Common and distinct neural mechanisms of aversive and appetitive pain-related learning

1354 

Abstract Submission 

Jialin Li^1,2, Katharina Schmidt², Lea Busch², Tamás Spisák², Katja Wiech^3,2, Ulrike Bingel²

¹Max Planck School of Cognition, Leipzig, Germany, ²University Medicine Essen, Essen, Germany, ³Wellcome Centre for Integrative Neuroimaging, Oxford, United Kingdom

Jialin Li  
Max Planck School of Cognition|University Medicine Essen
Leipzig, Germany|Essen, Germany

Katharina Schmidt  
University Medicine Essen
Essen, Germany

Lea Busch  
University Medicine Essen
Essen, Germany

Tamás Spisák  
University Medicine Essen
Essen, Germany

Katja Wiech  
Wellcome Centre for Integrative Neuroimaging|University Medicine Essen
Oxford, United Kingdom|Essen, Germany

Ulrike Bingel  
University Medicine Essen
Essen, Germany

Learning from signals of upcoming aversive and appetitive events is crucial for humans. Several studies have investigated the mechanisms underlying these two types of learning (Martin-Soelch et al., 2007; Fullana et al., 2016; Klein et al., 2022; van der Schaaf et al., 2022), but mostly using different sensory modalities and paradigms (e.g., pain vs. monetary reward), which limits interpretations (Klein et al., 2022). A recent study (van der Schaaf et al., 2022) addressed this issue by comparing aversive and appetitive learning within the same sensory modality (tonic heat pain increase vs. decrease) and found marked behavioural differences including stronger aversive than appetitive acquisition but no differences in extinction. However, little is known about the neural underpinnings of such findings. This study aims to elucidate common and distinct neural mechanisms of aversive and appetitive pain-related learning in healthy individuals.

N = 62 healthy individuals participated in an fMRI paradigm in which moderate continuous pain was elicited by a novel capsaicin-induced tonic heat pain model (van der Schaaf et al., 2022). Individually calibrated phasic heat pain (unconditioned stimulus, US) was experimentally induced and paired with predictive cues (conditioned stimulus, CS). Following habituation, three geometrical cues (CSincrease, CSdecrease, CSmedium) signaled pain exacerbation (USincrease), pain decrease (USdecrease) and no change in temperature (USmedium) respectively in the acquisition phase. In the subsequent extinction phase, all cues were followed by the USmedium (Fig. 1). Neuroimaging data were preprocessed using the fMRIprep pipeline. Univariate general linear model analyses were performed separately for each phase using SPM12. First-level regressors included all CS and US events during acquisition and CS events during extinction, as well as nuisance variables (ratings, head movement). To identify brain regions that show increasing engagement with learning over time, we applied time modulation to CS regressors (Forkmann et al., 2023), and individual-wise contrasts modeling time x condition interactions (e.g., aversive: CSincrease x time > CSmedium x time) were entered into group-level one-sample t-tests. Results were followed up by functional connectivity analyses (gPPI; McLaren et al., 2012) for key regions identified. Results were thresholded at p < .05 family-wise error (FWE) corrected at peak level with small volume correction (SVC) for brain regions included in our a priori hypotheses (SFB1280 A11 project).

The time x condition interaction indicated that compared to appetitive events, the acquisition of aversive events was associated with a stronger increase in medial thalamus activity (Fig. 2a). Conjunction analysis revealed that both types of learning were accompanied by increasing responses in the superior frontal gyrus and medial occipital cortex (Fig. 2b).
Functional connectivity analysis using the occipital cortex as the seed region showed positive coupling with the frontal operculum during both types of learning (Fig. 2c), suggesting a similar connectivity pattern regardless of valence during acquisition.
In the extinction phase, we found a greater decrease in activity in the lateral occipital cortex for the aversive than the appetitive condition, whereas the parahippocampal gyrus showed a stronger decrease in the appetitive condition (Fig. 2d). Both conditions showed a decrease in response over the course of the extinction phase in the vmPFC (Fig. 2e), which was negatively correlated with the activity in the periaqueductal gray (Fig. 2f).

Our results show that in healthy individuals, aversive and appetitive learning (within the same sensory modality) involve a common set of brain regions but can also be distinguished in areas such as the medial thalamus during acquisition and the occipital cortex, parahippocampal gyrus and vmPFC during extinction.

Learning and Memory Other

Activation (eg. BOLD task-fMRI) ¹

Connectivity (eg. functional, effective, structural)

Univariate Modeling

Perception: Pain and Visceral ²

Acquisition

Cognition

FUNCTIONAL MRI

Learning

Pain

Somatosensory

Univariate

Other - extinction; pain relief