Reduced putamen activity precedes placebo analgesia: individual participant data meta-analysis

2522 

Abstract Submission 

Balint Kincses¹, Tamás Spisák², Tor Wager³, Ulrike Bingel⁴, Placebo Imaging Consortium⁵

¹University Hospital Essen, Essen, NRW, ²University Medicine Essen, Essen, Germany, ³Dartmouth College, Hanover, NH, ⁴University Medicine Essen, Essen, NRW, ⁵https://github.com/placebo-imaging-consortium/placebo-imaging-consortium.github.io, -, -

Balint Kincses  
University Hospital Essen
Essen, NRW

Tamás Spisák  
University Medicine Essen
Essen, Germany

Tor Wager, PhD  
Dartmouth College
Hanover, NH

Ulrike Bingel  
University Medicine Essen
Essen, NRW

Placebo Imaging Consortium  
https://github.com/placebo-imaging-consortium/placebo-imaging-consortium.github.io
-, -

Placebo analgesia (PA) is a widely investigated phenomenon, with a significant and considerable reduction of pain (Forsberg et al., 2017). A better understanding of its underlying neural mechanisms has promising implications for the exploitation of this effect in various clinical contexts. The neural correlates of PA during the pain delivery phase have previously been extensively investigated (Atlas & Wager, 2014; Zunhammer et al., 2018, 2021). However, brain activity preceding the pain stimulation, i.e. the anticipatory brain activity, has not thoroughly investigated. As anticipatory brain activity may reflect placebo-related expectations, its characterisation is an essential step towards the better understanding of mechanisms underlying PA. Here we investigated PA-related anticipatory brain activation changes with an unprecedented sample size in an individual participant data (IPD) meta-analysis.

IPD neuroimaging and behavioral data was collected on behalf of the Placebo Imaging Consortium, as previously described in (Zunhammer et al., 2018, 2021). The current analysis is based on an extended database, including 570 participants from 18 studies and focused on the differences in the anticipatory activity before pain stimulation with and without placebo-induction. Study plans were pre-registered at the Open Science Framework (https://osf.io/z4qu2). The preprocessing of IPD first-level anticipatory activation maps (ß or contrast images) consisted of re-slicing individual maps to MNI152 template, winsorizing voxel level data (outlier mitigation), and calculating meta-analytic summary scores with the voxel level generic inverse variance (GIV) approach, with studies as random effect. We investigated mean activation differences during the anticipation phase (placebo-control), as well as the correlation of these changes with behavioral placebo analgesia (pain rating differences). Permutation based (n permutation=5000) max statistics with probabilistic threshold free cluster enhancement (Spisák et al., 2019) was used to correct for the family-wise error rate with an alpha-value of 0.05 (two tailed). The analysis was done in MATLAB2021b and SPM12.

With the extended dataset, we observed behavioural PA effects similar to the previously reported estimates (Zunhammer et al., 2018) (Hedge's g: -0.63). In terms of mean placebo-related differences in anticipatory brain activity, we observed a statistically significant decrease in three clusters in the occipital lobe: MNI peak coordinates: x=-5.7, y=-87.9, z=-9.8; x=-4.2, y=-74.6, z=26.3; x=-18.2, y=-64.3, z=13.8, FWER corrected, p<0.05) and in the cerebellum (x= -21.9, y=-79.8, z=-23.8, FWER corrected, p<0.05). No significant FWER corrected activity increase was observed. Placebo-related decreases in anticipatory brain activity were negatively correlated with behavioral PA in the left putamen (Figure 1, MNI peak coordinates: x=-24.1, y=0.6, z=2.0, FWER corrected, p<0.05, two-tailed). Specifically, decreased anticipatory activity in the placebo condition led to higher behavioral PA (Figure 2).

Our previous meta-analytic results on the neural correlates of PA (Zunhammer et al., 2021) suggested that placebo is a multifaceted phenomenon, possibly driven by multiple complementary/parallel neural processes (e.g. in the presence and absence of conditioning). This heterogeneity may also apply to anticipatory activity and explain the lack of coherent increases across the investigated studies. The correlation analysis suggests that the involvement of the putamen is a common characteristic of different PA-related neural mechanisms during the anticipatory phase. It is therefore a promising target for further research aiming at the exploitation of PA in clinical contexts.

Higher Cognitive Functions Other

Activation (eg. BOLD task-fMRI) ²

Univariate Modeling

Perception: Pain and Visceral ¹

FUNCTIONAL MRI

Meta- Analysis

MRI

Pain

Other - placebo; placebo analgesia; pain anticipation; individual participant data