Figures
Abstract
The purpose of this study was to use functional near-infrared spectroscopy (fNIRS) to examine patterns of both activation and deactivation that occur in the frontal lobe in response to noxious stimuli. The frontal lobe was selected because it has been shown to be activated by noxious stimuli in functional magnetic resonance imaging studies. The brain region is located behind the forehead which is devoid of hair, providing a relative ease of placement for fNIRS probes on this area of the head. Based on functional magnetic resonance imaging studies showing blood-oxygenation-level dependent changes in the frontal lobes, we evaluated functional near-infrared spectroscopy measures in response to two levels of electrical pain in awake, healthy human subjects (n = 10; male = 10). Each subject underwent two recording sessions separated by a 30-minute resting period. Data collected from 7 subjects were analyzed, containing a total of 38/36 low/high intensity pain stimuli for the first recording session and 27/31 pain stimuli for the second session. Our results show that there is a robust and significant deactivation in sections of the frontal cortices. Further development and definition of the specificity and sensitivity of the approach may provide an objective measure of nociceptive activity in the brain that can be easily applied in the surgical setting.
Citation: Aasted CM, Yücel MA, Steele SC, Peng K, Boas DA, Becerra L, et al. (2016) Frontal Lobe Hemodynamic Responses to Painful Stimulation: A Potential Brain Marker of Nociception. PLoS ONE 11(11): e0165226. https://doi.org/10.1371/journal.pone.0165226
Editor: André Mouraux, Université catholique de Louvain, BELGIUM
Received: April 6, 2016; Accepted: October 7, 2016; Published: November 2, 2016
Copyright: © 2016 Aasted et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: Research reported in this article was supported by the Mayday Foundation, the Anesthesia Research Distinguished Trailblazer Award, and the National Institute of General Medical Science of the National Institutes of Health under award number R01GM104986. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Competing interests: D.A.B. is an inventor on a technology licensed to TechEn, a company whose medical pursuits focus on noninvasive optical brain monitoring. D.A.B.’s interests were reviewed and are managed by Massachusetts General Hospital and Partners HealthCare in accordance with their conflict of interest policies. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
Introduction
While subjective measures of pain may be helpful in the clinic, there are a number of clinical conditions where subjects are either drowsy or unconscious during tissue damage, such as surgery. Providing adequate anesthesia during surgery is routine, but we do not have a good measure of nociceptive stimuli that may act on brain systems. Our prior report suggests that during anesthesia, activation of certain brain regions may take place during nociceptive stimulation[1]. Frontal lobe activation during nociceptive stimulus has been reported by others. [2–4]. Here we wished to extend these findings using functional near infrared spectroscopy (fNIRS) with more complete coverage of the frontal lobe in order to determine an optimal frontal area in which such signals are most easily defined. As summarized in our prior report[1], frontal lobe activity is involved in nociceptive signaling. This has implications for placement of probes with regard to signal detection and ease of placement in conditions such as evaluation of pain in clinical conditions (e.g., under anesthesia). We also have previously reported activation in the somatosensory cortex to nociceptive stimuli, confirming the nature of the parallel observations in the frontal lobe for the same stimulus using fNIRS [5].
fNIRS provides a non-invasive approach for the study of cortical cerebral hemodynamic fluctuations by passing two harmless wavelengths of light, such as 690 nm and 830 nm[5], through the scalp, cerebral spinal fluid, and cortical volumes. As this light travels, it is altered by oxygenated and deoxygenated hemoglobin, which are among the dominant absorbers of near-infrared light in biological tissue [6]. The backscattered light is detected by photodiodes and through the application of the modified Beer-Lambert Law it is possible to calculate changes in cerebral concentrations of deoxygenated (HbR) and oxygenated (HbO) hemoglobin, as well as total hemoglobin concentration (HbT) [1].
The experience of pain involves a number of brain regions, several of which are located in the cortical regions observable with fNIRS. The rationale for evaluating the frontal lobes was also considered in light of multiple functional magnetic resonance imaging (fMRI) studies of pain or nociception reporting activation in the frontopolar prefrontal cortex, including Brodamann Area 10. [7, 8–12]. Decreased activation to noxious painful stimuli has been observed in such nociceptive evoked pain studies (viz., [11]). Furthermore, our previous work indicated that a frontal lobe cortical response could be used to differentiate between low and high pain intensity stimuli, as we were able to do over the somatosensory cortex[5]. Placement of fNIRS sensors on the forehead would make the application of the technology easy to use in clinical situations (e.g., sedation or anesthesia). The objective of the present study was to evaluate responses to a noxious stimulus and determine if (1) specific regional activation using fNIRS over two polarfrontal cortical (medial and lateral) regions for the same noxious stimulus rated for two levels of pain intensity (low and high); and (2) whether there are any differences between the responses to noxious stimuli in these two regions.
