Edited by: Luke Henderson, University of Sydney, Australia
Reviewed by: Guilherme Lucas, University of São Paulo, Brazil; Chris David Keating, University of Hertfordshire, UK
Specialty section: This article was submitted to Autonomic Neuroscience, a section of the journal Frontiers in Neurology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Chronic pain is a pathological developing course of pain. In clinic, an objective indicator is needed for diagnosing and better controlling chronic pain. The abnormal neural responses in chronic pain are reflected by multiple event-related potentials (ERPs) in time, frequency, and location domain, respectively. However, multiple changes in ERPs are not applicable in clinic. So, the principal feature covered the most informative changes extracted from these three domains of ERP during the development of chronic pain is needed. In the present study, a parallel factor analysis method was employed to extract time–frequency–channel features of laser-evoked potential (LEP) simultaneously from rats with chronic inflammatory pain. Results showed that the main feature of LEP in channel domain locates in the frontal brain region in rats with chronic inflammatory pain while in the parietal brain region in control rats. In the frequency domain, the main frequency of LEP was significantly higher in chronic inflammatory pain rats than that in control rats. These findings indicate that the frontal region with higher frequency response to nociceptive information is the principal feature in the chronic pain state. Our study provided not only a principal feature of LEP but also a promising strategy for chronic pain, which is potential for clinic application.
Chronic pain is a pathological pain state characterized by pain persistence (
Event-related potentials (ERPs) are a measurement that reflects neuronal processes by frequency, time course, and topography changes (
Effective reduction methods such as principal component analysis (PCA) are traditionally applied for ERPs to explore the principal components and to characterize ERPs (
Therefore, in the present study, in order to explore the principal feature of LEP during the development of chronic inflammatory pain in three domains, we recorded LEP obtained from the electrocorticogram (ECoG) of rats with chronic pain model and applied the PARAFAC method to decompose multichannel LEP data. Components of each rat are the main characteristics of the time–frequency–channel information for brain oscillatory activities.
Thirteen male Sprague-Dawley adult rats (weight 300–350 g) were used. These animals were provided by the Department of Experimental Animal Sciences, Peking University Health Science Center. The animals were housed individually in cages at room temperature of 22 ± 1°C and kept on a 12 h light and dark cycle. Food and water were available
Rats were anesthetized with sodium pentobarbital (50 mg/kg,
A chronic inflammatory pain model of monoarthritis in rats was established according to a previously described method by Butler (
On the day prior to intra-articular injection (D0) and 1, 7, 14, and 28 days after intra-articular injection (D1, D7, D14, and D28), rats received thermal nociceptive stimulation with a laser beam (wavelength 10.6 µm, beam diameter 2.5 mm, pulse width 20 ms), which was delivered by a CO2-laser stimulator (DIMEI-300, Changchun Optics Medical Apparatus Co., Ltd., China). Laser stimuli were applied to the plantar of hind paw when rats were awake and quiet. The appropriate intensity of the laser beam for each individual rat was determined by using an ascending series of laser beam intensities with a 1 W increment. The intensity that generated four to five hind paw withdrawal responses out of six stimuli was selected as the intensity of stimulation. Each rat received 15 stimuli that could induce hind paw withdrawal responses. Each stimulus was targeted at slightly different positions. The interstimulation interval varies from 40 to 150 s.
The EEG/ERP system (CogniTrace ERP, ANT Inc., The Netherlands) was used for EEG data collection. Twelve recording electrodes and one reference electrode were connected to the digital preamplifier, and the ground electrode was connected the GND connector of the amplifier. All signals were referenced to the electrode that was located 2 mm caudal to lambda (the #13 electrode in Figure
The preparation and preprocessing of data were carried out as follows. The duration of each epoch was set as 1,500 ms (500 ms before and 1,000 ms after the laser stimulation onset). Large baseline drift was checked and removed for all trials and channels. Then, the epoch signals were re-referenced to an average of all channel recordings. Finally, the laser-evoked ECoG data were preprocessed with a band-pass filter of 1–70 Hz (eegfilter.m at EEGLAB software:
Wavelet transformation and PARAFAC were performed as described by Wang (
Then, the time points with biggest coefficients and the frequencies with biggest coefficients were selected. For further channel domain analysis, the coefficients of every channel were plotted.
Two-way analysis of variance was used to compare the frequency and time difference between the two groups.
