Background & Aims
Although translational neuroimaging has identified several brain markers of pain, these markers are far from being suitable for clinical use due to lack of rigorous validation and standardization1. There is a need for developing robust brain imaging models of pain using large sample sizes and multivariate models2. A promising approach recently developed is the Connectome-Based Predictive Model (CPM) which predicts brain-behavior relationships using brain connectivity data3. Here, we employed a CPM analysis of experimental pain during functional magnetic resonance imaging (fMRI) to predict clinical pain in fibromyalgia (FM) patients.
Methods
104 female FM patients (42.3 ±12.6 years old) underwent a fMRI paradigm (6-10 minutes duration) in which pseudorandomized levels of pressure-pain were applied to the left thumbnail in 5-25s blocks interleaved with rest periods. Clinical pain intensity was assessed immediately prior to fMRI using a visual analogue scale (VAS).
Preprocessing in fMRIPrep5, SPM, and AFNI included physiological noise removal, motion correction, co-registration to the structural T1, normalization to MNI standard space, smoothing, brain extraction, and high pass filtering. Connectomes were generated using the Brainnetome atlas4.
CPM followed the protocol of Shen et al.3 with a leave-one-subject-out framework to predict VASscores. Thresholding significance was set at 0.01. Goodness-of-fit was measured as the correlation between predicted and actual pain intensity scores. Permutation assessed significance with 1000 iterations with randomized labels to generate a null distribution.
Results
The correlation between actual and predicted values for VAS scores was found to be significant for positive edges of the connectome (R=0.21, p=0.03), which indicates that increased connectivity during painful stimulation was associated with increased clinical pain. No significant results were found using negative edges. The highest degree nodes (> 25) that contributed to the predictive model included perigenual anterior cingulate cortex, anterior, mid, and posterior insula, and putamen.
Conclusions
Results showed that multivariate connectome predictive modeling using experimental pain in FM patients predicts the severity of clinical pain. Significant features in anterior cingulate and insular cortices, which have been implicated in augmented pain processing in FM, were major contributors to the predictive model. Data collection is ongoing to assemble a novel dataset for validation of this signature in FM and other chronic pain conditions. Ultimately, with rigorous development and validation, these neuroimaging markers may garner clinical utility.
References
1.Davis KD, Flor H, Greely HT, et al. Brain imaging tests for chronic pain: medical, legal and ethical issues and recommendations. Nat Rev Neurol. 2017;13(10):624-638.
2.Woo CW, Chang LJ, Lindquist MA, Wager TD. Building better biomarkers: brain models in translational neuroimaging. Nat Neurosci. 2017;20(3):365-377.
3.Shen X, Finn ES, Scheinost D, et al. Using connectome-based predictive modeling to predict individual behavior from brain connectivity. Nat Protoc. 2017;12(3):506-518.
4.Glover GH, Li TQ, Ress D. Image-based method for retrospective correction of physiological motion effects in fMRI: RETROICOR. Magn Reson Med. 2000;44(1):162-167.
5.Esteban O, Markiewicz CJ, Blair RW, et al. fMRIPrep: a robust preprocessing pipeline for functional MRI. Nat Methods. 2019;16(1):111-116.
Presenting Author
Noah Waller
Poster Authors
Eric Ichesco BS
University of Michigan Chronic Pain and Fatigue Research Center
Lead Author
Noah Waller BS
University of Michigan
Lead Author
Scott Peltier PhD
University of Michigan
Lead Author
Ishtiaq Mawla
PhD
University of Michigan
Lead Author
Chelsea Kaplan
PhD
University of Michigan
Lead Author
Tony Larkin
University of Michigan
Lead Author
Andrew Schrepf
University of Michigan
Lead Author
Anson Kairys PhD
University of Michigan
Lead Author
Steven Harte
PhD
University of Michigan Chronic Pain and Fatigue Research Center
Lead Author
Daniel Clauw
University of Michigan Chronic Pain and Fatigue Research Center
Lead Author
Richard Harris PhD
University of California Irvine
Lead Author
Topics
- Pain Imaging