We obtained Ethical approval from the Health & Disability Ethics Committee, New Zealand (Reference 2023 EXP 19190). Research consultation with Māori was obtained from the Ngāi Tahu Research Consultation Committee (Reference 24402). The trial has been prospectively registered in the Australian New Zealand Clinical Trials Registry (ACTRN12625000706471).

All eligible participants will receive an information sheet with consent form outlining the study design and purpose. Participants who are enrolled will complete a digital consent form on the day before their baseline assessment and provide a signed paper copy upon attending in person. Participants may opt-out at any point during the study. All data for primary and secondary outcomes will be deidentified before analysis. In recognition of their time and contribution to the research, participants will receive NZ $200 (US $115) supermarket vouchers upon completion of all study sessions.

A single-arm, open-label, pilot-feasibility clinical trial was designed according to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT; Multimedia Appendix 1) and the Consolidated Standards of Reporting Trials (CONSORT; Multimedia Appendix 2) extension for pilot and feasibility trials. The trial will be conducted in Dunedin, New Zealand, and the study phases are summarized in Figure 1 [56,57]. This study was designed as an open-label trial to assess the feasibility, acceptability, safety, and preliminary data of the developed EEG-NF training. Given the pilot-feasibility design and exploratory nature of the study, inclusion of a control group was not necessary at this stage. The primary focus is to gather preliminary data on EEG-NF training implementation and variability in clinical outcomes, which will inform a future fully powered clinical trial [58-60]. A structured description of the intervention is summarized in Multimedia Appendix 3, following the Template for Intervention Description and Replication (TiDiR) guide (Multimedia Appendix 3) [61].

Convenience sampling will be used to recruit participants from the wider Dunedin community. Periodic advertising in community newspapers and social media will be carried out. Posters will be placed throughout the local community to advertise the study for participant recruitment.

Interested participants will undergo eligibility screening by telephone or through a secure online survey using Research Electronic Data Capture (REDCap; Vanderbilt University) web application [62]. As this is a pilot-feasibility study, a formal power-based sample size calculation was not undertaken. Instead, the sample size was determined as 6-12 participants over a 4-month recruitment period, in accordance with recommendations for feasibility studies and previous literature, ensuring adequacy to evaluate safety and feasibility outcomes while also permitting exploratory analyses [63].

Participants aged 35-75 years with MSK pain, averaging a score of ≥4 on the 11-point Numeric Pain Rating Scale over the past 3 months, with a minimum pain duration of 3 months, and a score of at least 19 on the PainDETECT questionnaire will be eligible to participate in this study [6,14,64-67].

Participants will be excluded if they have one of the following conditions or situations: systemic rheumatic conditions, neurological conditions, major psychiatric illness, cervical or lumbar radiculopathy, peripheral entrapment neuropathies, cognitive impairment, have undergone any surgery or intra-articular steroid injections in the 3 months before the study, have undergone any intra-articular hyaluronic acid injections in the 6 months before the study, are scheduled to undergo any surgery within 6 months of enrollment in the study, have undergone previous neurological procedures of the brain, are pregnant or up to 6 months postpartum, cardiovascular disease, peripheral arterial disease of limbs, uncontrolled hypertension (≥150/95 mm Hg), or uncontrolled diabetes.

Interested participants will undergo a brief preliminary screening for initial eligibility by telephone or through an online questionnaire using the REDCap online survey platform [68]. If the initial eligibility criteria are satisfied, participants will undergo a detailed screening based on their health information through a secure online survey. Participants will then be given an appointment for the final confirmatory screening.

A paper-based Montreal Cognitive Assessment will be carried out to screen for cognitive impairments [69]. The maximum Montreal Cognitive Assessment score is 30 points, and only those who score 21 or above will be eligible to participate in this study, based on normative data for the population included in the study [69,70]. If the participant meets this threshold and is found to be eligible, baseline assessment will continue.

Eligible participants will undergo 12 sessions (30 minutes each; 3 sessions per week) of ISF EEG-NF training and two 90-minute baseline and postintervention assessments at the Department of Anatomy–Research Clinic facility at the University of Otago, New Zealand. The intervention will be provided by a researcher experienced in the delivery of neurofeedback. Participants will be required to abstain from alcohol and caffeinated beverages for 24 hours and from food and drinks for at least 1 hour before any assessment sessions [71].

