activation in response to rewards correlates with increases in dopamine in the striatum, suggesting that the fMRI signal in the SN/VTA is related to dopaminergic neuron activity (Schott et al., 2008; Düzel et al., 2009). fMRI signals in the CB and VTA may, therefore, reflect incoming reward-related signals, prediction error processing, or maybe most likely, both.

The origin of the afferent reward signals to the cerebellum remains unclear; while an early study showed afferent connections from the VTA to the cerebellum (Ikai et al., 1992), this finding has not been replicated in more recent research (Li et al., 2023). However, preliminary evidence from Guarque-Chabrera et al., 2024 suggests the existence of a connection from the VTA to the cerebellum, although its strength and functional relevance remain uncertain. Given this limited evidence, the VTA-to-cerebellum interactions detected with DCM may not be driven primarily by a direct connection. Instead, they may reflect indirect network-level interactions within the mesolimbic and mesocortical systems.

In rodents, a subset of VTA dopamine neurons has been found to be active when the US is expected and does not occur, and are therefore considered to represent (reward-like) prediction error signals driving extinction learning (Salinas-Hernández et al., 2018; Salinas-Hernández et al., 2023). The VTA, however, also contains neurons which are active during the presentation and prediction of threat (Jo et al., 2018; Moriya et al., 2018), and likewise we observed fMRI activation of the VTA related to threat prediction and presentation. Cerebellar activations and their interactions are likely not limited to reward-like prediction errors; they may also be related to unexpected presentation of aversive stimuli. Because aversive stimulus presentation results in pronounced cerebellar activations and the response to the US habituates over time, we were unable to separate cerebellar activation related to the unexpected (initial acquisition trials) and the expected (late acquisition trials) presentation of the US.

Although the present findings are consistent with the interpretation of prediction error-related activations in the VTA (and cerebellum), based on the present fMRI data, we cannot decide whether activations reflect surprise signals (i.e. unsigned prediction error-related signals) or signed prediction error signals (Pearce and Hall, 1980; Rouhani and Niv, 2021; Montague et al., 1996; Barto, 2018). Rodent data show that cerebellar neurons in the cortex and nuclei increase activations both related to the unexpected presentation and omission of reward signals (Hull, 2020), which would indicate that the cerebellum primarily contributes to unsigned (reward) prediction errors. Likewise, a meta-analysis of human prediction error showed that activations in the cerebellum were more consistently related to unsigned prediction errors (Corlett et al., 2022).

Results showed significant functional interaction between the cerebellum and VTA during unexpected US omission in extinction training, and interactions were also observed during spontaneous recovery and reacquisition. This suggests that cerebellum-VTA coupling consistently emerges in conditions where prediction errors occur. Predictive signals of the cerebellum may be linked to the timing of the unconditioned response and, therefore, its omission and may interact with VTA-related prediction error processing.

Although interpretation must remain cautious given BOLD and DCM limitations outlined in more detail below, we detected phase-specific differences in cerebellum-VTA connectivity. Extinction learning showed bidirectional coupling, spontaneous recovery showed primarily VTA-to-cerebellum influences, and reacquisition showed predominantly cerebellum-to-VTA influences. A tentative interpretation is that a bidirectional loop may support the formation of an extinction memory during active learning, whereas in spontaneous recovery the extinction memory is mainly retrieved, possibly reducing the need for cerebellar output. In reacquisition, reactivation of the fear memory may again engage cerebellar outputs that modulate omission-related processing in the VTA, while the reactivated fear memory during reinforced trials may reduce the detectability of VTA-to-cerebellum influences.

Notably, the unexpected US omission area in the cerebellum was more prominent on the left, while we applied the US to the right. In our previous work, it was also the left cerebellum which was mainly activated, however, stimulation was always done on the left (Ernst et al., 2019; Batsikadze et al., 2024). As outlined above, we believe that cerebellar activation is reward-related. In a recent meta-analysis of fMRI data in monetary reward tasks, reward outcome was also related to the left cerebellum (i.e. lobule VI in the cerebellar hemisphere and vermis) (Kruithof et al., 2023). There is good evidence that emotional information is most prominently processed in the right cerebral hemisphere (Gainotti, 2019; Güntürkün et al., 2020). The present data suggest that emotional processing may also be lateralized in the cerebellum, with the left cerebellar hemisphere projecting to the right cerebral cortex and vice versa (Pakusch et al., 2025). Lateralization of positive emotions to left cerebellar lobule VI has also been reported by Liu et al., 2018. Furthermore, emotion-related tasks resulted in increased activity of the left cerebellar hemisphere in patients with anorexia nervosa compared to controls (Zhong et al., 2023). Meta-analyses of fMRI studies of emotional tasks, however, report mostly bilateral cerebellar activations with no clear lateralization (Keren-Happuch et al., 2014; Stoodley and Schmahmann, 2009; Pierce et al., 2023). These studies primarily focused on emotion recognition tasks. Future research should explore the lateralization of emotional processing in the cerebellum using a variety of emotional tasks. In addition to emotions, there is also a preference of the right cerebral hemisphere for the ventral fronto-parietal attention network, which has been related to the detection of potentially harmful events (Bartolomeo and Seidel Malkinson, 2019). Thus, as outlined above, attention-related processes involved in fear conditioning may also contribute to (lateralized) cerebellar activations.

In the following, we discuss potential limitations of our study. First, while we demonstrated functional connectivity in the DCM analysis, the results were weaker with PPI. PPI is known to frequently lack power and to have a high proportion of false negatives (Friston et al., 1997; O’Reilly et al., 2012), particularly for event-based designs. Additionally, the low number of unexpected events in our experiment, due to the 100% reinforcement/non-reinforcement rates in acquisition training (which was designed to induce maximum prediction error in the initial extinction trials), may have made detecting activation challenging. Furthermore, as the PPI regressor is demeaned, any persistent activation (e.g. due to an anatomical connection) will decrease detected connectivity. Because cerebellar and VTA activation co-occur not only in response to unexpected US omissions, but also to US presentation and prediction, this lack of specificity could have weakened our results. The same reasons apply for the lack of observation in functional connectivity between the DCN and the VTA.

Second, in the DCM analysis, we observed strong modulation related to unexpected US omissions when using the CB and VTA as volumes of interest (VOIs), but observed only trend-level results when substituting the CB with the DCN. This is likely explained due to the high iron content in the DCN causing susceptibility artefacts and reducing signal (Diedrichsen et al., 2011). Although we opted for a whole-brain sequence with a conventional echo-time in order to generate a dataset of general value, future studies could opt for lower echo times and optimize for detection of signal in the DCN (de Hollander et al., 2017).

Third, the extent of the VTA varies depending on the MRI atlas used. The VTA in the atlas from Pauli et al., 2018 is about 4 times smaller than the VTA in the atlas by Trutti et al., 2021 due to different VTA definitions used. Our average volumes are in good accordance with Trutti et al., 2021, which in turn is in good accordance with histological data. The VTA and its boundaries, particularly laterally, are not clearly visible on MRI scans and were identified by using landmarks, such as the substantia nigra, red nucleus, interpeduncular fossa, and the cerebral aqueduct. We have approximated the VTA region as closely as we could, but cannot rule out imperfect masks, especially at both lateral ends of the VTA.

In conclusion, our study provides evidence that interactions between the cerebellum and VTA contribute to fear extinction learning. These interactions may support reward-like prediction error processing, a mechanism thought to drive fear extinction learning. Previous studies found that the cerebellum modulates addictive and social behavior via direct projection to VTA in rodents (Carta et al., 2019; Baek et al., 2022). In humans, our findings extend this framework by highlighting cerebellum-VTA interactions involved in emotional processing. Incorporating these interactions into models of the fear extinction network may provide new ways of improving exposure therapy by targeting the dopaminergic system or stimulating the cerebellum specifically during unexpected learning events.

Methods

Preregistration

This study was preregistered on OSF on 16/11/2023 (Nio et al., 2023). The preregistration document can be viewed at https://doi.org/10.17605/OSF.IO/2PXWE.

The preregistration outlines the methods explained here and includes the main hypotheses. Adjustments were made to specific parts of the methods, which include the optimized localization of the VTA and DCN by manual drawing by an expert annotator, and the inclusion as volumes of interest (VOIs) in the functional connectivity analyses.

Deviations from preregistration

A primary deviation from the preregistration involved the event-based fMRI analysis of unexpected US omissions. The preregistered analysis plan specified event-based contrasts using early versus late halves of extinction, recall, reacquisition, reextinction, and unexpected US phases. In addition, the preregistration described an exploratory analysis focusing on the initial three unexpected omission trials within these phases. In the final analyses, this preregistered exploratory first-three-trial contrast was used as the main event-based fMRI analysis, replacing the early/late block contrasts. This decision was supported by the rapid decay of behavioral differentiation and model-derived prediction error estimates observed under the full (non-) reinforcement structure of the paradigm (100% reinforcement during acquisition and 0% during extinction and recall), which indicated that prediction errors were largely confined to the initial trials (see Figure 2 for SCRs, Figure 3 for PSRs, and Figure 9 for prediction error estimates).

Furthermore, several preregistered exploratory analyses were not performed. In particular, a preregistered exploratory multi-node DCM of the broader fear extinction network was omitted as it was out of scope for the current VOI-specific manuscript centered on cerebellum-VTA interactions. Other preregistered exploratory analyses were likewise not performed, as they were not necessary to address the primary hypotheses of the present study.

In addition, the preregistration contains a reporting error in the sample size rationale: the alpha level was listed as p<0.05, whereas the underlying power analysis conducted during grant preparation used p<0.005. This affects the estimated minimum sample size but not the target sample size, which was 50 participants in both the preregistration and the original manuscript. The correct parameters are reported below.

