Print  Close

Improved performance of within- and between-individual pain prediction using multi-echo fMRI

Poster No:

2500 

Submission Type:

Abstract Submission 

Authors:

Eui-Jin Jung1,2, Suhwan Gim1,2,3, Dong Hee Lee1, Jae-Joong Lee1, Choong-Wan Woo1,2,3,4

Institutions:

1Center for Neuroscience Imaging Research, Institute for Basic Science, Suwon, Korea, Republic of, 2Department of Biomedical Engineering, Sungkyunkwan University, Suwon, Korea, Republic of, 3Department of Intelligent Precision Healthcare Convergence, Sungkyunkwan University, Suwon, Korea, Republic of, 4Life-inspired Neural Network for Prediction and Optimization Research Group, Suwon, Korea, Republic of

First Author:

Eui-Jin Jung  
Center for Neuroscience Imaging Research, Institute for Basic Science|Department of Biomedical Engineering, Sungkyunkwan University
Suwon, Korea, Republic of|Suwon, Korea, Republic of

Co-Author(s):

Suhwan Gim  
Center for Neuroscience Imaging Research, Institute for Basic Science|Department of Biomedical Engineering, Sungkyunkwan University|Department of Intelligent Precision Healthcare Convergence, Sungkyunkwan University
Suwon, Korea, Republic of|Suwon, Korea, Republic of|Suwon, Korea, Republic of
Dong Hee Lee  
Center for Neuroscience Imaging Research, Institute for Basic Science
Suwon, Korea, Republic of
Jae-Joong Lee  
Center for Neuroscience Imaging Research, Institute for Basic Science
Suwon, Korea, Republic of
Choong-Wan Woo  
Center for Neuroscience Imaging Research, Institute for Basic Science|Department of Biomedical Engineering, Sungkyunkwan University|Department of Intelligent Precision Healthcare Convergence, Sungkyunkwan University|Life-inspired Neural Network for Prediction and Optimization Research Group
Suwon, Korea, Republic of|Suwon, Korea, Republic of|Suwon, Korea, Republic of|Suwon, Korea, Republic of

Introduction:

Pain has multidimensional features, leading to individual differences in the experience of pain. To understand of these multidimensional features of pain, it is crucial to quantify pain as a subjective experience. Previous studies have developed neuroimaging biomarkers of pain using subjective pain reports (Wager et al., 2013; Woo et al., 2017; Lee et al., 2021). Recently, multi-echo fMRI (ME-fMRI) has emerged as an effective tool to alleviate the effect of non-neural signal such as head motion, physiological, and thermal noise. It notably improves the temporal signal-to-noise ratio (tSNR) by optimizing T2* decay rates at each voxel, thereby improving the reliability of functional connectivity in individuals (Lynch et al., 2020) and increasing effect size estimates and statistical power in task-fMRI (Lombardo et al., 2016). In addition, Multi-Echo Independent Component Analysis (ME-ICA), a denoising technique (Kundu et al., 2012), uses multiple echoes to distinguish between TE-dependent BOLD-like and TE-independent non-BOLD-like fluctuations based on T2* decay. Therefore, ME-fMRI combined with ME-ICA may better capture the brain representation of individual differences in pain than the previous fMRI sequence (i.e., single-echo fMRI). However, no previous studies have investigated the effects of the combination of ME-fMRI and ME-ICA on the brain representation of pain. Here we compared the performance of the pain prediction model between ME-fMRI and the single-echo fMRI.

Methods:

Thirty-seven participants (19 female) underwent a thermal pain induction task during fMRI. Participants received thermal stimulation with fixed temperatures ranging from 45°C to 47.5°C, increasing in 0.5°C increments, and rated the level of intensity. Functional images were collected using a whole-brain T2* weighted echo-planar sequence (TR: 1sec; TE1: 13.00ms, TE2: 31.26ms, TE3: 49.52ms; FOV: 210mm; voxels: 3mm; 52 slices; in-plane acceleration factor: 2; multi-band factor: 4). In this study, there were three experimental conditions: single echo based on second echo (SE), multi-echo optimal combination (OC) and denoised by ME-ICA (ME). Initially, we compared tSNR to examine the effect of ME-fMRI. To understand the effect of ME-fMRI and ME-ICA on the brain representation of pain, we then performed two pain prediction models: 1) within-individual pain prediction model and 2) between-individual pain prediction.