Methods
The study was approved by the Institutional Review Board (IRB) of the Massachusetts General Hospital and met the scientific and ethical guidelines for human pain research of the Helsinki Accord and the International Association for the Study of Pain.
NIRS Probe and System
Cerebral hemodynamic activity was recorded using a multichannel functional near-infrared spectrometer operating at 690 and 830 nanometer wavelengths (TechEn Inc. MA, USA, CW7 System). The probe contained 12 sources, 12 standard separation detectors, and 12 short separation detectors. Fig 1 illustrates the layout of the sources and standard separation detectors. Standard separation detectors were positioned 30 millimeters from adjacent sources and short separation detectors were located 8 millimeters from a single source.
The sensitivity of each probe to detecting brain hemodynamics is represented on a logarithmic color scale ranging from 0 dB (red) to -40 dB (blue) times the maximum sensitivity. The source locations (red numbers), detector locations (blue numbers), and channels (green lines) are further detailed in Fig 2.
Study protocol and subject population
Ten healthy subjects were included in the study (right handed, male, mean age 24.4 ± 5.8 years). Each subject gave informed written consent prior to the experiments. Subjects with a history of neurological trauma or psychiatric disorders, or who were unable to remain still for 6 consecutive minutes, were excluded.
Pain Stimuli and Ratings
Prior to performing the experiment, a 5 Hz electrical stimulation was applied to each subject’s left thumb (Neurometer CPT, Neurotron, Baltimore, MD). For subjective measures of pain, we used the verbal rating scale—verbal analogue scale[13], which has been shown to be a reliable scale[14] as measure of pain intensity. Details of the pain ratings were explained to the subjects prior to the stimuli being applied: “on an 11 point scale where 0 = no pain and 10 = maximal pain please define the stimulus that produces a 3/10 pain level and in a separate rating a 7/10 pain level to an increasing electrical stimulus”. Thus, we evaluated two levels of pain intensity: low (3/10) and high (7/10) intensities in each individual. The levels of stimulus for each rating in each individual were then used during the NIRS recordings. During the experiment, semi-randomized sequences of the low and high pain stimuli, lasting 5 seconds in duration, separated by 25-second periods of rest, were applied in two 6-minute sets, separated by a 30-minute resting period. This resulted in a total of 12 low intensity (3/10) and 12 high intensity (7/10) noxious stimuli in a sequence, with equal numbers of each per 3-minute section, for use in the analysis.
Data Processing Stream
Analysis was carried out using the open source software Homer2, which is implemented in Matlab (Mathworks, Natick, MA), and the process described in Yücel et al., 2015[5]. The processing stream parameters that were used are detailed in Table 1. The sensor channels were split into two groups to evaluate whether central or lateral regions respond more to nociceptive stimulus, as depicted in Fig 2. The ‘central’ section was composed of the 12 source-detector pairs listed in Table 2 and predominately covered the superior frontal gyri, while the ‘lateral’ results included the remaining 8 source-detector pairs on each side, for a total of 16 channels (Table 3), predominately covering middle and inferior frontal gyri. Some overlap occurs because the probe channels were selected independent of their cortical projections. Each of the 6-minute runs was split to separately analyze the first and second three minutes to assess whether a habituation effect occurs.
Please see Homer2 documentation for further information on function parameters (http://homer-fnirs.org/tutorials/).
Statistics
Statistical analysis was performed using MATLAB to perform paired t-tests to determine whether there was a significant difference between the two stimulus intensities across subjects for each test segment. Statistical significance was assessed with a p-value of 0.05 corrected for multiple comparisons (p<0.0125) across medial and lateral averaged source-detector pair signals and across 2 time points.
