Chronic pain results in an enormous societal and financial burden, not only from direct and indirect healthcare costs, but also from the toll of the opioid epidemic, in part arising from the lack of alternatives to treatment.1,2
Nonpharmacological, neuromodulation strategies like deep brain stimulation could be applied to refractory chronic pain if safe and effective brain targets are identified. The conscious experience of pain comprises sensory-discriminative, affective-motivational, and cognitive-evaluative dimensions, each of which are processed in distinct cerebral locations. Traditionally, regions that encode the sensory-discriminative components of pain, including the periaqueductal/periventricular gray matter and the somatosensory thalamus, have been targeted with deep brain stimulation, but the long-term success of these procedures has been inconsistent. It is possible that a more effective strategy would be to target areas that are involved with the psychological processes that modulate the context and interpretation of pain.3,4
Similar to the anterior cingulate cortex that has been targeted to treat patients with refractory chronic pain, the anterior part of the insula is also a part of the limbic system and is involved in affective-motivational and cognitive-evaluative brain functions that have fundamental roles in human awareness and pain perception.5–8 Intracranial electroencephalogram (EEG) recordings have demonstrated that nociceptive inputs are processed in the insula along the posterior-to-anterior direction, with each subarea related to a different dimension of pain;9 thus, whereas the posterior insula is involved in the sensory processing of pain, the anterior insula is likely to be involved in processing the subsequent affective and cognitive dimensions of pain.10–12 Neuroimaging studies have shown that neural underpinnings for the anticipation of pain, the determination of whether a stimulus is painful, and subjective pain intensity relate to the anterior part of the insula.6,13,14 Furthermore, patients with extensive damage within the insular region display asymbolia for pain, indicating a sensory-limbic disconnection of supraspinal pain processing.15
This pilot study aimed to determine whether direct stimulation of the anterior insula can alter cerebral pain processing in humans. The effects of anterior insula stimulation were evaluated by measuring heat pain thresholds, nociceptive-specific cutaneous laser-evoked potentials (LEPs), and continuous intracranial EEG recordings in a sample of patients with epilepsy who were undergoing intracranial monitoring for the localization of uncontrolled seizures.
Methods
Subjects and Intracerebral Electrode Implantation
This was an early-phase, nonrandomized, serial accrual trial in which subject responses after insular stimulations were evaluated against their own control responses obtained from baseline, sham, and noninsular cortical stimulations. Six patients with epilepsy without sensory dysfunction were recruited to participate in this study. Each subject was implanted with intracerebral electrodes customized to evaluate the presumed epileptogenic target for subsequent epilepsy surgery. Each subject was implanted with a depth electrode (Ad-Tech RD10R-SP05X-000: 10 contacts, 5-mm contact spacing, 0.86-mm diameter; AD-Tech Medical Instrument Corp.) in the left and/or right anterior insula—specifically the anterior insular gyrus.16 Continuous intracranial EEG recordings were collected at a 4096-Hz sampling rate (Natus Quantum LTM Amplifier; Natus Medical, Inc.). The study protocol was approved by the institutional review board at the University of Virginia, and all subjects gave their informed consent prior to the study.
Anterior Insula Stimulation
The direct anterior insula stimulations were performed in the inpatient epilepsy monitoring unit while subjects were fully awake, comfortable, and without sedating medications. Electrode contacts used for the anterior insula stimulation were localized using both postoperative T1-weighted MR and CT images (Fig. 1A). Insular stimulations consisted of three 10-second, bipolar stimulation trains separated by 2- to 5-second observation periods to evaluate for possible evoked seizures, resulting in a cumulative stimulation of 30 seconds (monophasic, 50 Hz or 100 Hz, 250-µsec pulse width, Nicolet Cortical Stimulator; Natus Neurology, Inc.). Stimulations at 2, 4, 8, and 10 mA were given, although some series were terminated at 8 mA because of concerns about evoking atypical seizures. The identical stimulation protocol was administered as control stimulations of noninsular cortex to evaluate stimulation specificity. The location of control stimulations varied by subject according to customized electrode placement, but eloquent cortex was avoided. Sham stimulations, during which no current was delivered, were also performed to control for potential placebo effects. The safety of these stimulations was evaluated by bedside physicians, real-time intracranial EEG monitoring, and electrocardiogram (ECG) recordings. Finally, cardiac rhythm was assessed for the last subject (case 6) following a 5-minute anterior insula stimulation at 10 mA to mimic future prolonged insular stimulations that may be used in further clinical study. ECG recordings were embedded as an independent channel in the EEG recordings.