Results:

In this study, the ME condition showed significantly higher tSNR compared to SE and OC (Fig. 1a). We also examined the correlation between frame-wise displacement (FD) and tSNR difference between the OC minus SE and the ME minus OC conditions (Fig. 1b). In the OC minus SE condition, a negative correlation was observed between FD and tSNR difference (r=-0.66). Conversely, a positive correlation was observed in the ME minus OC condition (r=0.58). This implies that tSNR was increased with ME-fMRI and especially with ME-ICA, which can effectively reduce non-neuronal signals such as head movements. Within-individual pain prediction utilizing leave-one-subject-out cross-validation displayed significant predictive performance across all conditions (Fig. 2b). Notably, ME showed the highest predictive performance (r=0.53) compared to SE (r=0.49) and OC (r=0.51). Furthermore, in the between-individual pain prediction, ME condition (r=0.47) outperformed both SE (r=0.42) and OC (r=0.41;Fig. 2e). These results imply that the ME-fMRI combined with ME-ICA enhances tSNR throughout overall brain regions and reduces the influence of non-neuronal signals, thus capturing brain representation of pain than single-echo fMRI.
Supporting Image: Figure1Temporalsignal-to-noiseratio.png
   ·Figure 1. Temporal signal-to-noise-ratio
Supporting Image: Figure2Within-andbetween-personpainpredictionresults.png
   ·Figure 2. Within- and between-person pain prediction results
 

Conclusions:

This research indicates a promising way to capture pain intra- and inter-individual variability in pain using ME-fMRI rather than single-echo fMRI. The utilization of ME-fMRI holds the potential to enhance personalized pain assessment and management strategies, consequently refining our understanding of the diverse range of pain experiences among individuals.

Modeling and Analysis Methods:

Classification and Predictive Modeling 2
Methods Development
Multivariate Approaches

Novel Imaging Acquisition Methods:

BOLD fMRI

Perception, Attention and Motor Behavior:

Perception: Pain and Visceral 1

Keywords:

FUNCTIONAL MRI
Multivariate
Pain

1|2Indicates the priority used for review

Provide references using author date format

1. Wager, T. D., Atlas, L. Y., Lindquist, M. A., Roy, M., Woo, C. W., & Kross, E. (2013). An fMRI-based neurologic signature of physical pain. New England Journal of Medicine, 368(15), 1388-1397.
2. Woo, C. W., Schmidt, L., Krishnan, A., Jepma, M., Roy, M., Lindquist, M. A., ... & Wager, T. D. (2017). Quantifying cerebral contributions to pain beyond nociception. Nature communications, 8(1), 14211.
3. Lee, J. J., Kim, H. J., Čeko, M., Park, B. Y., Lee, S. A., Park, H., ... & Woo, C. W. (2021). A neuroimaging biomarker for sustained experimental and clinical pain. Nature medicine, 27(1), 174-182.
4. Lynch, C. J., Power, J. D., Scult, M. A., Dubin, M., Gunning, F. M., & Liston, C. (2020). Rapid precision functional mapping of individuals using multi-echo fMRI. Cell reports, 33(12).
5. Lombardo, M. V., Auyeung, B., Holt, R. J., Waldman, J., Ruigrok, A. N., Mooney, N., ... & Kundu, P. (2016). Improving effect size estimation and statistical power with multi-echo fMRI and its impact on understanding the neural systems supporting mentalizing. Neuroimage, 142, 55-66.
6. Kundu, P., Inati, S. J., Evans, J. W., Luh, W. M., & Bandettini, P. A. (2012). Differentiating BOLD and non-BOLD signals in fMRI time series using multi-echo EPI. Neuroimage, 60(3), 1759-1770.