Participants

Experiment power estimate and number of participants

Experiment power was estimated based on previous fear conditioning data acquired at 7T fMRI (Ernst et al., 2019) using the FMRIPower toolbox (fmripower.org) (Mumford, 2012). Considering first level unexpected US omission contrasts during a partially reinforced acquisition training and aiming for a power of 80% at p<0.005 for a one-sided hypothesis test, group sizes were estimated at 41 participants for Crus I ipsilateral to US presentation (effect size 0.56 in units of standard deviations). To account for potential dropouts and outliers, 50 participants were initially recruited. However, only 44 successfully completed the 3 day study, with reasons for non-completion, including Covid-related issues (one participant), problems with the Digitimer equipment (three participants), and failure to attend on day 3 (two participants). Additionally, one participant dropped out due to a shimming failure, resulting in a final sample size of 43 participants.

Exclusion criteria and demographic information

Participants were between 20 and 35 years old (23 men, 20 women, mean age: 24.7 (SD = 3.4) years), fluent in German, non-smokers, reported no intake of medication, or illicit drugs affecting the central nervous system, had no history of neurological or mental disorders for themselves or their first-degree relatives, and had no previous participation in similar learning experiments. Additionally, female participants were not pregnant, breastfeeding, or using hormonal contraceptives. Before the experiment, participants completed the Depression Anxiety Stress Scale (DASS21G) (Henry and Crawford, 2005). Participants who scored higher than 20 for depression, 14 for anxiety, or 25 for stress components were excluded from the study. All participants were right-handed based on the Edinburgh handedness inventory (Oldfield, 1971). Participants were asked to refrain from alcohol consumption the night before the experiment. Informed consent and consent to publish were obtained from all participants. The study was approved by the Ethics Committee of the University Hospital Essen (approval number: 20–9664-BO) and conducted in accordance with the Declaration of Helsinki.

Fear conditioning paradigm

Participants underwent a three-day differential fear conditioning paradigm. The paradigm presentation was controlled by a computer running Presentation (version 21.0, Neurobehavioral System Inc, Berkeley, CA). Participants were shown two visual stimuli (square and diamond) as CS (CS+ and CS-) and informed that electrical stimulations would be applied during the experiment. Three types of trials were used: reinforced CS+ (CS+ paired with the US; CS+/US), unreinforced CS+ (no US), and unreinforced CS-. The CS- was never reinforced.

The paradigm consisted of seven phases separated over three days (Figure 1A). On day 1, the experiment started with a habituation phase consisting of 8 unreinforced CS- and CS+ trials (four CS+ trials, four CS- trials, presented in semi-randomized order). This phase is intended to decrease novelty effects and enable participants to acclimate to the scanner environment and presentation of visual stimuli. The habituation phase was followed by acquisition training consisting of 32 trials with fully reinforced CS+ (16 CS+/US, 16 CS- trials), during which the fear association was learned. On day 2, extinction training was applied and consisted of 32 unreinforced trials (16 CS+, 16 CS-). On day 3, the recall test, reacquisition, reextinction, and the unexpected US phase were applied. The recall test consisted of 16 unreinforced trials (8 CS+, 8 CS-). This phase tested the recall of extinguished fear responses and showed spontaneous recovery (see Results). Reacquisition, reextinction, and the unexpected US phases were done in a single continuous fMRI run without pauses between sequences. The partially reinforced reacquisition phases consisted of 36 trials (15 CS+/US trials, 3 CS+, 18 CS-) with three unexpected omissions of the US, reextinction consisted of 24 trials (12 CS+, 12 CS-), and the unexpected US phase consisted of 28 trials (3 CS+/US trials, 11 CS+, 14 CS-) with three unexpected reinforced CS+ trials (Figure 1A). Learned fear responses were measured bySCRs and pupil size reactions (PSRs) in each trial. The reacquisition, reextinction, and unexpected US phases were included to increase the number of trials with high prediction errors (i.e. discrepancy between US expectation and actual US presentation).

We used the model from Batsikadze et al., 2022 to optimize our experimental design. Values for US prediction and US prediction error produced by the model were recorded during simulation. Simulations were repeated for different paradigm CS sequences, and the sequence which required a short number of trials and still yielded high prediction errors was selected.

Figure 9

Download asset Open asset

Mean absolute prediction error values for each trial estimated from skin conductance responses (SCRs), with CS+ and CS- responses paired in blocks.

Reinforcement of the CS+ by a US (CS+/US) is indicated by filled squares. Prediction and prediction error values were estimated over 200 iterations. CS: conditioned stimulus; US: unconditioned stimulus.

The CS+ and CS- were presented in equal numbers during the first and second half of each phase, regardless of whether the CS+ was followed by a US or not. Counterbalancing was implemented for the visual CS (CS+ square and CS- diamond, or CS+ diamond and CS- square) as well as for the order of presentation of the CS during the unreinforced habituation, extinction training, and recall test phases (i.e. CS+ presented first or CS- presented first). The reinforced acquisition training and reacquisition phases always started with a CS+. The two orders of CS presentation were pseudorandomly generated, similar to Ernst et al., 2019. Participants were instructed to pay attention to any possible connection between the CS and US presentations. They were also informed that the patterns of CS and US presentations would stay the same across all phases. However, they were not told about the CS/US contingencies or the timing and occurrence of the US.

In the analysis, the focus was on trials containing unexpected omissions of the US (CS+/no US) with high prediction error values. These trials occurred during early extinction, early recall, reacquisition, and the reextinction phase.

Visual conditioned stimuli (CS)

Visual stimuli were displayed on an fMRI monitor (BOLD-screen 32, Cambridge Research Systems Ltd., Rochester, UK) placed at the end of the scanner bore which was projected to the participant using a mirror system (Figure 1B). Two pictures of black geometric figures (a square and a diamond shape) on a gray background were used as CS+ and CS-. Both CS were presented for 6 s. The period of the CS was shortened from 8 s used in our previous study (Ernst et al., 2019) to 6 s to decrease the length of the experiment, while still being able to assess SCRs (Lonsdorf et al., 2017). In the periods between visual stimuli a fixation cross was shown (time of ITI: varies between 9 s and 13.5 s).

Aversive unconditioned stimulus (US)

In reinforced CS+ trials, the visual stimulus co-terminated with the presentation of the aversive US, an electrical stimulation (100 ms) applied to the first dorsal interosseous muscle area of the right hand via a 6.5 mm concentric surface electrode (WASP electrode, Specialty Developments, Bexley, UK). The US was produced by a high voltage DC stimulator (DS7A, Digitimer Ltd., London, UK) and consisted of four consecutive 500 µs current pulses (maximum output voltage: 400 V) with an interpulse interval of 33 ms. The location of the US electrode was kept constant across all days of measurement by marking the electrode position with a permanent marker on days 1 and 2. Before the first experimental phase began, the individual electrical stimulation intensity threshold was calibrated to be perceived as ‘very unpleasant but not painful’. To reach the threshold, the current strength was progressively increased and modulated according to each participant’s feedback. The calibrated US intensity was increased by 20% to compensate for possible habituation to the US, which could result in a weakening of the conditioned responses as successfully done in previous studies of our group (Batsikadze et al., 2022; Inoue et al., 2020). The US intensity remained the same for the three days of the experiment (mean US intensity: 3.8 (SD = 2.9) mA, range 0.6–19.2 mA).

Physiological data acquisition

During each phase, SCRs and pupil size were recorded. SCRs were acquired with appropriate hardware filters sampling at 2 kHz through an MP160 Data Acquisition Hardware unit (BIOPAC Systems Inc, Goleta, CA). Two SC electrodes were attached to the left-hand hypothenar eminence. Pupil size was measured using an MRI-compatible eye tracking system (Avotec Inc, Stuart, FL). Calibration of the eye tracking system was performed prior to each phase to track gaze position on the screen.

Self-reports

After each learning phase, participants completed questionnaires using a 4-button fiberoptic response device (Current Designs, Haverford, PA). The Likert scale questionnaires assessed self-reports, such as arousal (rated on a 1–9 scale from 1: ‘very calm’ to 9: ‘very nervous’), valence (rated on a 1–9 scale from 1: ‘comfortable’ to 9: ‘uncomfortable’), and fear (rated on a 1–9 scale from 1: ‘not afraid’ to 9: ‘very afraid’) related to the visual stimuli. The questionnaires also included questions about US perception and CS-US contingency. Additionally, participants completed standardized questionnaires before the start of the experiment to assess handedness, depression, anxiety, and stress levels. All participants were screened for contraindications of 7T MRI.

Skin conductance analysis

To eliminate high frequency artefacts, skin conductance data was first low-pass filtered at 10 Hz (62^nd-order Blackman FIR filter) in MATLAB (Release 2022b, RRID:SCR_001622, The MathWorks Inc, Natick, MA). SCRs were defined as the maximum trough-to-peak amplitude within a given time interval after CS onset using semi-automated peak detection implemented in a MATLAB-based EDA-Analysis App (Otto et al., 2023). A minimum amplitude criterion of 0.01 μs was used as the SCR detection threshold. In each trial, SCRs were evaluated for two distinct time windows: the conditioned response within a time window of 1.0–5.9 s after CS onset and the unconditioned response within a time window of 6–10 s after CS onset (irrespective of whether a US was presented). To account for between-subject variance, the resulting raw SCR amplitudes were increased by 1 μs and normalized through a logarithmic transformation (LN(1+SCR)) (Boucsein, 2012).

