Automatic pain quantification using autonomic parameters.

Abstract

The objective measurement of subjective, multi-dimensionally experienced pain is a problem for which there has not been an adequate solution. Although verbal methods (e.g., pain scales and questionnaires) are commonly used to measure clinical pain, they tend to lack objectivity, reliability, or validity when applied to mentally impaired individuals. Biopotential and behavioral parameters may represent a solution. Such coding systems already exist, but they are either very costly or time-consuming or have not been sufficiently evaluated. In this context, we collected a database of biopotentials to advance an automated pain recognition system, determine its theoretical testing quality, and optimize its performance. For this purpose, participants were subjected to painful heat stimuli under controlled conditions. One hundred thirty-five features were extracted from the mathematical groupings of amplitude, frequency, stationarity, entropy, linearity, and variability. The following features were chosen as the most selective: (1) electromyography corrugator peak to peak, (2) corrugator shannon entropy, and (3) heart rate variability slope RR. Individual-specific calibration allows the adjustment of feature patterns, resulting in significantly more accurate pain detection rates. The objective measurement of pain in patients will provide valuable information for the clinical team, which may aid the objective assessment of treatment (e.g., effectiveness of drugs for pain reduction, information on surgical indication, and quality of care provided to patients).

Methods for Person-Centered Continuous Pain Intensity Assessment From Bio-Physiological Channels

“…The system extracts features from the SCL, ECG and EMG channels. Based on the work of Walter et al [25], an initial set of 5 features were selected with regards to their individual ability to differentiate between pain intensities: stationary standard deviation (SCL), peak to peak mean (EMG), slope of R peaks interval (ECG), peak (EMG) and zero crossings (EMG). Following, an early fusion architecture is used to continuously predict the 5 different pain levels.…”

“…These measures represent the rate of vibration contained in the signal. Furthermore, the bandwidth as well as the amount of zero crossings was also computed [25].…”

“…The features were the mean and root mean square deviation of the heart rate variability during the 5.5 seconds. Finally, the slope of a linear regression that was fit to the intervals and their occurrence in time was computed as well [25].…”

Exploring Deep Physiological Models for Nociceptive Pain Recognition

“…Finally, 154 a third order bandpass Butterworth filter with a frequency range of [0.1, 250] Hz was 155 applied on the ECG signal. Furthermore, the data is segmented as proposed in [33], but 156 rather than using 5.5 sec windows with a shift of 3 sec from the elicitations' onset, the 157 preprocessed signals were segmented into windows of length 4.5 sec, with a shift from 158 the elicitations' onset of 4 sec (see Fig 2(a)) based on the data driven signal 159 segmentation approach that was recently proposed in [27]. Each signal extracted within 160 this window constitutes a 1-D array of size 4.5 × 256 = 1152 and is later on used in 161 combination with the corresponding level of pain elicitation to optimize and assess the 162 designed deep classification architectures.…”

“…Martinéz et al [12] were able to 28 significantly outperform standard approaches built upon hand-crafted features by using 29 a deep learning algorithm for affect modelling based on physiological signals (two 30 physiological signals consisting of Skin Conductance (SC) and Blood Volume Pulse 31 (BVP) were used in this specific work). The designed approach consisted of a 32 multi-layer Convolutional Neural Network (CNN) [13] combined with a single-layer 33 perceptron (SLP). The parameters of the CNN were trained in an unsupervised manner 34 using denoising auto-encoders [14].…”