Results
Frontal Lobe Responses
Using the automated signal-to-noise ratio and artifact detection algorithms in Homer2, the data from three subjects were identified as not meeting the minimum criteria for signal quality based on the criteria in Table 1 and were excluded from the group analysis. In most cases of this type, exclusion occurs as the result of weak signal strength or channel saturation (raw detector signal values under 1e4 or over 1e7) or excessive subject motion (standard deviation of raw detector signal greater than 50 or amplitude change greater than 5 over a 0.5-second interval). The remaining seven data sets contained a total of 65 usable low pain intensity stimuli and 67 high pain intensity stimuli (see Table 4), and were then analyzed to determine the group mean hemodynamic response over the central and lateral probe sections presented in Fig 3. The pattern of hemodynamic activity across the central channels during the first three stimuli of each type for the first run produced the hemodynamic response that best replicated the results obtained in Yücel et al., 2015[5]. Averaging the change in oxygenated hemoglobin across the time interval 10 to 14 seconds post-stimulus produces a statistically significant differentiation between the two stimuli (i.e., VAS 3/10 vs. VAS 7/10).
Top Left: The figure shows the group average hemodynamic response observed from central frontal channels, predominantly positioned over the superior frontal cortices. Bottom Left: Hemodynamic response observed from lateral channels, predominantly positioned over the bilateral medial frontal cortices. Oxygenated (red) and deoxygenated hemoglobin (blue) are depicted for the response to high (7/10; solid line) and low (3/10; dashed line) stimuli with standard error bars. Right top and bottom: The mean oxygenated hemoglobin response for each subject over the period 10–14 seconds post-stimulus, observed from central frontal channels (top) and from lateral channels (bottom). The number of stimuli in the analysis for each subject can be found in Table 4. The black line is a visual aid with a slope of one and zero y-intercept. A verbal analogue scale (VAS) was used to determine pain levels corresponding to a 3/10 (low pain intensity) and 7/10 (high pain intensity).
Changes in Frontal Response Over Time
To further investigate the effect of habituation on the brain’s response to noxious stimuli, the same process that was used to distinguish high pain intensity from low pain intensity stimuli during the first three stimuli of each type, was applied to each of the four time segments, separately analyzing the central and lateral regions. From this analysis it was determined that there was a distinguishable difference between the two stimulation intensities over the central channels for both the first and second sets of stimuli during the first 6-minute session, but that the difference lost significance during the second 6-minute period of stimuli (Table 5). No statistically significant difference was found between the HbO response for the lateral channels during any of the time segments (Table 6).
Discussion
In this study we demonstrated that the frontal lobe metric is robust and reproducible and we determined that the signal of interest is most strongly present in the centrally positioned optodes. Specifically, we determined that the relative change in oxygenated hemoglobin concentration over the superior frontal cortices sampled by the central channels is correlated with the perception of pain from electrical nerve stimulation.
Activation of the Frontal Lobes
As shown in the methods and results (see Fig 1), measures of the fNIRS signals captured with the central channels cover the surface of the superior frontal gyrus. Prior PET and fMRI studies have reported activation in the superior frontal lobe in experimental pain studies[15]. The function of the region includes non-pain processing, including working memory[16], and its role in pain has been established from a number of studies, for example, subdural electrophysiological measures using evoked painful stimuli[17] and fMRI studies on pain [18]. Of these studies, the electrophysiological study is perhaps the most salient as it comes closest to the approach we have taken here. The function in pain/nociceptive processing remains unknown but putative processes include cognitive evaluating prediction errors for aversive stimuli[19]. In this study we observed a decrease in activity correlated with noxious stimulation as well as a reduction in this effect that may be attributed to stimulus salience. Periodontal pain stimulation has been reported to produce a similar decrease in deoxygenated hemoglobin concentration in the frontal cortices [3]. However, studies utilizing a cold pressor test or mechanical force to induce pain have indicated the hemodynamic response may be specific to the type of nociceptive stimulus [4, 20]. Differences between the electrical stimuli that produced low and high pain reports were significant for the central channels but fell short of significance in the laterally placed channels situated on the forehead. Furthermore, the use of short separation channels enhanced the signal compared to using standard separation channels alone. Specifically, using the short separation channels for noise regression substantially improved the ability to distinguish the brain response to noxious stimuli versus autonomic signals in the skin[21].
Given that the frontal regions are involved in a number of processes (viz., multimodal threats and salient stimuli), it is possible that the response we observe may not be specific to nociception. The main frontal region evaluated included the polarfrontal prefrontal cortex or Brodmann Area 10 (see Caveats). Brodman area 10 (BA10) located in the anterior medial frontal cortex. The region is located inferior to Brodmann 9 (superior frontal cortex) and Brodmann 11 (the anterior extension of oribitofrontal cortex). BA 10, also known as the anterior prefrontal cortex, frontopolar prefrontal cortex or rostral prefrontal cortex, is an association cortex is involved in a wide variety of functions including risk and decision making [22], odor evaluation, reward and conflict/threats [23, 24], pain [10], and working memory [25, 26]. While arousal is probably not involved in the response, or only partially contributory in the awake patients and even in the sedated patients, it is unlikely to be so in the fully anesthetized patients (see Caveats below). Furthermore, concomitant with frontal lobe deactivation we observe a parallel increased activation in the primary somatosensory cortex[5, 27].