Behavioral response to anterior insula stimulation. A: Intracerebral electrode placement for anterior insula stimulation. T1-weighted MR images indicate the electrode contacts (white arrowheads) used for stimulation and recording. B: Quantitative sensory testing for heat pain thresholds during baseline and following anterior insula stimulations. All 6 subjects had baseline and sham (i.e., 0-mA) stimulations as well as varying increments of stimulation intensity as described in Table 1. C: Control stimulation was performed in 3 subjects by stimulating either the superior frontal gyrus, left middle frontal gyrus, or middle temporal gyrus; see also Table 1. Neither sham nor control stimulations altered heat pain thresholds from baseline. D: Prolonged, 5-minute, 50-Hz and 100-Hz anterior insula stimulations in case 6. Data are summarized using the mean ± 95% confidence level.
Assessments of Heat Pain Threshold and Pain-Related Neurophysiology Analysis
Heat pain threshold was assessed with a computerized quantitative sensory testing system—a psychophysical method that uses a thermal probe (i.e., 3 × 3–cm thermode, TSA-II Quantitative NeuroSensory Analyzer; Medoc Ltd.) placed on the subject's volar forearm, contralateral to the side of brain stimulation. This thermode was programmed to increase in temperature at a rate of 1°C per second from a baseline temperature of 32°C. During heat pain threshold assessment, subjects were asked to push a button when the heat sensation first reached a painful level. Testing occurred within 10–30 seconds following designation of nonstimulation baseline or sham assessments or after each stepwise increase in insular or control stimulation. Subjects were given time to practice prior to the study. Heat pain thresholds were averaged across 5–10 trials, with probe repositioning to prevent sensitization, and approximately 20 seconds elapsed between heat pain threshold assessment trials.17 Changes in heat pain thresholds following stimulation were compared using 1-way ANOVA (α = 0.05). For each subject, a nonparametric Mann-Whitney test was performed during post hoc comparison for the heat pain threshold recorded first during baseline and then following anterior insula stimulation with the highest stimulation intensity (α = 0.05).
Time-frequency spectral analysis of the intracranial EEG from the anterior insula was conducted to reveal changes in ongoing brain rhythms following stimulations. This was performed using Fourier spectral decomposition (downsampled from 4096 Hz to 512 Hz, 1-second window, Hanning tapering window, and 50% overlap).18 Changes in stimulation-related EEG frequency band powers (i.e., delta [δ, < 4 Hz], theta [θ, 4–7 Hz], alpha [α, 8–12 Hz], beta [β, 13–35 Hz], and gamma [γ, > 40 Hz]) were compared between baseline and poststimulation intervals by using 1-way ANOVA (α = 0.05). Note that the anterior insula EEG recordings were obtained from the same electrode contact used for stimulations.
Painful cutaneous laser stimulation (Nd:YAP laser, DEKA Stimuli 1340; El-En, Inc.) was used to study nociceptive-related brain responses following the anterior insula stimulations, because the resulting LEP is a direct neurophysiology readout of nociceptive-related neural activity (i.e., the sum of excitatory and inhibitory postsynaptic potentials) for cerebral nociceptive processing. For each subject, the intensity of the laser was titrated prior to the study to 4–5 of 10 on the Numeric Pain Rating Scale (NPRS; 0 is no pain and 10 is the maximum pain imaginable). For each laser stimulation session, 40 laser stimuli (pulse duration 4 msec, spot diameter 5 mm, interstimulation interval 6–8 seconds, and output power 1.25–2.25 mJ) were delivered to the dorsal hand area contralateral to the brain stimulation side. The laser beam was slightly moved randomly within the dorsal hand area to avoid sensitization. Subjects were asked to rate the overall laser pain intensity at the end of the laser stimulation session by using the NPRS. Trials contaminated by artifacts were removed manually, and LEPs for each subject were computed by averaging the extracted laser stimulation epochs starting 200 msec before and 800 msec after the stimulation.19
Results
Six patients with refractory epilepsy were recruited to participate in this study (see Table 1 for patient information). Each subject had been implanted with intracerebral electrodes in the presumed epileptogenic targets for monitoring seizure activity, including a depth electrode in the left and/or right anterior insula (see Fig. 1A). The decision to perform invasive seizure monitoring was made solely based on clinical considerations. None of the subjects in this study had a history of somatosensory dysfunction, and no subject was found to have an epileptic focus within the insula. Patient toleration or the appearance of epileptic discharges limited stimulation protocol in some instances.
