Elsevier

NeuroImage

Volume 81, 1 November 2013, Pages 283-293
NeuroImage

A novel approach to predict subjective pain perception from single-trial laser-evoked potentials

https://doi.org/10.1016/j.neuroimage.2013.05.017Get rights and content

Highlights

  • We proposed a novel and practice-oriented way to predict pain from single-trial LEPs.

  • Single-trial LEP features were rapidly and reliably estimated using CSP and MLR.

  • Naïve Bayes classifier discretely predicted low and high pain accurately.

  • Linear prediction model continuously predicted the intensity of pain reliably.

Abstract

Pain is a subjective first-person experience, and self-report is the gold standard for pain assessment in clinical practice. However, self-report of pain is not available in some vulnerable populations (e.g., patients with disorders of consciousness), which leads to an inadequate or suboptimal treatment of pain. Therefore, the availability of a physiology-based and objective assessment of pain that complements the self-report would be of great importance in various applications. Here, we aimed to develop a novel and practice-oriented approach to predict pain perception from single-trial laser-evoked potentials (LEPs). We applied a novel single-trial analysis approach that combined common spatial pattern and multiple linear regression to automatically and reliably estimate single-trial LEP features. Further, we adopted a Naïve Bayes classifier to discretely predict low and high pain and a multiple linear prediction model to continuously predict the intensity of pain perception from single-trial LEP features, at both within- and cross-individual levels. Our results showed that the proposed approach provided a binary prediction of pain (classification of low pain and high pain) with an accuracy of 86.3 ± 8.4% (within-individual) and 80.3 ± 8.5% (cross-individual), and a continuous prediction of pain (regression on a continuous scale from 0 to 10) with a mean absolute error of 1.031 ± 0.136 (within-individual) and 1.821 ± 0.202 (cross-individual). Thus, the proposed approach may help establish a fast and reliable tool for automated prediction of pain, which could be potentially adopted in various basic and clinical applications.

Introduction

Pain is an unpleasant multidimensional experience associated with real or potential tissue damage (Loeser and Treede, 2008). Therefore, pain experience does not simply reflect sensory information but can be substantially influenced by various psycho-social contexts (e.g., the gender of experimenter) (Aslaksen et al., 2007) and psycho-physiological factors (e.g., fluctuations in vigilance and attention). Since pain is a subjective first-person experience, self-report (e.g., Visual Analog Scales [VAS] and Numeric Rating Scales) is the gold standard for the determination of the presence, absence, and intensity of pain perception in clinical practice (Cruccu et al., 2010, Haanpaa et al., 2010). While self-report of pain provides important clinical information for the adequate treatment of pain patients in most situations (Brown et al., 2011), it fails to be used in some vulnerable populations (e.g., patients with disorders of consciousness, including coma, vegetative state, and minimally conscious state) (Schnakers and Zasler, 2007). Lack or any inaccuracy of pain assessment can lead to inadequate or suboptimal treatment of pain in these vulnerable patients, which may lead to various additional clinical problems (e.g., psychological distress or depression, the development of chronic pain) (Roulin and Ramelet, 2012, Zwakhalen et al., 2006). Although reliable pain assessment is important to adequately treat patients suffering from persistent pain (Gagliese and Melzack, 1997, Zwakhalen et al., 2006), it is extremely difficult to detect and monitor pain in specific clinical populations, such as non-communicative patients with disorders of consciousness (Schnakers et al., 2010). In addition, the brain damage of these patients may lead to confused, stereotyped, and uncoordinated behaviors (Roulin and Ramelet, 2012, Schnakers and Zasler, 2007), which is the barrier to adopt behavioral responses as pain indicators. Therefore, the availability of a physiology-based and objective assessment of pain that complements the self-report of pain would be of great importance in basic and clinical applications.

Nowadays, electroencephalographic (EEG) responses elicited by nociceptive laser heat pulses that selectively excite nociceptive free nerve endings in the epidermis (Bromm and Treede, 1984) are widely adopted to investigate the peripheral and central processing of nociceptive sensory input (Iannetti et al., 2003, Treede et al., 2003). Such laser-evoked potentials (LEPs) are mediated by the activation of type-II Aδ mechano-heat nociceptors (Treede et al., 1995) and spinothalamic neurons in the anterolateral quadrant of the spinal cord (Treede et al., 2003). LEPs consist of several transient responses that are time locked and phase locked to the onset of laser stimuli. The largest LEP response consists of a biphasic negative–positive complex (N2 and P2 waves), maximal at the scalp vertex (Bromm and Treede, 1984) and largely reflecting the activity of the bilateral operculoinsular and anterior cingulate cortex (Garcia-Larrea et al., 2003). The strong relationship between the N2 and P2 amplitudes in LEPs and the intensity of pain perception have been well characterized (Bromm and Treede, 1991, Garcia-Larrea et al., 1997, Iannetti et al., 2005, Kakigi et al., 1989), and the correlations between N2 and P2 latencies and the intensity of pain perception were also reported (Iannetti et al., 2005). All these findings inspire us to explore the possibility of objective assessment of pain based on the single-trial LEP features related to latencies and amplitudes of N2 and P2 waves.

