Reducing Pain in Complex Regional Pain Syndrome Using Personalized Brain Stimulation: A Feasibility Study

Complex regional pain syndrome (CRPS) is a chronic pain condition that is debilitating and dramatically decreases the quality of life. CRPS is characterized by a constellation of sensory, motor, and autonomic dysfunction. CRPS patients present with allodynia and hyperalgesia. There are also trophic changes of the affected limb, changes in skin colour and temperature, alterations in motor control, the presence of tremors, dystonia and edema, and feelings of hate towards the CRPS affected limb. Cortical representations of the affected limb in the somatosensory cortex are smaller and possess greater overlap with adjacent areas. This is speculated to be a result of a decrease in afferent input originating from the periphery. Additionally, this reorganization is reinforced by the minimal usage of the affected limb by patients in order to avoid pain. The worldwide incidence of CRPS is 26.2 per 100,000 people per year and is 3 to 4 times more prevalent in females. Based on these reports, it is estimated that Canada will experience ~10,000 new cases of CRPS each year. CRPS creates an economic burden; healthcare costs are increased by ~20% and two-thirds of individuals diagnosed with CRPS will not return to work. Taken together, these findings demonstrate the existing impact and increasing burden of CRPS on the Canadian workforce, economy, and healthcare. Pain relief is an unmet need in patients with CRPS. Options for medical management of CRPS include medications such as opioids, anticonvulsants, antidepressants, sympathetic blockade and invasive spinal cord stimulation (SCS). However, evidence for these treatments come from medium to low quality randomized control trials (RCTs). SCS, though effective for many patients with CRPS, is associated with significant complications and costs. At present, CRPS is managed through continued physiotherapy with supportive treatment for pain and mental health . However, this method does not effectively alleviate pain in CRPS.

Repetitive transcranial magnetic stimulation (rTMS) applied to the motor cortex has been shown to effectively alleviate pain in CRPS along with several other pain conditions. However, rTMS does not effectively alleviate pain for all individuals and there exists a large degree of interindividual variability in the efficacy of rTMS as a treatment for pain. This interindividual variability is thought to be a result of individual differences in patterns of brain activity. Indeed, in depression, tailoring rTMS to individual differences in depression-related brain activity recorded non-invasively via electroencephalography (EEG) resulted in more consistent and enhanced improvements in depression symptoms. Specifically, altering the frequency and location of rTMS according to alpha, beta, and theta power enhanced improvements in depression symptoms. In CRPS, one promising EEG marker is the movement-related cortical potential (MRCP) and is generated by real or imagined movement. Nerve stimulation delivered at the peak-negativity of the MRCP generated by attempted dorsiflexions in stroke patients resulted in improvements in a 10-meter walking task but not in a finger tapping task. Further, these patients had enhanced excitability of the cortical representation of the stimulated muscle. This is thought to have occurred because nerve stimulation caused cortical activity via afferent feedback within a couple hundred milliseconds of movement intention. Additionally, the peak-negativity of the MRCP is closely timed to alpha desynchronization, an EEG marker of cortical excitability.

MRCPs occur in paraplegic individuals, individuals suffering from stroke, individuals with amyotrophic lateral sclerosis, and Parkinson's disease. Given that MRCPs persist in a variety of conditions whereby movement is limited or even abolished, it is probable that MRCPs are quantifiable in CRPS. Using the MRCP in CRPS, one can theoretically reinforce connections used in sensorimotor control of the affected limb and the sensorimotor cortex. In addition to pain, CRPS is associated with structural and functional changes of the somatosensory, primary motor, and supplementary motor areas. MRCPs which are obtained from the sensorimotor cortices will provide a window into the movement planning processes that can be enhanced through MRCP tailored brain stimulation. Further, even if an individual cannot move, imagined movements that are furnished by the same sensorimotor cortices can be used to tailor brain stimulation.