Pupil size analysis

Preprocessing of the raw pupil size data was performed to detect and remove blinks (Kret and Sjak-Shie, 2019). Trials with fewer than 50% of their data points remaining after blink removal were excluded from the analysis. For each trial, the baseline was computed as the mean pupil size during the 500 ms period prior to CS onset. The baseline was subtracted from the corresponding pupil size (Jentsch et al., 2020; Mathôt et al., 2018) and the result was divided by the baseline to compute the PSR. The mean pupil size response for CS+ and CS- trials were calculated in the time windows from 4–5.9 s and 6.0–8.0 s for the CS and US PSR, respectively. PSRs during possible US presentation were excluded (5.9–6 s). The time window for the CS was chosen based on Jentsch et al., 2020 who observed the largest differentiation between CS+ and CS- 2 s prior to the US. The US time window was chosen to match the CS time window. For valid trials, a missing pupillometric data point was considered as ‘not a number value’ to not influence the pupil size mean computation (Mathôt et al., 2018).

Behavioral statistical analysis

For physiological data, due to non-normal distribution (SCR and PSR: Shapiro-Wilk test, p<0.001), statistical analysis was performed using a non-parametric ANOVA-type statistic (ATS) implemented via the ANOVAF option in the PROC Mixed procedure in SAS Studio 3.8 (SAS Institute Inc, Cary, NC, USA). This method is suggested for addressing skewed distributions, outliers, or small sample sizes. To ensure more reliable results, an ATS was applied, with the denominator degrees of freedom set to infinity (Brunner et al., 2002; Shah and Madden, 2004; Noguchi et al., 2012) as the use of finite denominator degrees of freedom can result in increased type I errors (Bathke et al., 2009). The ATS was calculated for each phase (habituation, acquisition training, extinction training, recall test, reacquisition, reextinction, unexpected US phase), with stimulus type (CS+, CS-) and block (early vs. late, defined as the first and second halves of the phase, respectively) as within-subjects factors and SCRs or PSRs as dependent variables.

For analyses of self-reports, an ATS with repeated measures was calculated with stimulus type and phase as within-subjects factors and ratings (valence, arousal, fear, and US expectancy) as dependent variables.

For behavioral results, a p<0.05 criterion was used to determine if results were significantly different from those expected if the null hypotheses were correct. Post hoc comparisons were calculated using least square means tests and were adjusted for multiple comparisons using the Tukey-Kramer method.

Computational modeling

An artificial agent was trained to predict the likelihood of a US for a given visual input in a virtual version of the experiment as in Batsikadze et al., 2022. The model was based on reinforcement learning (Sutton and Barto, 2018) and consisted of a deep neural network (DNN) (LeCun et al., 2015). The model hyperparameters were fit to SCRs recorded in the experiment, which served as a proxy for participants’ expectation of a US.

The same simplified visual stimuli for trial stimuli $s_{t}$ and the same encoding for reinforcement signals $r_{t}$ were used as in Batsikadze et al., 2022. We employed the same network architecture, which comprised two hidden fully connected layers with 64 units each and an output layer with activation function $φ$ . To account for the counterbalanced order of CS+ and CS- presentations in the experimental design, two trial sequences were modeled separately: In trial sequence 1, habituation, extinction, and recall began with a CS+ trial, whereas in trial sequence 2 these phases began with a CS- trial. For each of the two trial sequences experienced by the participants, 25 randomly initialized agents were trained at each point of the hyper-parameter grid search. Their outputs were averaged and compared to the group-averaged SCR learning curves, and the final set of hyperparameters were selected by minimizing a cost function that quantified the difference between model predictions and SCR data. After selecting the best-fitting hyperparameters, 100 randomly initialized agents were trained anew, and their averaged outputs provided the trial-by-trial prediction $v_{t}$ and prediction errors $δ_{t}$ used in the fMRI analyses.

On each trial t, the agent stored an experience tuple $e_{t} = (s_{t}, r_{t}, δ_{t})$ in memory for later replay (Lin, 1991). The agents were trained using the backpropagation algorithm (Rumelhart et al., 1986) on batches of experiences of size b, which were sampled randomly from memory with a probability that was proportional to a priority score p:

P (e_{k}) = \frac{p_{k}}{\sum_{j = 1}^{T} p_{j}}

Priority scores depended on the experiences’ recency $λ^{τ}$ , where τ is the time passed since the experience and λ is a decay factor. Optionally, the priority could additionally depend on the magnitude of US prediction error, i.e., $p = | δ | λ^{τ}$ (Schaul et al., 2016). The parameter RPE ∊ {Yes, No} indicates whether the replay priorities also depended on the magnitude of the US prediction error. The number of replays $i$ was varied to control the degree of learning on each trial.

While the previous model could account for ABA renewal (Batsikadze et al., 2022), it cannot account for spontaneous recovery in the AAA paradigm, because, in the model, extinction in the same context overwrites the association formed during acquisition. Hence, we extended the model by an additional replay phase, which takes place between habituation, acquisition, extinction, recall, and reacquisition phases, and serves to recover the initially acquired association. This replay phase prioritized experiences according to the reinforcement $r_{k}$ received in a trial and consisted of a total of 100 replays of batches of size 128. Reactivation probabilities for these replays were computed as follows:

P (e_{k}) = \frac{e^{β r_{k}}}{\sum_{j = 1}^{T} e^{β r_{j}}}

where β denotes the inverse temperature and controls the relative difference of reactivation probabilities.

Hyper-parameter fitting

We averaged SCRs from CS+ and CS- trials separately for each trial sequence and applied min-max scaling. The averaged SCRs were accordingly defined as ${\bar{Y}}_{l} = ({\bar{y}}_{+, 1}, . . ., {\bar{y}}_{+, N}, {\bar{y}}_{-, 1}, . . ., {\bar{y}}_{-, N})$ , where ${\bar{y}}_{+, n}$ and ${\bar{y}}_{-, n}$ are the averaged SCRs for the n-th CS+ and n-th CS- presentations across all participants who completed a given trial sequence $l$ , respectively. Analogously, the averaged US predictions of the model were defined as ${\bar{V}}_{l} (b, λ, i, β, R P E, φ) = ({\bar{v}}_{+, 1}, \dots, {\bar{v}}_{+, N}, {\bar{v}}_{-, 1}, \dots, {\bar{v}}_{-, N})$ , where ${\bar{v}}_{+, n}$ and ${\bar{v}}_{-, n}$ are the averaged US predictions for the n-th CS+ and CS- presentations across all model instances who were trained on a given trial sequence l, respectively. For each trial sequence, 25 randomly initialized agents were trained with the same hyperparameters, and their outputs were averaged to reduce variability and obtain stable trial-wise estimates of US predictions which are able to show a good fit with the population data and prediction errors. The goodness of fit was defined as:

\bar{F} (b, λ, i, β, R P E, φ) = \sum_{l = 1}^{L} w_{l} (‖ {\bar{Y}}_{l} - {\bar{V}}_{l} (b, λ, i, β, R P E, φ) ‖ + P_{A c q} + P_{G e n} + P_{E x t} + P_{R e n} + P_{R e x})

where $w_{l}$ is the number of participants who experienced trial sequence l. To ensure that the overall learning curve in the model resembled that of the participants, we added the following penalty terms, if the model failed to

learn that CS+ is followed by the US: $P_{A c q} = {\begin{cases} 0, & {\bar{v}}_{A c q, E n d} \geq 0.5 \\ 1, & {\bar{v}}_{A c q, E n d} < 0.5 \end{cases}$ where ${\bar{v}}_{A c q, E n d}$ is the average US prediction over the last two CS+ presentations of acquisition training.
show the CR at the start of extinction training: $P_{G e n} = {\begin{cases} 0, & {\bar{v}}_{E x t, S t a r t} \geq 0.5 \\ 1, & {\bar{v}}_{E x t, S t a r t} < 0.5 \end{cases}$ where ${\bar{v}}_{E x t, S t a r t}$ is the average US prediction over the first two CS+ presentations of extinction training.
successfully extinguish the CR during extinction training: $P_{E x t} = {\begin{cases} 0, & {\bar{v}}_{E x t, E n d} < 0.1 \\ 1, & {\bar{v}}_{E x t, E n d} \geq 0.1 \end{cases}$ where ${\bar{v}}_{E x t, E n d}$ is the average US prediction over the last two CS+ presentations of extinction training.
show renewal of the CR: $P_{R e n} = {\begin{cases} 0, & {\bar{v}}_{R e c, S t a r t} \geq 0.5 \\ 1, & {\bar{v}}_{R e c, S t a r t} < 0.5 \end{cases}$ where ${\bar{v}}_{R e c, S t a r t}$ is the average US prediction over the first two CS+ presentations of recall.
successfully extinguish the CR during recall: $P_{R e x} = {\begin{cases} 0, & {\bar{v}}_{R e c, E n d} \leq 0.1 \\ 1, & {\bar{v}}_{R e c, E n d} > 0.1 \end{cases}$ where ${\bar{v}}_{R e c, E n d}$ is the average US prediction over the last two CS+ presentations of recall.

A grid search was conducted over the hyper-parameter sets shown in Table 1. The model with the best goodness-of-fit was then chosen to derive trial-by-trial values for US predictions and prediction errors. Prediction errors were used for parametric modulation of fMRI data to test hypotheses related to prediction error processing of the cerebellum and VTA.

Table 1

Modeling.

Parameters for which the grid search was run.

Grid search parameters
b (batch size)	{1, 16, 32, 64}
λ (decay factor)	{0.5, 0.525, 0.55,..., 0.95, 0.975, 1}
i (training repeats)	{1, 2, 3, 4, 5, 6, 7, 8, 9, 10}
β (inverse temperature)	{0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5}
RPE	{Yes, No}
φ	{ELU, Sigmoid}

MRI acquisition

All MRI imaging data were collected while participants were lying supine in the 7T MRI scanner (MAGNETOM Terra, Siemens Healthineers AG, Forchheim, Germany). A 1-channel transmit/32-channel receive head RF coil (Nova Medical Inc, Wilmington, MA, USA), was used. Two dielectric pads for signal homogenization were placed on either side of each participant’s upper neck (Teeuwisse et al., 2012). Depending on head size, further cushioning was added to prevent participant head movement and discomfort.