Temporal Evaluation of Responses
While we expected to see no differences between the stimulation periods, we observed a diminished response to pain during the second stimulus period, following the 30-minute period of rest. The ability to distinguish the two intensity stimuli in the frontal cortex persists past the first three minutes, unlike the response observed in the somatosensory cortex in earlier studies, where the signal diminished during the second three minutes due to a habituation effect when using the same stimuli[5]. The neurological processes that lead to this decreased response from repeated exposure are well known and include enhanced modulation via frontal-periaqueductal connections [28].
Benefits of Data acquisition from Forehead
In the clinical settings measures over the frontal area important have substantial advantages over a somatosensory-only monitoring system. Potentially the most significant of these is ease of access. fNIRS requires a reliable interface between the optical equipment and the subject’s scalp, which is complicated by the presence of hair. When hair is present it may require extensive adjustment of the equipment in order to create a reliable optical connection. This could limit widespread use of the technology in medical applications. A second advantage is redundancy. Because measurements from biological processes are notoriously noisy and are often prone to artifacts and unexpected behaviors, by monitoring two brain regions instead of one, when possible, fNIRS has the potential to self-verify and reduce the prevalence of false-positives.
Caveats
We wish to point out a number of caveats related to the study. (1) Specificity of Response: We do not know that the response is specific to pain. Since we did not utilize non-nociceptive control stimuli, it is not possible to exclude that the observed differences in the responses could be related to differences in stimulus intensity and/or salience for the frontal response in this study. However, while we have reported a similar frontal lobe response to noxious stimuli under anesthesia (see conclusions), the issue of salience during these states is still not well understood. It has been reported that there is a breakdown of functional connectivity during anesthesia but not sensory connectivity[29]. We are unaware of electrophysiological measures of neurons in the frontal lobe regions evaluated in this study. Having said this, it is clear that further evaluation of the experimental process to determine the contribution of salient processing in this paradigm is necessary. Furthermore, additional features may contribute to the observed signal including arousal (salience) or anxiety. (2) Potential Variability of Measures in a Clinical Setting: While we have reported on the use of NIRS in measuring pain under anesthesia[30] or sedation[1], there are still issues related to the measures of evoked pain that need clarification including: 2.1. Habituation of a pain response; 2.2. Differences in evoked vs. ongoing pain; 2.3. The depth of anesthesia or use of opioid analgesics; 2.4. Consistency of measures including signal quality in individual patients/subjects. (3) The need for further Studies: This is a preliminary study and further studies are needed to control for highly salient non-nociceptive stimuli. (4) Specificity of Brain Region: Connectivity based parcellation of human frontal polar cortex has previously been described [31]. We did not define the specific area, but our medial probes covered the region and the lateral probes may have included frontal regions just outside the brain area.
Conclusion
While we currently interpreted these results as a primary response in the frontal lobe region to a noxious stimuli, it may be a secondary response affecting other networks (e.g., Salience Network[32]). Clearly further evaluation of the sensitivity and specificity of this measure, if successful, may provide a potential marker for evoked nociceptive activity in humans.
Author Contributions
- Conceptualization: DB LB.
- Data curation: SCS.
- Formal analysis: CMA LB.
- Funding acquisition: DB DAB.
- Investigation: SCS.
- Methodology: CMA MAY DAB.
- Project administration: DB DAB LB.
- Resources: DB DAB.
- Software: DAB.
- Supervision: DB LB.
- Writing – original draft: CMA DB.
- Writing – review & editing: CMA MAY SCC KP DAB LB DB.
References
- 1. Becerra L, Aasted CM, Boas DA, George E, Yucel MA, Kussman BD, et al. Brain measures of nociception using near-infrared spectroscopy in patients undergoing routine screening colonoscopy. Pain. 2016 Apr;157(4):840–8. pmid:26645550. Pubmed Central PMCID: 4794375.
- 2. Barati Z, Shewokis PA, Izzetoglu M, Polikar R, Mychaskiw G, Pourrezaei K. Hemodynamic response to repeated noxious cold pressor tests measured by functional near infrared spectroscopy on forehead. Annals of biomedical engineering. 2013 Feb;41(2):223–37. pmid:22956158.