The aim of the present study was to develop a novel approach to rapidly and reliably predict pain from single-trial LEP features (Fig. 1), which can be achieved through the following two major steps. First, a novel method that combines common spatial pattern (CSP) and multiple linear regression (MLR) was proposed to achieve an automated and reliable single-trial estimate of LEP features. Second, a Naïve Bayes classifier and a multiple linear prediction model were trained to respectively distinguish low and high pain and predict the intensity level of pain perception from single-trial LEP features. Such training and predicting were achieved at both within-individual level, where the classifier and prediction model were trained on and applied to single-trial LEPs from the same individual, and cross-individual level, where the classifier and prediction model were trained on a cohort of individuals and applied to another individual.

Section snippets

Experimental design and EEG recording

Twenty-nine healthy participants (9 females and 20 males) aged 17–25 years (mean 22.2 ± 1.9), without a history of chronic pain, participated in the study. All participants gave written informed consent, and the local ethics committee approved the experimental procedures.

Radiant-heat stimuli were generated by an infrared neodymium yttrium aluminum perovskite (Nd:YAP) laser with a wavelength of 1.34 μm (Electronical Engineering, Italy). Laser pulses activate directly nociceptive terminals in the

Laser-evoked brain responses

Laser stimuli elicited a clear pinprick perception and reproducible time-locked LEPs in all participants (mean VAS = 5.8 ± 1.2), related to the activation of Aδ fibers (Bromm and Treede, 1984). Fig. 2 shows the group-level average LEP waveforms at Cz (nose reference) and the scalp topographies of N2 and P2 waves for each level of pain perception intensity (I1–I4). Across subjects, latencies and amplitudes of N2 and P2 peaks for each level of pain perception intensity were as follows: N2 latency

Discussion

In the present study, we aimed at developing a fast, automated, and reliable approach for (1) estimating single-trial LEP features and (2) predicting pain perception from estimated single-trial LEP features. We applied a novel single-trial analysis approach that combined CSP and MLR to automatically and reliably estimate single-trial LEP features, and adopted a Naïve Bayes classifier to discretely predict low and high pain and a multiple linear prediction model to continuously predict the

Acknowledgments

LH and PX were supported by the National Natural Science Foundation of China (31200856), Natural Science Foundation Project of CQ CSTC, Special Financial Grant from the China Postdoctoral Science Foundation (2012T50755), and New Teacher Fund of Ministry of Education of China (20120182120002). GH, YSH and ZGZ were partially supported by a Grant (HKU 762111M) from the Hong Kong SAR Research Grants Council and a Grant (201211159074) from HKU CRCG. The collaboration between LH and GDI is generously

References (73)

  • L. Hu et al.

    A novel approach for enhancing the signal-to-noise ratio and detecting automatically event-related potentials (ERPs) in single trials

    NeuroImage

    (2010)
  • G.D. Iannetti et al.

    Operculoinsular cortex encodes pain intensity at the earliest stages of cortical processing as indicated by amplitude of laser-evoked potentials in humans

    Neuroscience

    (2005)
  • R. Kakigi et al.

    Pain-related somatosensory evoked potentials following CO2 laser stimulation in man

    Electroencephalogr. Clin. Neurophysiol.

    (1989)
  • Z.J. Koles

    The quantitative extraction and topographic mapping of the abnormal components in the clinical EEG

    Electroencephalogr. Clin. Neurophysiol.

    (1991)
  • V. Kunde et al.

    Topography of middle-latency somatosensory evoked potentials following painful laser stimuli and non-painful electrical stimuli

    Electroencephalogr. Clin. Neurophysiol.

    (1993)
  • V. Legrain et al.

    Attentional modulation of the nociceptive processing into the human brain: selective spatial attention, probability of stimulus occurrence, and target detection effects on laser evoked potentials

    Pain

    (2002)
  • V. Legrain et al.

    The role of working memory in the attentional control of pain

    Pain

    (2011)
  • J.D. Loeser et al.