Whole-brain functional MRI acquisition was performed with a 3-dimensional echo planar image sequence (Stirnberg and Stöcker, 2021) with an isotropic voxel size of 1.5 mm. The sequence was run for each phase (habituation, 98 volumes; acquisition training, 354 volumes; extinction training, 350 volumes; recall test, 182 volumes; reacquisition, reextinction, and unexpected US phases; 938 volumes). Imaging parameters were as follows: TR/TE, 1620/20 ms; flip angle, 11°; phase encoding acceleration factor, 2; 3D acceleration factor, 3; slice partial Fourier factor, 7/8; acquisition matrix, 140×140; number of slices, 96.

After acquisition training on day 1, a transversal QSM ASPIRE sequence was acquired (Eckstein et al., 2018; Eckstein et al., 2021) with an isotropic voxel size of 0.7 mm. Further imaging parameters were as follows: TR/TE1/TE2/TE3/TE4/TE5, 28/5/10/15/20/25 ms; flip angle 15°; phase encoding acceleration factor, 3; slice partial Fourier factor, 6/8; acquisition matrix, 320×280; number of slices, 256; TA, 10:52 min.

After the acquisition of the QSM ASPIRE sequence, a sagittal MP2RAGE sequence, including fat navigators (Gallichan and Marques, 2017; Marques et al., 2010; Gallichan et al., 2016) was acquired with an isotropic voxel size of 0.75 mm. Further imaging parameters were as follows: TR/TE, 6000/1.85 ms; TI1/TI2, 800/2750 ms; flip angles 1/2, 4°/5°; phase encoding acceleration factor, 3; acquisition matrix, 340×340; number of slices, 256; TA, 14:58 min.

Image processing

Motion correction of MP2RAGE volumes, including fat navigators was performed using offline reconstruction with the Retro-MoCo toolbox for MATLAB provided by David Gallichan (version 23 0.9.0 dev, https://github.com/dgallichan/retroMoCoBox, Gallichan, 2025). A T1 map was generated from the motion corrected MP2RAGE with the MP2RAGE-utils package implemented in MATLAB (release 1.0, https://github.com/srikash/MP2RAGE-utils, Kashyap, 2021). The MP2RAGE was normalized to MNI-space using the CAT12 (CAT12, release 1450) (Gaser et al., 2024) toolbox in SPM12 (Wellcome Department of Cognitive Neurology, London, UK).

Functional MRI volumes were brain extracted with BET (Brain Extraction Tool) (Jenkinson et al., 2005) in FSL (Release 6.0.1, RRID:SCR_002823, Centre for Functional MRI of the Brain, Oxford, UK). Motion and distortion correction was performed for each fMRI run with ANTs (Version 2.3.5, RRID:SCR_004757, University of Pennsylvania, Philadelphia, USA) (Tustison et al., 2014). A 5 volume fMRI sequence preceding each run with phase encoding in the opposite direction was used for the distortion correction. All volumes were coregistered to a T1 map derived from the acquired MP2RAGE with ANTs. The CAT12 processing of the MP2RAGE was used to normalize fMRI data. A 4.5 mm Gaussian kernel was used for smoothing. Motion nuisance regressors (three translations, three rotations) were derived from ANTs affine transformation matrices output.

fMRI analysis

The first level analysis was done in MNI space, defining conditions for both the CS and US window (i.e. CS+, CS-, US post CS+, no US post CS+, no US post CS-) separately for each phase. Regressors were calculated from the conditions in SPM12 by convolving a canonical hemodynamic response function with events, which were modeled as delta functions in an event-based design (i.e. the duration of each event is 0 s). Specifically, CS events were time-locked to CS onset (appearance of the visual stimulus), whereas US related events were time-locked to CS offset (6 s after CS onset), which coincided with either delivery or omission of the US. This event-related separation allowed us to test CS related contrasts at CS onset, including CS differentiation (CS+ > CS-). US-related contrasts were tested at CS offset and included contrasts reflecting unexpected omission of the US (e.g. first 3 no US post CS+ > no US post CS-) and US presentation (US post CS+ > no US post CS-). Beta weights were fitted for each regressor by a general linear model for each voxel fMRI time series. Contrasts were computed as linear combinations of beta weights. Values in contrast maps were converted to t-statistic values in first level analysis. Contrast maps are combined and used in a one-sample t-test for second level analysis in SPM12. Tests for significance (p<0.05) were done after threshold-free cluster enhancement (TFCE toolbox in SPM12, R174, http://dbm.neuro.uni-jena.de/tfce/) and family-wise error (FWE) corrections. Any trend-level results refer to significance tests without TFCE and FWE corrections (p<0.05). For the CB, results were visualized on cerebellar flatmaps using the SUIT toolbox in SPM12 (Diedrichsen and Zotow, 2015). For the cerebellar nuclei and VTA, results were visualized in coronal slices and 3D renderings generated in MRIcroGL (Rorden, 2025) using custom probabilistic atlases (described in ‘Volume of interest definition’). To acquire cerebellar anatomical region labels, activation maps were projected onto the SUIT atlas volume (Cerebellum-SUIT.nii, Diedrichsen, 2006).

Event-based analysis

In addition to main effects of conditions (e.g. during acquisition training), events were additionally separated over time for each experimental phase. For fear acquisition training, extinction training, recall test, reacquisition, reextinction, and the unexpected US phase, events were grouped in two equal-size blocks representing early and late halves of those phases (e.g. the eight first CS+ trials of fear acquisition training correspond to early acquisition and the eight last CS+ trials of fear acquisition training correspond to late acquisition), with equal numbers of CS+ and CS− trials in each block. Trials with unexpected events—such as unexpected US presentations or omissions—were treated as separate single-trial regressors. These included the initial three trials of acquisition training, extinction training, recall test, reacquisition, and reextinction and specific events, such as the three unexpected US omissions in reacquisition. When CS+ trials were modeled as single trials in the GLM, CS- trials were modeled in the same way to maintain a consistent model structure. This modeling choice did not translate into paired or numerically matched CS+ and CS− trials being used in the event-based contrasts. Rather, event-based contrasts for unexpected US omissions contrasted the first three unexpected omissions following CS+ against all ‘expected’ omissions following CS- (e.g. [first 3 no US post CS+ > no US post CS-]).

Parametric modulation analysis

Parametric modulation of fMRI data was performed with prediction and prediction error values in all learning phases derived from our computational model. CS+ and CS- events were modulated with prediction values, while US omission and presentation events were modulated with prediction error values (Figure 9). Events were not further divided within phases (e.g. no early or late halves or separation of the first three trials). Modulations were done separately for each phase, except for reacquisition, reextinction, and the unexpected US phase which were modulated together as they were part of the same fMRI run.

Volumes of interest (VOI) definition

A global conjunction between unexpected US omission parametric modulation contrasts during extinction, the recall test, and reacquisition was multiplied with a cerebellum mask and showed a region in lobule VI and Crus I. This region was used as a CB (lobule VI and Crus I; Figure 8—figure supplement 2) VOI. DCN and VTA VOIs were drawn for each participant. Drawing of the DCN (i.e. the left and right dentate, globose, emboliform, and fastigial nuclei) was done by an expert annotator using both MP2RAGE and QSM information in ITK-SNAP (version 3.8.0) (Yushkevich et al., 2006). The DCN were combined into a single bihemispheric VOI. Drawing of the VTA was done by adjusting estimated masks with both MP2RAGE and QSM information, using the red nuclei, substantia nigra, the cerebral aqueduct, and interpeduncular fossa as landmarks. Initial masks were estimated for the VTA by inversely transforming the probabilistic atlas by Trutti et al., 2021 thresholded at a probability of p>0.4. All drawn masks were transformed into MNI space. After transformation to MNI space, a probabilistic atlas was generated by averaging binary masks for each structure across participants. The probabilistic atlas was thresholded so that the volume of each structure matched the average volume across participants. For drawing of globose, emboliform, and fastigial nuclei, the thresholded atlas was inversely transformed from MNI to T1 space in order to add missing nuclei for each participant with lower quality QSM data. The VTA VOI was used as a seed region in the PPI analysis. In addition, the CB, DCN, and VTA VOIs were used as nodes in the 2-node DCM networks. The DCN and VTA VOIs were used as masks for displaying fMRI activations in coronal slices and 3D renderings.

Functional connectivity: PPI

For the PPI analysis, the VTA VOI was used to extract time series with SPM12 in MNI space. Time series were used as physiological regressors, and psychophysiological regressors were added for each condition using the gPPI toolbox (McLaren et al., 2012). Both first- and second-level analyses were performed with the same contrasts as defined in the parametric modulation (e.g. no US × prediction error during extinction) analysis. To summarize connectivity in the parametric modulation analysis, all contrasts related to unexpected US omissions were combined in a single contrast within the gPPI toolbox.

Functional connectivity: DCM

To calculate DCMs, the CB, and VTA VOIs were used to extract time series in a concatenated version of the event-based and parametric modulation models using SPM12 (Friston et al., 2003; Stephan et al., 2010; Zeidman et al., 2019a). Each condition was defined as input to the CB and VTA nodes (i.e. CS+, CS-, US post CS+, no US post CS+, no US post CS-), while each prediction or prediction error related event-based contrast (i.e. CS+ > CS-, first 3 no US post CS+ > no US post CS-) or parametric modulation (i.e. no US prediction error) could modulate the reciprocal intrinsic connections. Recurrent connections were active but were not modulated. Parametric Empirical Bayes (PEB) was used for analysis on the second level (Zeidman et al., 2019b). All DCM and PEB analyses were done within SPM12.