- 3. Sakuma S, Inamoto K, Higuchi N, Ariji Y, Nakayama M, Izumi M. Experimental pain in the gingiva and its impact on prefrontal cortical hemodynamics: a functional near-infrared spectroscopy study. Neuroscience letters. 2014 Jul 11;575:74–9. pmid:24878385.
- 4. Lee CH, Sugiyama T, Kataoka A, Kudo A, Fujino F, Chen YW, et al. Analysis for distinctive activation patterns of pain and itchy in the human brain cortex measured using near infrared spectroscopy (NIRS). PloS one. 2013;8(10):e75360. pmid:24098378. Pubmed Central PMCID: 3789686.
- 5. Yucel MA, Aasted CM, Petkov MP, Borsook D, Boas DA, Becerra L. Specificity of hemodynamic brain responses to painful stimuli: a functional near-infrared spectroscopy study. Scientific reports. 2015;5:9469. pmid:25820289. Pubmed Central PMCID: 4377554.
- 6. Pourshoghi A, Danesh A, Tabby DS, Grothusen J, Pourrezaei K. Cerebral reactivity in migraine patients measured with functional near-infrared spectroscopy. European journal of medical research. 2015;20:96. pmid:26644117. Pubmed Central PMCID: 4672549.
- 7. Porro CA, Cettolo V, Francescato MP, Baraldi P. Temporal and intensity coding of pain in human cortex. Journal of neurophysiology. 1998 Dec;80(6):3312–20. pmid:9862924.
- 8. Hautvast RW, Ter Horst GJ, DeJong BM, DeJongste MJ, Blanksma PK, Paans AM, et al. Relative changes in regional cerebral blood flow during spinal cord stimulation in patients with refractory angina pectoris. The European journal of neuroscience. 1997 Jun;9(6):1178–83. pmid:9215701.
- 9. Svensson P, Minoshima S, Beydoun A, Morrow TJ, Casey KL. Cerebral processing of acute skin and muscle pain in humans. Journal of neurophysiology. 1997 Jul;78(1):450–60. pmid:9242293.
- 10. Cauda F, Sacco K, Duca S, Cocito D, D'Agata F, Geminiani GC, et al. Altered resting state in diabetic neuropathic pain. PloS one. 2009;4(2):e4542. pmid:19229326. Pubmed Central PMCID: 2638013.
- 11. Lobanov OV, Quevedo AS, Hadsel MS, Kraft RA, Coghill RC. Frontoparietal mechanisms supporting attention to location and intensity of painful stimuli. Pain. 2013 Sep;154(9):1758–68. pmid:23711484. Pubmed Central PMCID: 3755961.
- 12. Kucyi A, Moayedi M, Weissman-Fogel I, Goldberg MB, Freeman BV, Tenenbaum HC, et al. Enhanced medial prefrontal-default mode network functional connectivity in chronic pain and its association with pain rumination. The Journal of neuroscience: the official journal of the Society for Neuroscience. 2014 Mar 12;34(11):3969–75. pmid:24623774.
- 13. Hjermstad MJ, Fayers PM, Haugen DF, Caraceni A, Hanks GW, Loge JH, et al. Studies comparing Numerical Rating Scales, Verbal Rating Scales, and Visual Analogue Scales for assessment of pain intensity in adults: a systematic literature review. Journal of pain and symptom management. 2011 Jun;41(6):1073–93. pmid:21621130.
- 14. Williamson A, Hoggart B. Pain: a review of three commonly used pain rating scales. Journal of clinical nursing. 2005 Aug;14(7):798–804. pmid:16000093.
- 15. Peyron R, Laurent B, Garcia-Larrea L. Functional imaging of brain responses to pain. A review and meta-analysis (2000). Neurophysiologie clinique = Clinical neurophysiology. 2000 Oct;30(5):263–88. pmid:11126640.
- 16. du Boisgueheneuc F, Levy R, Volle E, Seassau M, Duffau H, Kinkingnehun S, et al. Functions of the left superior frontal gyrus in humans: a lesion study. Brain: a journal of neurology. 2006 Dec;129(Pt 12):3315–28. pmid:16984899.
- 17. Lenz FA, Rios M, Zirh A, Chau D, Krauss G, Lesser RP. Painful stimuli evoke potentials recorded over the human anterior cingulate gyrus. Journal of neurophysiology. 1998 Apr;79(4):2231–4. pmid:9535984.