    The Kyoto protocol of IASP basic pain terminology

    Pain

    (2008)
  • A. Marquand et al.

    Quantitative prediction of subjective pain intensity from whole-brain fMRI data using Gaussian processes

    NeuroImage

    (2010)
  • S.D. Mayhew et al.

    Automated single-trial measurement of amplitude and latency of laser-evoked potentials (LEPs) using multiple linear regression

    Clin. Neurophysiol.

    (2006)
  • A. Mouraux et al.

    A multisensory investigation of the functional significance of the “pain matrix”

    NeuroImage

    (2011)
  • J. Muller-Gerking et al.

    Designing optimal spatial filters for single-trial EEG classification in a movement task

    Clin. Neurophysiol.

    (1999)
  • K.N. Ochsner et al.

    Neural correlates of individual differences in pain-related fear and anxiety

    Pain

    (2006)
  • M. Prato et al.

    A regularization algorithm for decoding perceptual temporal profiles from fMRI data

    NeuroImage

    (2011)
  • M.J. Roulin et al.

    Pain indicators in brain-injured critical care adults: an integrative review

    Aust. Crit. Care

    (2012)
  • D.A. Seminowicz et al.

    Cortical responses to pain in healthy individuals depends on pain catastrophizing

    Pain

    (2006)
  • Y. Tousignant-Laflamme et al.

    Establishing a link between heart rate and pain in healthy subjects: a gender effect

    J. Pain

    (2005)
  • I. Tracey et al.

    The cerebral signature for pain perception and its modulation

    Neuron

    (2007)
  • R.D. Treede et al.

    Clinical usefulness of laser-evoked potentials

    Neurophysiol. Clin.

    (2003)
  • E. Valentini et al.

    The primary somatosensory cortex largely contributes to the early part of the cortical response elicited by nociceptive stimuli

    NeuroImage

    (2012)
  • H. Zhang

    The optimality of naive Bayes

    (2004)
  • U. Baumgärtner et al.

    Multiple somatotopic representations of heat and mechanical pain in the operculo-insular cortex: a high-resolution fMRI study

    J. Neurophysiol.

    (2010)
  • B. Blankertz et al.

    The Berlin Brain–Computer Interface: accurate performance from first-session in BCI-naive subjects

    IEEE Trans. Biomed. Eng.

    (2008)
  • B. Bromm et al.

    Nerve fibre discharges, cerebral potentials and sensations induced by CO2 laser stimulation

    Hum. Neurobiol.

    (1984)
  • B. Bromm et al.

    Laser-evoked cerebral potentials in the assessment of cutaneous pain sensitivity in normal subjects and patients

    Rev. Neurol. (Paris)

    (1991)
  • J. Brooks et al.

    REVIEW: from nociception to pain perception: imaging the spinal and supraspinal pathways

    J. Anat.

    (2005)
  • Cited by (100)

    • Separation from a bonded partner alters neural response to inflammatory pain in monogamous rodents

      2022, Behavioural Brain Research
      Citation Excerpt :

      Human studies using brain imaging such as positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) have clarified brain regions activated by noxious stimuli causing pain, and the term “pain matrix” has been established to describe those brain regions [40,41]. The pain matrix commonly includes somatosensory cortex, prefrontal cortex, cingulate cortex, insula cortex, amygdala, thalamus, and many other brain regions, and activation of the pain matrix seems to be related to the perceived intensity of pain within and among individuals [42,43]. In animal studies, neuronal Fos, which is the product of the immediate-early gene c-fos, responds to direct stimulation by neurotransmitters and transsynaptic stimulation [44–46], and therefore Fos expression has long been used extensively as a marker of neuronal activation during the expression of various types of behavior including pain behaviors [47–52].

    • Learning Spatial-Spectral-Temporal EEG Representations with Deep Attentive-Recurrent-Convolutional Neural Networks for Pain Intensity Assessment

      2022, Neuroscience
      Citation Excerpt :

      Current research has explored various hand-crafted features to learn effective EEG representations for pain intensity assessment. Event-related Potential (ERP) occurs in response to laser stimulation has been proven effective for detecting pain (Huang et al., 2013), but it cannot adequately simulate clinical pain symptoms due to the brief stimulation (Hadjileontiadis, 2015). Moreover, Afrasiabi et al. (2020) found an increase in the coupling between cortical areas as pain increases, demonstrating that the amount of synchronization between EEG electrodes can also reveal pain-related events.

    View all citing articles on Scopus
    1

    These authors contributed equally.

    View full text