Appendix 1

SCRs for first-three and last-three trial analysis

To assess the robustness of behavioral results when restricting analyses to a small number of trials, we additionally re-binned trials into blocks comprising the first three and last three trials of each phase, instead of halves of each phase. Overall, the results were largely consistent with the primary analysis. The corresponding statistical results are summarized in Supplementary file 2.

PSRs for first-three and last-three trial analysis

Data availability

The anonymized dataset generated and analysed during the current study has been deposited in the ReSeeD Repository (SFB 1280 Extinction Learning, Ruhr University Bochum). Public access is provided via: https://doi.org/10.60517/1831cq60q. MRI preprocessing, fMRI analyses, and visualizations were performed using in-house pipelines based on ANTs, CAT12, SUIT, SPM12, and MATLAB. Custom scripts used for event extraction, model specification, statistical analyses, and figure generation are publicly available at: https://github.com/spennioo/code-nio-et-al-elife-2026, copy archived at Nio, 2026.

The following data sets were generated

1. Nio E
2. Pais Pereira P
3. Petrenko M
4. Doubliez A
5. Ernst TM
6. Timmann D
(2026) ReSeeD (Research Data Management System)
A three-day 7T fMRI study of Fear Conditioning, Extinction, and Recall in healthy adults aged 20-35 years.

https://doi.org/10.60517/1831cq60q

References

(2022) VTA-projecting cerebellar neurons mediate stress-dependent depression-like behaviors
eLife 11:e72981.

https://doi.org/10.7554/eLife.72981
- PubMed
- Google Scholar
Book
1. Barto AG
(2018) Adaptive critics and the basal ganglia
In: Houk JC, Davis JL, Beiser DG, editors. Models of Information Processing in the Basal Ganglia. The MIT Press. pp. 215–232.

https://doi.org/10.7551/mitpress/4708.003.0018
- Google Scholar
1. Bartolomeo P
2. Seidel Malkinson T
(2019) Hemispheric lateralization of attention processes in the human brain
Current Opinion in Psychology 29:90–96.

https://doi.org/10.1016/j.copsyc.2018.12.023
- PubMed
- Google Scholar
(2009) Greenhouse–geisser adjustment and the ANOVA-type statistic: cousins or twins?
The American Statistician 63:239–246.

https://doi.org/10.1198/tast.2009.08187
- Google Scholar
1. Batsikadze G
2. Diekmann N
3. Ernst TM
4. Klein M
5. Maderwald S
6. Deuschl C
7. Merz CJ
8. Cheng S
9. Quick HH
10. Timmann D
(2022) The cerebellum contributes to context-effects during fear extinction learning: a 7T fMRI study
NeuroImage 253:119080.

https://doi.org/10.1016/j.neuroimage.2022.119080
- PubMed
- Google Scholar
1. Batsikadze G
2. Pakusch J
3. Klein M
4. Ernst TM
5. Thieme A
6. Nicksirat SA
7. Steiner KM
8. Nio E
9. Genc E
10. Maderwald S
11. Deuschl C
12. Merz CJ
13. Quick HH
14. Mark MD
15. Timmann D
(2024) Mild deficits in fear learning: evidence from humans and mice with cerebellar cortical degeneration
eNeuro 11:ENEURO.0365-23.2023.

https://doi.org/10.1523/ENEURO.0365-23.2023
- PubMed
- Google Scholar
Book
1. Boucsein W
(2012) Electrodermal Activity
Springer.

https://doi.org/10.1007/978-1-4614-1126-0
- Google Scholar
1. Bouton ME
(2002) Context, ambiguity, and unlearning: sources of relapse after behavioral extinction
Biological Psychiatry 52:976–986.

https://doi.org/10.1016/s0006-3223(02)01546-9
- PubMed
- Google Scholar
Book
(2002)
Nonparametric Analysis of Longitudinal Data in Factorial Experiments

John Wiley & Sons.
- Google Scholar
1. Cai LX
2. Pizano K
3. Gundersen GW
4. Hayes CL
5. Fleming WT
6. Holt S
7. Cox JM
8. Witten IB
(2020) Distinct signals in medial and lateral VTA dopamine neurons modulate fear extinction at different times
eLife 9:e54936.

https://doi.org/10.7554/eLife.54936
- PubMed
- Google Scholar
(2018) Definition and characterization of an extended multiple-demand network
NeuroImage 165:138–147.

https://doi.org/10.1016/j.neuroimage.2017.10.020
- PubMed
- Google Scholar
1. Carta I
2. Chen CH
3. Schott AL
4. Dorizan S
5. Khodakhah K
(2019) Cerebellar modulation of the reward circuitry and social behavior
Science 363:eaav0581.

https://doi.org/10.1126/science.aav0581
- PubMed
- Google Scholar
(2022) Meta-analysis of human prediction error for incentives, perception, cognition, and action
Neuropsychopharmacology 47:1339–1349.

https://doi.org/10.1038/s41386-021-01264-3
- PubMed
- Google Scholar
1. Craske MG
2. Liao B
3. Brown L
4. Vervliet B
(2012) Role of inhibition in exposure therapy
Journal of Experimental Psychopathology 3:322–345.

https://doi.org/10.5127/jep.026511
- Google Scholar
(2008) BOLD responses reflecting dopaminergic signals in the human ventral tegmental area
Science 319:1264–1267.

https://doi.org/10.1126/science.1150605
- PubMed
- Google Scholar
(2017) Comparing functional MRI protocols for small, iron-rich basal ganglia nuclei such as the subthalamic nucleus at 7 T and 3 T
Human Brain Mapping 38:3226–3248.

https://doi.org/10.1002/hbm.23586
- PubMed
- Google Scholar
1. Diedrichsen J
(2006) A spatially unbiased atlas template of the human cerebellum
NeuroImage 33:127–138.

https://doi.org/10.1016/j.neuroimage.2006.05.056
- PubMed
- Google Scholar
1. Diedrichsen J
2. Maderwald S
3. Küper M
4. Thürling M
5. Rabe K
6. Gizewski ER
7. Ladd ME
8. Timmann D
(2011) Imaging the deep cerebellar nuclei: a probabilistic atlas and normalization procedure
NeuroImage 54:1786–1794.

https://doi.org/10.1016/j.neuroimage.2010.10.035
- PubMed
- Google Scholar
1. Diedrichsen J
2. Zotow E
(2015) Surface-based display of volume-averaged cerebellar imaging data
PLOS ONE 10:e0133402.

https://doi.org/10.1371/journal.pone.0133402
- PubMed
- Google Scholar
(2023) The cerebellum and fear extinction: evidence from rodent and human studies
Frontiers in Systems Neuroscience 17:1166166.

https://doi.org/10.3389/fnsys.2023.1166166
- PubMed
- Google Scholar
1. Doubliez A
2. Köster K
3. Müntefering L
4. Nio E
5. Diekmann N
6. Thieme A
7. Albayrak B
8. Nicksirat SA
9. Erdlenbruch F
10. Batsikadze G
11. Ernst TM
12. Cheng S
13. Merz CJ
14. Timmann D
(2025) Dopaminergic drugs modulate fear extinction-related processes in humans, but effects are mild
Brain Communications 7:fcaf333.

https://doi.org/10.1093/braincomms/fcaf333
- PubMed
- Google Scholar
(2009) Functional imaging of the human dopaminergic midbrain
Trends in Neurosciences 32:321–328.

https://doi.org/10.1016/j.tins.2009.02.005
- PubMed
- Google Scholar
1. Eckstein K
2. Dymerska B
3. Bachrata B
4. Bogner W
5. Poljanc K
6. Trattnig S
7. Robinson SD
(2018) Computationally efficient combination of multi-channel phase data from multi-echo acquisitions (ASPIRE)
Magnetic Resonance in Medicine 79:2996–3006.

https://doi.org/10.1002/mrm.26963
- PubMed
- Google Scholar
1. Eckstein K
2. Bachrata B
3. Hangel G
4. Widhalm G
5. Enzinger C
6. Barth M
7. Trattnig S
8. Robinson SD
(2021) Improved susceptibility weighted imaging at ultra-high field using bipolar multi-echo acquisition and optimized image processing: CLEAR-SWI
NeuroImage 237:118175.

https://doi.org/10.1016/j.neuroimage.2021.118175
- PubMed
- Google Scholar
1. Ernst TM
2. Brol AE
3. Gratz M
4. Ritter C
5. Bingel U
6. Schlamann M
7. Maderwald S
8. Quick HH
9. Merz CJ
10. Timmann D
(2019) The cerebellum is involved in processing of predictions and prediction errors in a fear conditioning paradigm
eLife 8:e46831.

https://doi.org/10.7554/eLife.46831
- PubMed
- Google Scholar
1. Esser R
2. Korn CW
3. Ganzer F
4. Haaker J
(2021) L-DOPA modulates activity in the vmPFC, nucleus accumbens, and VTA during threat extinction learning in humans
eLife 10:e65280.

https://doi.org/10.7554/eLife.65280
- PubMed
- Google Scholar
1. Faul L
2. Stjepanović D
3. Stivers JM
4. Stewart GW
5. Graner JL
6. Morey RA
7. LaBar KS
(2020) Proximal threats promote enhanced acquisition and persistence of reactive fear-learning circuits
PNAS 117:16678–16689.

https://doi.org/10.1073/pnas.2004258117
- Google Scholar
1. Friston KJ
2. Buechel C
3. Fink GR
4. Morris J
5. Rolls E
6. Dolan RJ
(1997) Psychophysiological and modulatory interactions in neuroimaging
NeuroImage 6:218–229.