- 18. Symonds LL, Gordon NS, Bixby JC, Mande MM. Right-lateralized pain processing in the human cortex: an FMRI study. Journal of neurophysiology. 2006 Jun;95(6):3823–30. pmid:16554508.
- 19. Ploghaus A, Tracey I, Clare S, Gati JS, Rawlins JN, Matthews PM. Learning about pain: the neural substrate of the prediction error for aversive events. Proceedings of the National Academy of Sciences of the United States of America. 2000 Aug 1;97(16):9281–6. pmid:10908676. Pubmed Central PMCID: 16859.
- 20. Barati Z, Zakeri I, Pourrezaei K. Functional data analysis view of functional near infrared spectroscopy data. Journal of biomedical optics. 2013 Nov;18(11):117007. pmid:24247748.
- 21. Yucel MA, Selb J, Aasted CM, Petkov MP, Becerra L, Borsook D, et al. Short separation regression improves statistical significance and better localizes the hemodynamic response obtained by near-infrared spectroscopy for tasks with differing autonomic responses. Neurophotonics. 2015 Jul;2(3):035005. pmid:26835480. Pubmed Central PMCID: 4717232.
- 22. Anticevic A, Repovs G, Barch DM. Resisting emotional interference: brain regions facilitating working memory performance during negative distraction. Cognitive, affective & behavioral neuroscience. 2010 May;10(2):159–73. pmid:20498341. Pubmed Central PMCID: 3856369.
- 23. Wheelock MD, Sreenivasan KR, Wood KH, Ver Hoef LW, Deshpande G, Knight DC. Threat-related learning relies on distinct dorsal prefrontal cortex network connectivity. NeuroImage. 2014 Nov 15;102 Pt 2:904–12. pmid:25111474. Pubmed Central PMCID: 4252829.
- 24. Gold AL, Morey RA, McCarthy G. Amygdala-prefrontal cortex functional connectivity during threat-induced anxiety and goal distraction. Biological psychiatry. 2015 Feb 15;77(4):394–403. pmid:24882566. Pubmed Central PMCID: 4349396.
- 25. Minamoto T, Yaoi K, Osaka M, Osaka N. The rostral prefrontal cortex underlies individual differences in working memory capacity: An approach from the hierarchical model of the cognitive control. Cortex; a journal devoted to the study of the nervous system and behavior. 2015 Oct;71:277–90. pmid:26280275.
- 26. Gohil K, Dippel G, Beste C. Questioning the role of the frontopolar cortex in multi-component behavior—a TMS/EEG study. Scientific reports. 2016;6:22317. pmid:26924655. Pubmed Central PMCID: 4770427.
- 27. Becerra L, Harris W, Joseph D, Huppert T, Boas DA, Borsook D. Diffuse optical tomography of pain and tactile stimulation: activation in cortical sensory and emotional systems. NeuroImage. 2008 Jun;41(2):252–9. pmid:18394924. Pubmed Central PMCID: 2728450.
- 28. Mantyh PW. Forebrain projections to the periaqueductal gray in the monkey, with observations in the cat and rat. The Journal of comparative neurology. 1982 Apr 1;206(2):146–58. pmid:7085925.
- 29. Bonhomme V, Vanhaudenhuyse A, Demertzi A, Bruno MA, Jaquet O, Bahri MA, et al. Resting-state Network-specific Breakdown of Functional Connectivity during Ketamine Alteration of Consciousness in Volunteers. Anesthesiology. 2016 Aug 8. pmid:27496657.
- 30. Kussman BD, Aasted CM, Yucel MA, Steele SC, Alexander ME, Boas DA, et al. Capturing Pain in the Cortex during General Anesthesia: Near Infrared Spectroscopy Measures in Patients Undergoing Catheter Ablation of Arrhythmias. PloS one. 2016;11(7):e0158975. pmid:27415436. Pubmed Central PMCID: 4944937.
- 31. Moayedi M, Salomons TV, Dunlop KA, Downar J, Davis KD. Connectivity-based parcellation of the human frontal polar cortex. Brain structure & function. 2015 Sep;220(5):2603–16. pmid:24927720. Pubmed Central PMCID: 4549383.
- 32. Simons LE, Moulton EA, Linnman C, Carpino E, Becerra L, Borsook D. The human amygdala and pain: evidence from neuroimaging. Human brain mapping. 2014 Feb;35(2):527–38. pmid:23097300. Pubmed Central PMCID: 3920543.