Anterior insula stimulation and pain assessments
| Case No. (age in yrs, sex) | Anterior Insula Stimulation | Quantitative Sensory Testing | NPRS Score for Laser Pain Intensity† | ||||
|---|---|---|---|---|---|---|---|
| Side | Intensity (mA) | Frequency (Hz) | Control Stimulation | Avg Baseline Heat Pain Threshold (°C) | Avg Heat Pain Threshold After Stimulation (°C)* | ||
| 1 (31, F) | Lt | 2, 4 | 50 | ND | 45.5 | 48.0 | ND |
| 2 (47, M) | Lt | 2, 4, 8, 10 | 50 | SFG | 41.3 | 50.6 | 4–5 vs 2 |
| 3 (58, M) | Lt | 2, 8 | 50 | ND | 43.0 | 51.1 | 4–5 vs 2 |
| 4 (44, M) | Rt | 2, 4 | 50 | MFG | 45.5 | 49.6 | ND |
| 5 (33, F) | Rt | 2, 4, 8 | 50 | MTG | 42.0 | 46.7 | 4–5 vs 1 |
| 6 (43, F) | Lt or Rt | 4 | 50, 100 | ND | 45.5 | 47.7 | ND |
Avg = average; MFG = middle frontal gyrus; MTG = middle temporal gyrus; ND = not done; SFG = superior frontal gyrus.
Results obtained with highest stimulation intensity.
Presented as prestimulation versus poststimulation scores.
Heat pain thresholds increased from baseline following anterior insula stimulation (Fig. 1B, p < 0.001) and were correlated with the stimulation intensity (regression analysis: β = 0.5712, standard error 0.070, p < 0.001). The subjects were not aware of the stimulations and did not report abnormal sensations. Control stimulations showed no significant changes in the heat pain thresholds (Fig. 1C, p = 0.561). The subject in case 6 underwent stimulation at both 50 Hz and 100 Hz with increased heat pain thresholds (p < 0.001), and no significant difference was found between the heat pain thresholds following 50-Hz versus 100-Hz stimulations (Fig. 1D, p = 0.729).
Time-frequency spectral decomposition analysis of the anterior insula EEG recordings showed significant changes in frequency band power of ongoing brain rhythms following stimulations (Fig. 2A). These stimulations decreased the proportion of total power contributed by the frequency bands greater than the delta band (< 3 Hz), and these diminishments in the power of faster frequencies persisted for more than 3 minutes following stimulation (Fig. 2A–C, p < 0.01).
Neurophysiological response to anterior insula stimulation. A: Time-frequency spectral analysis for the intracranial EEG recordings obtained during insular stimulation (subject in case 2). The x-axis and y-axis denote time and frequency, respectively. The color bar translates the magnitude of frequency power (in dB) to the corresponding color. The white horizontal dashed line indicates 10 Hz on the time-frequency plot. In this example (subject in case 2), direct stimulations are depicted by the vertical dark red artifacts in the plot. B: Significant changes in EEG frequency band power were found following stimulations (most visibly evident by a decrease in red intensity at 10 Hz and below) and quantitatively shown, with the most significant decreases seen in the lines falling below the baseline 95th percentile power spectra following higher stimulation intensities. C: Changes in EEG spectral power normalized to baseline for low (< 3 Hz, white box) and high (≥ 3 Hz, gray box) frequency bands for all subjects in this study. Low-frequency band powers were found to increase and high-frequency band powers were found to decrease following anterior insula stimulation with high stimulation intensity. D: LEPs demonstrate loss of the N2 and P2 after anterior insula stimulations E: The average heart rate was compared in case 6 as a function of change following insular stimulation; no significant change occurred. Max. = maximum; min. = minimum; stim = stimulation.