https://doi.org/10.1006/nimg.1997.0291
- PubMed
- Google Scholar
(2003) Dynamic causal modelling
NeuroImage 19:1273–1302.

https://doi.org/10.1016/s1053-8119(03)00202-7
- PubMed
- Google Scholar
(2016) Neural signatures of human fear conditioning: an updated and extended meta-analysis of fMRI studies
Molecular Psychiatry 21:500–508.

https://doi.org/10.1038/mp.2015.88
- PubMed
- Google Scholar
(2022) Non-linear frequency dependence of neurovascular coupling in the cerebellar cortex implies vasodilation-vasoconstriction competition
Cells 11:1047.

https://doi.org/10.3390/cells11061047
- PubMed
- Google Scholar
1. Gainotti G
(2019) Emotions and the right hemisphere: can new data clarify old models?
The Neuroscientist 25:258–270.

https://doi.org/10.1177/1073858418785342
- PubMed
- Google Scholar
(2016) Retrospective correction of involuntary microscopic head movement using highly accelerated fat image navigators (3D FatNavs) at 7T
Magnetic Resonance in Medicine 75:1030–1039.

https://doi.org/10.1002/mrm.25670
- PubMed
- Google Scholar
1. Gallichan D
2. Marques JP
(2017) Optimizing the acceleration and resolution of three-dimensional fat image navigators for high-resolution motion correction at 7T
Magnetic Resonance in Medicine 77:547–558.

https://doi.org/10.1002/mrm.26127
- PubMed
- Google Scholar
Software
1. Gallichan D
(2025) RetroMoCoBox
Github.

https://github.com/dgallichan/retroMoCoBox
(2024) CAT: a computational anatomy toolbox for the analysis of structural MRI data
GigaScience 13:giae049.

https://doi.org/10.1093/gigascience/giae049
- PubMed
- Google Scholar
1. Grupe DW
2. Nitschke JB
(2013) Uncertainty and anticipation in anxiety: an integrated neurobiological and psychological perspective
Nature Reviews. Neuroscience 14:488–501.

https://doi.org/10.1038/nrn3524
- PubMed
- Google Scholar
Conference
(2024)
Exploring monosynaptic midbrain-to-deep cerebellar nuclei projections

Abstracts, Neuroscience (Society for Neuroscience, Chicago, IL).
- Google Scholar
(2018) Triple representation of language, working memory, social and emotion processing in the cerebellum: convergent evidence from task and seed-based resting-state fMRI analyses in a single large cohort
NeuroImage 172:437–449.

https://doi.org/10.1016/j.neuroimage.2018.01.082
- PubMed
- Google Scholar
(2020) Brain lateralization: a comparative perspective
Physiological Reviews 100:1019–1063.

https://doi.org/10.1152/physrev.00006.2019
- PubMed
- Google Scholar
1. Heffley W
2. Hull C
(2019) Classical conditioning drives learned reward prediction signals in climbing fibers across the lateral cerebellum
eLife 8:e46764.

https://doi.org/10.7554/eLife.46764
- PubMed
- Google Scholar
1. Henry JD
2. Crawford JR
(2005) The short-form version of the Depression Anxiety Stress Scales (DASS-21): construct validity and normative data in a large non-clinical sample
The British Journal of Clinical Psychology 44:227–239.

https://doi.org/10.1348/014466505X29657
- PubMed
- Google Scholar
1. Hull C
(2020) Prediction signals in the cerebellum: beyond supervised motor learning
eLife 9:e54073.

https://doi.org/10.7554/eLife.54073
- PubMed
- Google Scholar
1. Ikai Y
2. Takada M
3. Shinonaga Y
4. Mizuno N
(1992) Dopaminergic and non-dopaminergic neurons in the ventral tegmental area of the rat project, respectively, to the cerebellar cortex and deep cerebellar nuclei
Neuroscience 51:719–728.

https://doi.org/10.1016/0306-4522(92)90310-x
- PubMed
- Google Scholar
1. Inoue L
2. Ernst TM
3. Ferber II
4. Merz CJ
5. Timmann D
6. Batsikadze G
(2020) Interaction of fear conditioning with eyeblink conditioning supports the sensory gating hypothesis of the amygdala in men
eNeuro 7:1–15.

https://doi.org/10.1523/ENEURO.0128-20.2020
- PubMed
- Google Scholar
Conference
(2005)
BET2: MR-based estimation of brain, skull and scalp surfaces

Eleventh Annual Meeting of the Organization for Human Brain Mapping.
- Google Scholar
(2020) Temporal dynamics of conditioned skin conductance and pupillary responses during fear acquisition and extinction
International Journal of Psychophysiology 147:93–99.

https://doi.org/10.1016/j.ijpsycho.2019.11.006
- PubMed
- Google Scholar
(2018) Dopamine neurons reflect the uncertainty in fear generalization
Neuron 100:916–925.

https://doi.org/10.1016/j.neuron.2018.09.028
- PubMed
- Google Scholar
(2019) A dopaminergic basis for fear extinction
Trends in Cognitive Sciences 23:274–277.

https://doi.org/10.1016/j.tics.2019.01.013
- PubMed
- Google Scholar
Software
1. Kashyap S
(2021) MP2RAGE-utils
Github.

https://github.com/srikash/MP2RAGE-utils
(2014) Cerebellar contributions to different phases of visceral aversive extinction learning
Cerebellum 13:1–8.

https://doi.org/10.1007/s12311-013-0512-9
- PubMed
- Google Scholar
(2014) A meta-analysis of cerebellar contributions to higher cognition from PET and fMRI studies
Human Brain Mapping 35:593–615.

https://doi.org/10.1002/hbm.22194
- PubMed
- Google Scholar
(2017) Attentional bias for uncertain cues of shock in human fear conditioning: evidence for attentional learning theory
Frontiers in Human Neuroscience 11:266.

https://doi.org/10.3389/fnhum.2017.00266
- PubMed
- Google Scholar
(2019) Predictive and reactive reward signals conveyed by climbing fiber inputs to cerebellar Purkinje cells
Nature Neuroscience 22:950–962.

https://doi.org/10.1038/s41593-019-0381-8
- PubMed
- Google Scholar
1. Kostadinov D
2. Häusser M
(2022) Reward signals in the cerebellum: origins, targets, and functional implications
Neuron 110:1290–1303.

https://doi.org/10.1016/j.neuron.2022.02.015
- PubMed
- Google Scholar
1. Kret ME
2. Sjak-Shie EE
(2019) Preprocessing pupil size data: guidelines and code
Behavior Research Methods 51:1336–1342.

https://doi.org/10.3758/s13428-018-1075-y
- PubMed
- Google Scholar
(2023) The human cerebellum in reward anticipation and outcome processing: an activation likelihood estimation meta-analysis
Neuroscience and Biobehavioral Reviews 149:105171.

https://doi.org/10.1016/j.neubiorev.2023.105171
- PubMed
- Google Scholar
1. Labrenz F
2. Spisák T
3. Ernst TM
4. Gomes CA
5. Quick HH
6. Axmacher N
7. Elsenbruch S
8. Timmann D
(2022) Temporal dynamics of fMRI signal changes during conditioned interoceptive pain-related fear and safety acquisition and extinction
Behavioural Brain Research 427:113868.

https://doi.org/10.1016/j.bbr.2022.113868
- PubMed
- Google Scholar
(2015) Deep learning
Nature 521:436–444.

https://doi.org/10.1038/nature14539
- PubMed
- Google Scholar
1. Li C
2. Saliba NB
3. Martin H
4. Losurdo NA
5. Kolahdouzan K
6. Siddiqui R
7. Medeiros D
8. Li W
(2023) Purkinje cell dopaminergic inputs to astrocytes regulate cerebellar-dependent behavior
Nature Communications 14:37319-w.

https://doi.org/10.1038/s41467-023-37319-w
- Google Scholar
Conference
1. Lin LJJ
(1991) Self-improvement based on reinforcement learning, planning and teaching
Proceedings of the 8th International Workshop on Machine Learning, ICML. pp. 323–327.

https://doi.org/10.1016/B978-1-55860-200-7.50067-2
- Google Scholar
1. Linnman C
2. Zeidan MA
3. Furtak SC
4. Pitman RK
5. Quirk GJ
6. Milad MR
(2012) Resting amygdala and medial prefrontal metabolism predicts functional activation of the fear extinction circuit
The American Journal of Psychiatry 169:415–423.

https://doi.org/10.1176/appi.ajp.2011.10121780
- PubMed
- Google Scholar
1. Liu G
2. Zeng G
3. Wang F
4. Rotshtein P
5. Peng K
6. Sui J
(2018) Praising others differently: neuroanatomical correlates to individual differences in trait gratitude and elevation
Social Cognitive and Affective Neuroscience 13:1225–1234.

https://doi.org/10.1093/scan/nsy093
- PubMed
- Google Scholar
(2001) Neurophysiological investigation of the basis of the fMRI signal
Nature 412:150–157.

https://doi.org/10.1038/35084005
- PubMed
- Google Scholar
1. Logothetis NK
2. Wandell BA
(2004) Interpreting the BOLD signal
Annual Review of Physiology 66:735–769.

https://doi.org/10.1146/annurev.physiol.66.082602.092845
- PubMed
- Google Scholar
1. Logothetis NK
(2008) What we can do and what we cannot do with fMRI
Nature 453:869–878.