In order to test the accuracy of PARAFAC method, the feature extracted by PARAFAC was compared with original LEP. With PARAFAC analysis, one or more components that represent main characteristics of EEG are obtained for each data. Within each component, it contains three-way information in the time–frequency–channel domain. A representative example with the PARAFAC analysis is shown in Figure
Table
Time |
|||||
---|---|---|---|---|---|
Group | D0 (%) | D1 (%) | D7 (%) | D14 (%) | D28 (%) |
CFA group |
100 | −32.4 ± 12.1 | −28.0 ± 15.7 | −35.0 ± 19.0 | −48.9 ± 15.2 |
NS group | 100 | −4.9 ± 12.0 | −5.3 ± 7.6 | −6.2 ± 6.5 | −6.5 ± 5.8 |
After the PARAFAC analysis, there were several components extracted by each rats. Each component consisted of three coefficients matrices, i.e., matrix of time, frequency, and channel, reflecting the main characteristics in the time, frequency, and channel domain, respectively. We further collected all the components in each group and then compared each of the three domains between two groups.
To find out the difference in the channel domain between the CFA and NS groups, coefficients of each channel from both the NS and CFA groups at different days before and after pain induction were plotted (Figure
The difference in the time domain and the difference in the frequency domain between the two groups were plotted in Figures
In the frequency domain, there was a statistically significant higher frequency response in the CFA group than the NS groups [
Previous studies showed that the PARAFAC was a promising approach to process multiple channel EEG signals (
Laser-evoked potential is dependent on intensity of nociceptive perception (
Based on the definition of chronic pain that pain lasts even after the original tissue damage has cured, the nature of chronic pain is the persistence of pain. Traditional researches that study changes at one time point of chronic pain are limited for providing the ongoing and developmental characteristic of chronic pain. In this study, we explored longitudinal study to find the dynamic changes of ERP features during the development of chronic pain.
The frontal and parietal brain regions were found to be the principal feature response to the laser nociceptive stimulation at day 0 (Figure
Results showed that the coefficients in the frontal region increased in rats with chronic inflammatory pain but did not change in the control rats. It indicates rats with chronic inflammatory pain process the nociceptive information predominantly in the frontal regions. It was in line with previous functional magnetic resonance imaging and electrophysiology studies indicating activity in the prefrontal region was increased in chronic pain condition (
The contribution of the frontal region seemed to become smaller over recording days in the NS control group though statistic results showed no difference. This phenomenon was probably due to pain habitation, i.e., reduction of pain and pain-related response by repetitive stimulation (
In the frequency domain, the frequency response was found to be higher in rats with CFA than the control. It is consistent with previous finding that the frontal network predominantly oscillated at higher frequencies in chronic pain (
Frequency response is usually studied as frequency bands, which could eliminate the individual variation. However, the edge of frequency band is artificial, and the trend of researches is to apply the dominant peak frequency rather than frequency band (
Our exploratory research has several limitations. First, LEP from ECoG of rats with chronic pain could not be transformed to patients, though rats study allows longitudinal study and high ratio of signal to noise. Further experiment from patients and ongoing pain would be performed. Second, the limited spatial resolution of rat ECoG and PARAFAC method is not allowed to be focused on specific brain areas, which is hard to explain the underlying neurophysiology mechanism. Third, the casual and unique relationship between these features and chronic pain is hardly determined from the current study. Intervention research is needed in future. Fourth, small sample size is a weak point because of long-time recordings.
In conclusion, we applied the PARAFAC analysis in the multichannel LEPs during the development of chronic pain. It was found that the frontal region with higher frequency was the principal feature of neural response in the chronic pain condition. These features provide a potential neural network characteristic for chronic pain. Besides, our study provides a promising strategy of applying LEP combined with PARAFAC for assisting chronic pain diagnosis and treatment.
All animal experiments were conducted with approval of the Animal Care and Use Committee of our university and were in accordance with the Guidelines of International Association for the Study of Pain.
JW (first author) carried out the ECoG recording and data analysis, participated in the study design, and drafted the manuscript. JW (second author) carried out the data analysis. YW participated in the study design and helped to draft the manuscript. XL conceived of the study, participated in its design, and helped to draft the manuscript. All the authors read and approved the final manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors are grateful to Mr. Douglas Richards for critical reading of the manuscript. This work was supported by grants from the National Natural Science Foundation of China (81230023, 81571067, and 81521063), the Ministry of Science and Technology of China (“973” Program, 2013CB531905), the “111 Project”(B07001), and the National High Technology Research and Development Program of China (863 Program, 2015AA020514).
The first author JW is funded by Beijing Municipal Research Fund for Outstanding Young Scholar (2014000020124G155).