At the start of each session, participants will be seated in a chair with back support and asked to remain relaxed for 10 minutes while the trainer prepares them for neurofeedback training. The ISF EEG-NF training will be administered using a 21-channel DC-coupled amplifier (BrainMaster Technologies Inc) with a Comby EEG lead cap containing Ag/AgCl sensors of an appropriate size, secured to the participant’s head. Reference electrodes will be placed at the mastoids [72]. The impedance of the active electrodes will be continuously monitored to remain below 5 kΩ [49,52,72]. Participants will be advised to minimize eyeball, head, and neck movements, swallowing, and clenching of teeth to reduce motion artifacts in the EEG data [73]. Treatment adherence will be ensured by directly observing participants throughout the training sessions.

Participants will be instructed to relax, keep their eyes closed, and listen to the audio feedback. For this study, an ISF EEG-NF program to downtrain the RIns and dACC was developed using the BrainAvatar Live standardized low-resolution electromagnetic brain tomography projector software [52,74]. Standardized low-resolution electromagnetic brain tomography allows the selection of brain regions (region of interest [ROI]) for EEG-NF training based on the current density of the ROI and 3D brain mapping functionality [74-76]. A visual representation of the ROIs, direction of training, and an overview of outcome measures are provided in Figure 2. The figure was generated using BrainNet Viewer [77] and BioRender [78].

The BrainMaster Technologies software (BrainAvatar) will deliver real-time auditory feedback when participants’ brain activity is simultaneously downregulated within the ROIs—the dACC and RIns—in the ISF (0.0-0.1 Hz) band. The BrainAvatar auditory feedback is generated using Musical Instrument Digital Interface–based tones and sound effects. The auditory feedback serves as a primary reinforcement modality, providing immediate and intuitive cues that reflect real-time brain activity. The system incorporates a range of sounds, including simple tones, chimes, and naturalistic effects such as chirps or gongs, which vary in pitch, clarity, or occurrence depending on whether the individual’s brainwave activity meets the targeted thresholds. For example, reward tones or sound effects are triggered only when the desired neural patterns are achieved, whereas the absence, distortion, or reduction of sound indicates deviation from the training goals. This dynamic modulation allows the brain to receive continuous, nonverbal information about its own functioning, supporting operant conditioning and the self-regulation processes central to effective neurofeedback training [47,48].

The auditory feedback (reward) will be captured using Audacity, a free and open-source audio recording and editing software [79]. Audacity can use the computer’s microphone and sound card as its own audio-to-digital converter, eliminating the need for additional equipment. The auditory feedback will then be used to calculate the duration of successful neurofeedback training, defined as the cumulative time that the participant received the auditory reward feedback during the 30-minute neurofeedback session [80]. The duration of successful neurofeedback training from each participant across the 12 sessions will be included in the analysis.

The target ROIs were chosen based on previous literature related to EEG-NF and MSK pain conditions [30,31,50,52]. The reward threshold will be continuously monitored by the trainer and adjusted manually if cortical activity falls outside the desired threshold [49,52,72]. The duration and number of neurofeedback sessions were informed by recent studies using ISF EEG-NF for chronic pain management, which demonstrated changes in clinical pain outcomes, and were selected to ensure methodological rigor in this study [72,81].

Once written informed consent is obtained, participants will complete questionnaires that include demographics and general health information, such as age, sex, and ethnicity. Assessment of resting-state EEG and clinical and experimental pain outcomes will be conducted by a researcher who is adequately trained in all procedures.

The primary outcomes of this study will be the feasibility and safety measures [82]. These will be collected by the investigator throughout the study period, from recruitment until the last postintervention assessment session. Feasibility outcomes include the recruitment rate, defined as the number of participants attending a screening assessment each month; the enrollment rate, defined as the proportion of participants recruited from the total screened; the compliance rate, defined as the number of sessions attended by each participant out of the total sessions; and the dropout rate, defined as the number of participants who dropped out over course of the study [83,84]. All measures will be monitored through monthly audits of recruitment records during the first 4 months of the recruitment and intervention period.

An adverse event is defined as any harmful event or symptom from the trial that could reasonably be linked to the procedure or EEG-NF training. EEG-NF is a safe technique; however, all participants will be asked about any adverse events experienced from the previous session at each visit, and any worsened side effects compared with previous training will be recorded. All participants will be instructed to complete the Discontinuation Emergent Sign and Symptom scale [85], a 43-item checklist consisting of emotional, behavioral, cognitive, and physical symptoms. These symptoms can be considered possible side effects from neurofeedback training and have been used in previous literature as a measure of safety monitoring for similar studies [52,72].

