Machine learning–based prediction of clinical pain using multimodal neuroimaging and autonomic metrics

Abstract

While self-report pain intensity ratings are the gold standard in clinical pain assessment, they are highly variable, inherently subjective in nature, and significantly influenced by multidimensional factors. The lack of objective biomarkers for pain has contributed to suboptimal chronic pain management (e.g., opioid public health crisis) [26]. Thus, research focused on the development of quantitative, objective biomarkers/predictors alongside selfreport to aid diagnosis, estimate prognosis, and predict treatment efficacy is of increasing importance to combat chronic pain [24,30,32].Growing consensus has suggested that altered central nervous system processing can support and maintain abnormal pain perception in chronic pain, implicating aberrant activity and connectivity of multiple functional brain networks, including default mode, salience, and 2

Composite Pain Biomarker Signatures for Objective Assessment and Effective Treatment

“…This is what is needed and work has begun. A study used ML-based predictions of clinical pain with two types of neuroimaging (resting state and arterial spin labeling) plus autonomic metrics was recently published (Lee et al, 2018b). When all three multimodal parameters were combined, patient classification between lower and higher clinical pain intensity states had the best performance, illustrating the power of a ML-based composite approach.…”

Sensor Technologies to Manage the Physiological Traits of Chronic Pain: A Review

“…Functional magnetic resonance imaging (fMRI) is another computer vision technique that has been used in the analysis of the central nervous system response in relation to the pain perception and modulation processes, both chronic and acute ones [249,250]. Classification task in these systems is normally based on multivariate pattern analyses (MVPA) [47,251], and on the brain response in two recognizable situations [252]: baseline activity and pain state. Baseline activity is used as reference in the classification method.…”

“…Magnetic resonance imaging (MRI) sequences are used to construct voxel-based models in activity time series [253]. Images are previously preprocessed to eliminate motion artifacts, breathing and cardiac activity [252]. Through a standardized brain atlas, the brain is segmented into regions of interest (ROIs).…”

Interpretable machine learning models for classifying low back pain status using functional physiological variables

“…Increasingly, researchers are turning towards advance statistical learning techniques to develop accurate prediction models for people with LBP using information from high-dimensional, multivariate biological signals [3]. Existing studies have used biological signals such as surface electromyography (sEMG) [4,5], kinematics [6,7], brain neuroimaging [3], and spine neuroimages [8] as candidate predictors; feeding into statistical learning techniques such as support vector machine (SVM) [3][4][5][6], neural networks [7], and natural language processing [8]. The excellent predictive accuracy of current models developed using a spectrum of biological signals and statistical techniques has demonstrated the potential for such methods to assist clinical decision-making [3][4][5][6][7][8].…”

“…Despite their predictive value, prohibitive barriers exist towards a more generalized integration of advance predictive models into routine clinical practice. First, some biological signals, such as brain neuroimages [3], are not feasible to be collected as routine in most clinical settings. Second, candidate biological signals should ideally be collected during activities that are routinely assessed clinically, rather than more complex sporting manoeuvres such as golf [4].…”