https://doi.org/10.1038/nature06976
- PubMed
- Google Scholar
1. Lonsdorf TB
2. Menz MM
3. Andreatta M
4. Fullana MA
5. Golkar A
6. Haaker J
7. Heitland I
8. Hermann A
9. Kuhn M
10. Kruse O
11. Meir Drexler S
12. Meulders A
13. Nees F
14. Pittig A
15. Richter J
16. Römer S
17. Shiban Y
18. Schmitz A
19. Straube B
20. Vervliet B
21. Wendt J
22. Baas JMP
23. Merz CJ
(2017) Don’t fear “fear conditioning”: methodological considerations for the design and analysis of studies on human fear acquisition, extinction, and return of fear
Neuroscience and Biobehavioral Reviews 77:247–285.

https://doi.org/10.1016/j.neubiorev.2017.02.026
- PubMed
- Google Scholar
1. Luo R
2. Uematsu A
3. Weitemier A
4. Aquili L
5. Koivumaa J
6. McHugh TJ
7. Johansen JP
(2018) A dopaminergic switch for fear to safety transitions
Nature Communications 9:2483.

https://doi.org/10.1038/s41467-018-04784-7
- PubMed
- Google Scholar
(2017) Granular layer neurons control cerebellar neurovascular coupling through an NMDA Receptor/NO-dependent system
The Journal of Neuroscience 37:1340–1351.

https://doi.org/10.1523/JNEUROSCI.2025-16.2016
- PubMed
- Google Scholar
(2010) MP2RAGE, a self bias-field corrected sequence for improved segmentation and T1-mapping at high field
NeuroImage 49:1271–1281.

https://doi.org/10.1016/j.neuroimage.2009.10.002
- PubMed
- Google Scholar
(2018) Safe and sensible preprocessing and baseline correction of pupil-size data
Behavior Research Methods 50:94–106.

https://doi.org/10.3758/s13428-017-1007-2
- PubMed
- Google Scholar
1. McLaren DG
2. Ries ML
3. Xu G
4. Johnson SC
(2012) A generalized form of context-dependent psychophysiological interactions (gPPI): a comparison to standard approaches
NeuroImage 61:1277–1286.

https://doi.org/10.1016/j.neuroimage.2012.03.068
- PubMed
- Google Scholar
1. Milad MR
2. Quirk GJ
(2012) Fear extinction as a model for translational neuroscience: ten years of progress
Annual Review of Psychology 63:129–151.

https://doi.org/10.1146/annurev.psych.121208.131631
- PubMed
- Google Scholar
(1996) A framework for mesencephalic dopamine systems based on predictive Hebbian learning
The Journal of Neuroscience 16:1936–1947.

https://doi.org/10.1523/JNEUROSCI.16-05-01936.1996
- PubMed
- Google Scholar
1. Moriya S
2. Yamashita A
3. Kawashima S
4. Nishi R
5. Yamanaka A
6. Kuwaki T
(2018) Acute aversive stimuli rapidly increase the activity of ventral tegmental area dopamine neurons in awake mice
Neuroscience 386:16–23.

https://doi.org/10.1016/j.neuroscience.2018.06.027
- PubMed
- Google Scholar
1. Mumford JA
(2012) A power calculation guide for FMRI studies
Social Cognitive and Affective Neuroscience 7:738–742.

https://doi.org/10.1093/scan/nss059
- PubMed
- Google Scholar
(2024) A hierarchical atlas of the human cerebellum for functional precision mapping
Nature Communications 15:8376.

https://doi.org/10.1038/s41467-024-52371-w
- PubMed
- Google Scholar
(2023) 7T fMRI study on extinction learning in young and healthy participants
Center for Open Science 1:2PXWE.

https://doi.org/10.17605/OSF.IO/2PXWE
- Google Scholar
Software
1. Nio E
(2026) Code-nio-et-al-elife-2026, version swh:1:rev:3898cde630087c48adc7ccae2537a57fd889ab26
Software Heritage.

https://archive.softwareheritage.org/swh:1:dir:b690712efb5b3dc70edc6cb80f0932aecd95605a;origin=https://github.com/spennioo/code-nio-et-al-elife-2026;visit=swh:1:snp:bb8b394fb65b09908a4feb47303c1dff07439518;anchor=swh:1:rev:3898cde630087c48adc7ccae2537a57fd889ab26
(2012) NparLD: an R software package for the nonparametric analysis of longitudinal data in factorial experiments
Journal of Statistical Software 50:1–23.

https://doi.org/10.18637/jss.v050.i12
- Google Scholar
1. Oldfield RC
(1971) The assessment and analysis of handedness: the Edinburgh inventory
Neuropsychologia 9:97–113.

https://doi.org/10.1016/0028-3932(71)90067-4
- PubMed
- Google Scholar
(2012) Tools of the trade: psychophysiological interactions and functional connectivity
Social Cognitive and Affective Neuroscience 7:604–609.

https://doi.org/10.1093/scan/nss055
- PubMed
- Google Scholar
Software
1. Otto T
2. Wolf OT
3. Merz CJ
(2023) EDA-analysis app, version 5.11
Zenodo.

https://doi.org/10.5281/zenodo.7965376
1. Pakusch J
2. Nio E
3. Grosch T
4. Ernst TM
5. Batsikadze G
6. Güntürkün O
7. Timmann D
8. Mark MD
(2025) Evidence for lateralization of fear emotions in the cerebellum
Journal of Neurology 272:435.

https://doi.org/10.1007/s00415-025-13183-0
- PubMed
- Google Scholar
(2020) Dopamine: from prediction error to psychotherapy
Translational Psychiatry 10:164.

https://doi.org/10.1038/s41398-020-0814-x
- PubMed
- Google Scholar
(2018) A high-resolution probabilistic in vivo atlas of human subcortical brain nuclei
Scientific Data 5:180063.

https://doi.org/10.1038/sdata.2018.63
- PubMed
- Google Scholar
1. Pearce JM
2. Hall G
(1980) A model for Pavlovian learning: variations in the effectiveness of conditioned but not of unconditioned stimuli
Psychological Review 87:532–552.

https://doi.org/10.1037/0033-295X.87.6.532
- PubMed
- Google Scholar
1. Pierce JE
2. Thomasson M
3. Voruz P
4. Selosse G
5. Péron J
(2023) Explicit and implicit emotion processing in the cerebellum: a meta-analysis and systematic review
Cerebellum 22:852–864.

https://doi.org/10.1007/s12311-022-01459-4
- PubMed
- Google Scholar
1. Pisano TJ
2. Dhanerawala ZM
3. Kislin M
4. Bakshinskaya D
5. Engel EA
6. Hansen EJ
7. Hoag AT
8. Lee J
9. de Oude NL
10. Venkataraju KU
11. Verpeut JL
12. Hoebeek FE
13. Richardson BD
14. Boele H-J
15. Wang SS-H
(2021) Homologous organization of cerebellar pathways to sensory, motor, and associative forebrain
Cell Reports 36:109721.

https://doi.org/10.1016/j.celrep.2021.109721
- PubMed
- Google Scholar
1. Ploghaus A
2. Tracey I
3. Gati JS
4. Clare S
5. Menon RS
6. Matthews PM
7. Rawlins JN
(1999) Dissociating pain from its anticipation in the human brain
Science 284:1979–1981.

https://doi.org/10.1126/science.284.5422.1979
- PubMed
- Google Scholar
1. Ploghaus A
2. Tracey I
3. Clare S
4. Gati JS
5. Rawlins JNP
6. Matthews PM
(2000) Learning about pain: the neural substrate of the prediction error for aversive events
PNAS 97:9281–9286.

https://doi.org/10.1073/pnas.160266497
- Google Scholar
(2004) Prediction error for free monetary reward in the human prefrontal cortex
NeuroImage 23:777–786.

https://doi.org/10.1016/j.neuroimage.2004.07.028
- PubMed
- Google Scholar
Book
1. Rescorla RA
2. Wagner AR
(1972)
A theory of pavlovian conditioning: the effectiveness of reinforcement and non-reinforcement

In: Black AH, Prokasy WF, editors. Classical Conditioning: Current Research and Theory. Appleton-Century-Crofts. pp. 64–99.
- Google Scholar
1. Rorden C
(2025) MRIcroGL: voxel-based visualization for neuroimaging
Nature Methods 22:1613–1614.

https://doi.org/10.1038/s41592-025-02763-7
- PubMed
- Google Scholar
1. Rouhani N
2. Niv Y
(2021) Signed and unsigned reward prediction errors dynamically enhance learning and memory
eLife 10:e61077.

https://doi.org/10.7554/eLife.61077
- PubMed
- Google Scholar
(1986) Learning representations by back-propagating errors
Nature 323:533–536.

https://doi.org/10.1038/323533a0
- Google Scholar
(2018) Dopamine neurons drive fear extinction learning by signaling the omission of expected aversive outcomes
eLife 7:e38818.

https://doi.org/10.7554/eLife.38818
- PubMed
- Google Scholar
(2023) Functional architecture of dopamine neurons driving fear extinction learning
Neuron 111:3854–3870.

https://doi.org/10.1016/j.neuron.2023.08.025
- PubMed
- Google Scholar
Preprint
1. Schaul T
2. Quan J
3. Antonoglou I
4. Silver D
(2016) Prioritized experience replay
arXiv.

https://arxiv.org/abs/1511.05952
- Google Scholar
1. Schott BH
2. Minuzzi L
3. Krebs RM
4. Elmenhorst D
5. Lang M
6. Winz OH
7. Seidenbecher CI
8. Coenen HH
9. Heinze H-J
10. Zilles K
11. Düzel E
12. Bauer A
(2008) Mesolimbic functional magnetic resonance imaging activations during reward anticipation correlate with reward-related ventral striatal dopamine release
The Journal of Neuroscience 28:14311–14319.

https://doi.org/10.1523/JNEUROSCI.2058-08.2008
- PubMed
- Google Scholar
1. Schultz W
2. Dickinson A
(2000) Neuronal coding of prediction errors
Annual Review of Neuroscience 23:473–500.