To the best of the investigator's knowledge, MRCPs have not been used to trigger TMS by a real-time brain-computer interface (BCI) in healthy controls or in patient populations. The investigator's lab has developed a BCI that accurately detects MRCPs in healthy controls (Appendix Figure 1) with its best performance being an accuracy of 92%, true-positive-rate of 80% and a false-positive-rate of 7%. These results are comparable to other BCIs that detect MRCPs. Additionally, the investigator's lab has developed and validated a closed-loop EEG-triggered TMS framework that is compatible with the investigator's BCI (Appendix Figure 2). Ultimately, it is feasible for the investigator's lab to accurately tailor rTMS timing to MRCPs in healthy controls. However, the feasibility of doing so in a pain population and specifically in CRPS has yet to be determined. To the best of the investigator's knowledge MRCP morphology has not been studied in CRPS. However, MRCPs have been characterized in paraplegic individuals, individuals suffering from stroke, individuals with amyotrophic lateral sclerosis, and Parkinson's disease. Therefore, the feasibility of real-time MRCP tailored rTMS needs to be established in CRPS. Further, the effect of this approach on subjective and neurophysiological measures of pain have yet to be determined. Understanding the short-term effects of this novel approach would inform a future real-time MRCP tailored rTMS intervention for CRPS. The goal of the research is to determine the feasibility and effects of a novel BCI approach to delivering rTMS in individuals suffering from complex regional pain syndrome.

The goals of this study are to:

1. Determine the feasibility of triggering rTMS at MRCPs in individuals with CRPS. This aim will be assessed by measuring the accuracy of the BCI.
2. Determine the effects of real-time MRCP tailored rTMS on subjective pain in individuals with CRPS. This aim will be assessed using a numeric pain rating scale.
3. Determine the effects of real-time MRCP tailored rTMS on neurophysiological markers of pain. This aim will be assessed by EEG biomarkers of chronic pain.

Participants: 12 eligible participants will include males and females aged 20 through 80 years with a diagnosis of CRPS Type 1 affecting the upper or lower limb using the revised Budapest criteria. The investigators will recruit patients with a history of CRPS of at least 3 months since the start of symptoms or diagnosis. The investigators will exclude patients if they have had CRPS diagnosis greater than 3 years to avoid patients with chronic CRPS who may not respond to interventions. According to published literature, chronic CRPS patients (> 3 years) may be a subtype with specific phenotypic traits, including possible non-response to therapy.

Study Design: This research will be conducted in Dr. Nelson's Human Neurophysiology and Neuroimaging lab, Ivor Wynne Centre (IWC), room IWC AB-131, at McMaster University. This lab is equipped with all the needed infrastructure for this study. At the beginning of a session, participants will first be asked to fill out the paper copy of the TMS screening questionnaire and handedness questionnaire. Participants will also be provided with a paper copy of the consent form and will be given ample time to read and sign the form. Participants will be blinded for the duration of the study, unaware as to which stimulation is applied. The investigator team will be unblinded while collecting data.

This experiment will include 8 phases during a single session of ~3 hours (Figure 1). Pre-intervention pain rating measures will be acquired using the numeric rating scale (NRS). Then EEG will be prepared (~45 minutes), and 5 minutes of resting-state EEG will be measured. The motor hot spot and resting motor threshold (RMT) for TMS will be determined. Next, the BCI will be trained on 30 trials of the investigator's task (Figure 2). Following BCI training, the real-time MRCP tailored rTMS intervention will begin (~15 minutes). Then, 5 minutes of resting-state EEG will be measured again, and pain ratings via the NRS will be measured again.

Outcomes and Data Analysis:

Aim 1 (accuracy of BCI): Offline analysis of the EEG data recorded during intervention will be used to determine the true positive (TPR), false positive (FPR), true negative (TNR), and false negative rates (FNR) of the model. This aim will be achieved if 70% or more of the participants' trained BCI achieve a TPR > 65% and an FPR < 10%.

Aim 2 (pain ratings): Pain ratings will be measured before and after intervention in all 12 participants using the NRS. Ratings will be compared before and after intervention.

Aim 3 (neurophysiological markers of pain): Resting state EEG before and after the intervention will be analyzed offline and compared with the outcome from Aim 2 to provide a more comprehensive evaluation of pain severity and explore neurobiological changes associated with the intervention. Changes in EEG are observed across various chronic pain conditions, including elevated theta band power at rest and increased levels of event-related desynchronization (ERD). Importantly, the magnitude of these changes is associated with pain severity and can be used as objective markers of pain following an intervention. In this study, EEG markers of pain, including peak-alpha/theta frequency, alpha-gamma phase amplitude coupling, and power spectral density will be computed.

Significance: This study uses a cutting-edge approach through the most advanced technology in the field of neuromodulation to optimize rTMS as a therapeutic management option for CRPS. The proposed research will produce the preliminary data necessary for a larger study that is likely to make a significant impact in the treatment and management of CRPS in Canada and worldwide. The proposed research is an essential precursor to build a solid foundation for tailored rTMS in CRPS. The present research paves the way for this advancement in a population that is in desperate need of alternative pain-relief solutions. In the future, the methodology herein can be tested in and adapted to other chronic pain conditions.