Supraspinal LEPs typically consist of a negative peak at approximately 245 msec (N2 component) followed by a positive peak (P2) at approximately 310 msec following hand stimulation. Anterior insula stimulation attenuated LEP N2P2 peaks (i.e., N2-P2 amplitude; Fig. 2D) and reduced pain intensity ratings from NPRS 5 to NPRS 1–2 (Table 1).
ECG analysis was performed by comparing average heart rate per minute with and without anterior insula stimulation. No subject showed significant heart rate changes following stimulation (cases 1–5, p > 0.05) or during 10-second or 5-minute stimulations in case 6 (Fig. 2E, p = 0.134).
Discussion
In this pilot study, we demonstrated that direct stimulation of the anterior insula inhibited nociceptive information processing in the brain and profoundly increased the heat pain threshold in humans. Specifically, brief electrical stimulations over the anterior part of the insula increased the threshold of thermal nociception in a dose-dependent fashion without producing abnormal sensations or altering cardiac function. While the study was not designed to compare pain reduction modalities, the change of heat pain threshold observed exceeds those reported with morphine or other opioid analgesics. Moreover, laser acute pain intensity also decreased in all subjects, similar to those laser acute pain studies with analgesics,20 and cognitive tasks.21
The mechanisms of neuromodulation conferred by direct anterior insula stimulation have yet to be elucidated, but our neurophysiology findings of high-frequency EEG band power diminishment and nociceptive-specific evoked potential (i.e., LEP) attenuation following stimulation confirm a significant role of the anterior insula in pain processing.5–7 Given that the LEPs represent nociceptive-related information processing at the supraspinal level, the observed attenuation of LEPs is probably a result of inhibition of higher-order nociceptive information processing (perhaps within the anterior insula). Importantly, the observed behavioral and neurophysiological effects outlasted the stimulation, and direct stimulation of the anterior insula was safe and did not evoke sensations, mood changes, or cardiovascular abnormalities. Conversely, stimulation-evoked painful sensations have been reported with mid-insula and/or posterior insula stimulation.22
There are limitations to this pilot study. First, standardized neuropsychological evaluations of mood and cognition were not used for characterization of possible stimulation-related behavioral changes. Mood changes have been shown to occur immediately following stimulations of the anterodorsal part of the insular cortex, subgenual cingulate region, subthalamic nucleus, and globus pallidus. Second, the sample size of this study is limited by the number of patients with insular electrodes, and electrode placement beyond the anterior insula varied with each patient's clinical needs. We were not able to control the order of stimulation intensity; to compare the different effects of contralateral, ipsilateral, and bilateral stimulations; or to maintain a consistent control stimulation site. Despite the small sample size in the present study, an increased heat pain threshold following anterior insula stimulations was found by comparisons with sham and control stimulations, and was observed in all subjects. Finally, although the implication exists, it remains unknown whether acute anterior insula stimulation can be used chronically as a long-term neuromodulatory target for treatment of chronic pain. The possible treatment effects of anterior stimulation for refractory chronic pain remain to be demonstrated in a larger clinical study.
Conclusions
Safe, nonpharmacological therapies are imperative for the future management of chronic pain conditions and to mitigate the ongoing opioid crisis. Recent advances in neuromodulatory devices are capable of delivering complex stimulation paradigms in either open or closed-loop modalities, increasing the potential for brain stimulation as a treatment for refractory chronic pain. The anterior insula offers promise as a therapeutic site for investigation, given that this study provides direct validation of its role in pain perception.
Acknowledgments
This study was supported by the University of Virginia Brain Institute, Departments of Neurology and Neurosurgery, and the School of Medicine.
We thank Mr. Jeremy Winberry and Dr. Andrea Franzini for their assistance with data acquisition and helpful comments for our manuscript.
Disclosures
Dr. Elias is a consultant for Second Sight, and he received support of a non–study-related clinical or research effort that he oversaw from Insightec.
Author Contributions
Conception and design: Liu, Elias. Acquisition of data: Moosa, Quigg. Analysis and interpretation of data: all authors. Drafting the article: Liu. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Liu. Statistical analysis: Liu. Administrative/technical/material support: all authors. Study supervision: Liu, Quigg, Elias.