https://doi.org/10.1146/annurev.neuro.23.1.473
- PubMed
- Google Scholar
1. Shah DA
2. Madden LV
(2004) Nonparametric analysis of ordinal data
Phytopathology 94:33–43.

https://doi.org/10.1094/PHYTO.2004.94.1.33
- PubMed
- Google Scholar
(2024) Selective recruitment of the cerebellum evidenced by task-dependent gating of inputs
eLife 13:RP96386.

https://doi.org/10.7554/eLife.96386
- PubMed
- Google Scholar
1. Snider RS
2. Maiti A
(1976) Cerebellar contributions to the Papez circuit
Journal of Neuroscience Research 2:133–146.

https://doi.org/10.1002/jnr.490020204
- PubMed
- Google Scholar
(2010) Ten simple rules for dynamic causal modeling
NeuroImage 49:3099–3109.

https://doi.org/10.1016/j.neuroimage.2009.11.015
- PubMed
- Google Scholar
1. Stirnberg R
2. Stöcker T
(2021) Segmented K-space blipped-controlled aliasing in parallel imaging for high spatiotemporal resolution EPI
Magnetic Resonance in Medicine 85:1540–1551.

https://doi.org/10.1002/mrm.28486
- PubMed
- Google Scholar
1. Stoodley CJ
2. Schmahmann JD
(2009) Functional topography in the human cerebellum: a meta-analysis of neuroimaging studies
NeuroImage 44:489–501.

https://doi.org/10.1016/j.neuroimage.2008.08.039
- PubMed
- Google Scholar
Book
1. Sutton RS
2. Barto AG
(2018)
Reinforcement Learning: An Introduction

MIT Press.
- Google Scholar
(2012) Quantitative assessment of the effects of high-permittivity pads in 7 Tesla MRI of the brain
Magnetic Resonance in Medicine 67:1285–1293.

https://doi.org/10.1002/mrm.23108
- PubMed
- Google Scholar
(2021) A probabilistic atlas of the human ventral tegmental area (VTA) based on 7 Tesla MRI data
Brain Structure & Function 226:1155–1167.

https://doi.org/10.1007/s00429-021-02231-w
- PubMed
- Google Scholar
1. Tustison NJ
2. Cook PA
3. Klein A
4. Song G
5. Das SR
6. Duda JT
7. Kandel BM
8. van Strien N
9. Stone JR
10. Gee JC
11. Avants BB
(2014) Large-scale evaluation of ANTs and FreeSurfer cortical thickness measurements
NeuroImage 99:166–179.

https://doi.org/10.1016/j.neuroimage.2014.05.044
- PubMed
- Google Scholar
1. Wagner MJ
2. Kim TH
3. Savall J
4. Schnitzer MJ
5. Luo L
(2017) Cerebellar granule cells encode the expectation of reward
Nature 544:96–100.

https://doi.org/10.1038/nature21726
- PubMed
- Google Scholar
(2012) Whole-brain mapping of direct inputs to midbrain dopamine neurons
Neuron 74:858–873.

https://doi.org/10.1016/j.neuron.2012.03.017
- PubMed
- Google Scholar
1. Yágüez L
2. Coen S
3. Gregory LJ
4. Amaro E Jr
5. Altman C
6. Brammer MJ
7. Bullmore ET
8. Williams SCR
9. Aziz Q
(2005) Brain response to visceral aversive conditioning: a functional magnetic resonance imaging study
Gastroenterology 128:1819–1829.

https://doi.org/10.1053/j.gastro.2005.02.068
- PubMed
- Google Scholar
1. Yushkevich PA
2. Piven J
3. Hazlett HC
4. Smith RG
5. Ho S
6. Gee JC
7. Gerig G
(2006) User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability
NeuroImage 31:1116–1128.

https://doi.org/10.1016/j.neuroimage.2006.01.015
- PubMed
- Google Scholar
1. Zeidman P
2. Jafarian A
3. Corbin N
4. Seghier ML
5. Razi A
6. Price CJ
7. Friston KJ
(2019a) A guide to group effective connectivity analysis, part 1: First level analysis with DCM for fMRI
NeuroImage 200:174–190.

https://doi.org/10.1016/j.neuroimage.2019.06.031
- PubMed
- Google Scholar
1. Zeidman P
2. Jafarian A
3. Seghier ML
4. Litvak V
5. Cagnan H
6. Price CJ
7. Friston KJ
(2019b) A guide to group effective connectivity analysis, part 2: Second level analysis with PEB
NeuroImage 200:12–25.

https://doi.org/10.1016/j.neuroimage.2019.06.032
- PubMed
- Google Scholar
1. Zheng N
2. Gui Z
3. Liu X
4. Wu Y
5. Wang H
6. Cai A
7. Wu J
8. Li X
9. Kaewborisuth C
10. Zhang Z
11. Wang Q
12. Manyande A
13. Xu F
14. Wang J
(2023) Investigations of brain-wide functional and structural networks of dopaminergic and CamKIIα-positive neurons in VTA with DREADD-fMRI and neurotropic virus tracing technologies
Journal of Translational Medicine 21:543.

https://doi.org/10.1186/s12967-023-04362-6
- PubMed
- Google Scholar
1. Zhong S
2. Su T
3. Gong J
4. Huang L
5. Wang Y
(2023) Brain functional alterations in patients with anorexia nervosa: A meta-analysis of task-based functional MRI studies
Psychiatry Research 327:115358.

https://doi.org/10.1016/j.psychres.2023.115358
- PubMed
- Google Scholar

Article and author information

Author details

Enzo Nio
1. Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
Contribution
Data curation, Software, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

For correspondence
enzo.nio@uk-essen.de

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-8006-8993
Patrick Pais Pereira
1. Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
Contribution
Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Project administration

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0009-0004-8200-7931
Nicolas Diekmann

Institute for Neural Computation, Faculty of Computer Science, Ruhr University Bochum, Bochum, Germany

Contribution
Software, Formal analysis, Methodology, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-3638-7617
Mykola Petrenko
1. Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
Contribution
Data curation, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Alice Doubliez
1. Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
Contribution
Software, Validation, Investigation, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-1736-5458
Thomas Michael Ernst
1. Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
Contribution
Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Project administration, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-2170-9241
Giorgi Batsikadze
1. Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
Contribution
Supervision, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-0517-0119
Stefan Maderwald

Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany

Contribution
Resources, Methodology, Writing – review and editing

Competing interests
No competing interests declared
Cornelius Deuschl
1. Institute for Diagnostic and Interventional Radiology and Neuroradiology, Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Institute for Diagnostic and Interventional Radiology, Neuroradiology and Nuclear Medicine, University Hospital Knappschaftskrankenhaus Bochum, Ruhr University Bochum, Bochum, Germany
Contribution
Validation, Writing – review and editing, Neuroradiological evaluation of T1-weighted MRI scans for incidental findings

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-9262-7289
Metin Üngör

Department of Psychology, Philipps-University Marburg, Marburg, Germany

Contribution
Conceptualization, Funding acquisition, Methodology, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-5453-2860
Sen Cheng

Institute for Neural Computation, Faculty of Computer Science, Ruhr University Bochum, Bochum, Germany

Contribution
Conceptualization, Software, Formal analysis, Supervision, Funding acquisition, Methodology, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-6719-8029
Christian Josef Merz

Department of Cognitive Psychology, Institute of Cognitive Neuroscience, Ruhr University Bochum, Bochum, Germany

Contribution
Conceptualization, Funding acquisition, Methodology, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-5679-6595
Harald H Quick
1. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
2. High-Field and Hybrid MR Imaging, Essen University Hospital, University of Duisburg-Essen, Essen, Germany
Contribution
Resources, Funding acquisition, Writing – review and editing

Competing interests
No competing interests declared
Dagmar Timmann
1. Department of Neurology and Center for Translational Neuro- and Behavioral Sciences (C TNBS), Essen University Hospital, University of Duisburg-Essen, Essen, Germany
2. Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany
Contribution
Conceptualization, Resources, Supervision, Funding acquisition, Methodology, Writing – original draft, Project administration, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-1935-416X

Funding

Deutsche Forschungsgemeinschaft (SFB1280 - Project number 316803389)

Enzo Nio
Patrick Pais Pereira
Nicolas Diekmann
Mykola Petrenko
Alice Doubliez
Thomas Michael Ernst
Giorgi Batsikadze
Metin Üngör
Sen Cheng
Christian Josef Merz
Harald H Quick
Dagmar Timmann

European Union Horizon 2020

https://doi.org/10.3030/956414

Enzo Nio
Alice Doubliez
Dagmar Timmann

Deutsche Forschungsgemeinschaft (432647511)

Harald H Quick

Hans Böckler Stiftung (Individual scholarship)

Patrick Pais Pereira

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to thank our technician, Beate Brol, for the drawing and adjustment of DCN and VTA masks. Additionally, we would like to thank Greta Wippich for providing helpful illustrations. The MAGNETOM Terra 7T MRI system used in the study was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), grant number 432647511. This project has received funding from the DFG (project number 316803389 - SFB1280), the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No 956414 and an individual scholarship for Patrick Pais Pereira from the Hans-Böckler-Stiftung.

Ethics

Informed consent and consent to publish were obtained from all participants. The study was approved by the Ethics Committee of the University Hospital Essen (approval number: 20-9664-BO) and conducted in accordance with the Declaration of Helsinki.

Version history

Preprint posted: November 7, 2024
Sent for peer review: December 21, 2024
Reviewed Preprint version 1: January 31, 2025
Reviewed Preprint version 2: April 14, 2026
Version of Record published: July 13, 2026

Cite all versions

You can cite all versions using the DOI https://doi.org/10.7554/eLife.105399. This DOI represents all versions, and will always resolve to the latest one.

Copyright

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.