References
- 1↑
Seth P, Rudd RA, Noonan RK, Haegerich TM. Quantifying the epidemic of prescription opioid overdose deaths. Am J Public Health. 2018;108(4):500–502.
- 2↑
Dahlhamer J, Lucas J, Zelaya C, et al. Prevalence of chronic pain and high-impact chronic pain among adults−United States, 2016. MMWR Morb Mortal Wkly Rep. 2018;67(36):1001–1006.
- 3↑
Spooner J, Yu H, Kao C, et al. Neuromodulation of the cingulum for neuropathic pain after spinal cord injury. Case report. J Neurosurg. 2007;107(1):169–172.
- 4↑
Boccard SGJ, Prangnell SJ, Pycroft L, et al. Long-term results of deep brain stimulation of the anterior cingulate cortex for neuropathic pain. World Neurosurg. 2017;106(625):637.
- 5
Craig AD. How do you feel—now? The anterior insula and human awareness. Nat Rev Neurosci. 2009;10(1):59–70.
- 6↑
Coghill RC, Talbot JD, Evans AC, et al. Distributed processing of pain and vibration by the human brain. J Neurosci. 1994;14(7):4095–4108.
- 7
Starr CJ, Sawaki L, Wittenberg GF, et al. Roles of the insular cortex in the modulation of pain: insights from brain lesions. J Neurosci. 2009;29(9):2684–2694.
- 8
Greenspan JD, Lee RR, Lenz FA. Pain sensitivity alterations as a function of lesion location in the parasylvian cortex. Pain. 1999;81(3):273–282.
- 9↑
Frot M, Faillenot I, Mauguière F. Processing of nociceptive input from posterior to anterior insula in humans. Hum Brain Mapp. 2014;35(11):5486–5499.
- 10
Duerden EG, Albanese MC. Localization of pain-related brain activation: a meta-analysis of neuroimaging data. Hum Brain Mapp. 2013;34(1):109–149.
- 11
Schweinhardt P, Bushnell MC. Pain imaging in health and disease—how far have we come?. J Clin Invest. 2010;120(11):3788–3797.
- 12
Craig AD. Interoception: the sense of the physiological condition of the body. Curr Opin Neurobiol. 2003;13(4):500–505.
- 13↑
Craig AD, Chen K, Bandy D, Reiman EM. Thermosensory activation of insular cortex. Nat Neurosci. 2000;3(2):184–190.
- 14↑
Ploner M, Lee MC, Wiech K, et al. Prestimulus functional connectivity determines pain perception in humans. Proc Natl Acad Sci U S A. 2010;107(1):355–360.
- 15↑
Berthier M, Starkstein S, Leiguarda R. Asymbolia for pain: a sensory-limbic disconnection syndrome. Ann Neurol. 1988;24(1):41–49.
- 16↑
Tanriover N, Rhoton AL Jr, Kawashima M, et al. Microsurgical anatomy of the insula and the sylvian fissure. J Neurosurg. 2004;100(5):891–922.
- 17↑
Greenspan JD. Quantitative assessment of neuropathic pain. Curr Pain Headache Rep. 2001;5(2):107–113.
- 18↑
Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods. 2004;134(1):9–21.
- 19↑
Liu C-C, Ohara S, Franaszczuk PJ, et al. Attention to painful cutaneous laser stimuli evokes directed functional interactions between human sensory and modulatory pain-related cortical areas. Pain. 2011;152(12):2781–2791.
- 20↑
Nielsen JC, Bjerring P, Arendt-Nielsen L, Petterson KJ. A double-blind, placebo controlled, cross-over comparison of the analgesic effect of ibuprofen 400 mg and 800 mg on laser-induced pain. Br J Clin Pharmacol. 1990;30(5):711–715.
- 21↑
Beydoun A, Morrow TJ, Shen JF, Casey KL. Variability of laser-evoked potentials: attention, arousal and lateralized differences. Electroencephalogr Clin Neurophysiol. 1993;88(3):173–181.
- 22↑
Ostrowsky K, Magnin M, Ryvlin P, et al. Representation of pain and somatic sensation in the human insula: a study of responses to direct electrical cortical stimulation. Cereb Cortex. 2002;12(4):376–385.