Secondary outcomes of the study include self-reported questionnaires, resting-state EEG, electrocardiogram (ECG), and quantitative sensory testing (QST), assessed at baseline (T1) and postintervention (T2). Table 1 briefly describes each of the measurement tools used and the time points throughout the study. All tools are reliable and validated for use in individuals with persistent pain and are consistent with previous literature and the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) recommendations for clinical trials in chronic pain [86-90].

Resting-state EEG will be recorded (sampled at 256 Hz) using a 21-channel system (Fp1, Fp2, F7, F3, Fz, F4, F8, T7, C3, Cz, C4, T8, P7, P3, Pz, P4, P8, O1, O2, A1, A2; BrainMaster Technologies, Inc) [43]. Raw EEG recordings will be collected for 10 minutes with participants’ eyes closed. To minimize potential artifacts, participants will be instructed to avoid facial movements, head and neck movements, and swallowing. Alertness will be monitored by observing the presence of spindles and slowing of the alpha rhythm in the EEG stream to prevent drowsiness-related increases in theta power during the recording [116].

Each EEG file will be resampled to 128 Hz and bandpass filtered for 0.01-44 Hz in EEGLAB (MATLAB R2020a) [117]. Artifacts such as eye blinks, muscle artifacts, perspiration, and body movements will be filtered out using ICoN software [52,118].

Exact low-resolution brain electromagnetic tomography (eLORETA) software will be used to extract the following EEG variables. First, ROI current density, for which frequency-specific data will be collected by performing a voxel-by-voxel analysis containing 6239 voxels for each of the following frequency bands at the dACC and RIns: infraslow (0.01-0.10 Hz), slow (0.2-1.5 Hz), delta (2-3.5 Hz), theta (4-7.5 Hz), alpha (8-12 Hz), beta (12.5-30 Hz), and gamma (30.5-44 Hz). Comparisons will be made between participants’ pretreatment and posttreatment measures using eLORETA statistical contrast maps. Multiple voxel-by-voxel comparisons in the logarithm of t-ratio with a threshold of P≤.05 will be used to compute the cortical 3D distribution of current density [49].

Second, functional connectivity will be assessed using lagged linear connectivity, which is a statistical measure of coherence representing the phase synchronization between ROIs within the brain, as it combines the relationship between the phase and amplitude of the signals [23,24]. This measure reflects the level of communication between the ROIs [90]. Log-transformed lagged linear connectivity will be derived for all possible connections across the 7 frequency bands between the dACC and RIns.

eLORETA is an EEG source localization software that computes current density across the full brain volume [119] to localize cortical brain structures with fewer localization errors for EEG data extraction and analysis [120]. eLORETA has been widely used in previous literature for EEG analysis [121-123].

Heart rate variability (HRV) metrics will be obtained using a Polar V800 heart rate monitor and a Polar H10 Chest Pro Strap [124]. The ECG monitor will be placed below the sternum and secured to ensure a fixed position for the participant. Raw heart rate and R-R (peak-to-peak R wave) interval time series data will be downloaded from the Polar Flow software for further processing and analysis. Kubios HRV Premium (v3.2.0) software will be used to analyze HRV from the R-R interval record exported from the Polar Flow software [125]. The root-mean-square of successive differences between normal heartbeats and stress index (SI) will be extracted for HRV analysis. Root-mean-square of successive differences quantifies the variability between adjacent respiratory rate intervals, demonstrating a lesser susceptibility to influence by the respiratory system and, due to this, is regarded as the most dependable HRV for monitoring vagal response [125,126]. The SI will be calculated using the square root of the Baevsky SI model [127]. Given that the insular cortex and dACC are key regions involved in interoception—integrating internal bodily signals to regulate physiological and emotional states—any changes in sympathetic and parasympathetic activity can be effectively captured through HRV analysis [80,128,129].

The following QST measurements will be performed at T1 and T2. All QST measures were included based on previous literature demonstrating variability in these clinical measures among individuals with chronic MSK pain and NP [130-132].

The primary aim of this study is to pilot test and assess the feasibility of the novel ISF EEG-NF training protocol targeting the dACC and RIns for individuals with NP-like qualities. This study, for the first time, aims to simultaneously downregulate the ISF activity of the RIns and dACC in a chronic pain population. Therefore, a preliminary study assessing the feasibility, acceptability, and safety of the intervention is imperative before advancing to a fully powered clinical trial to evaluate its efficacy.

Previous studies, using comparable protocols have conducted dual region source-localized ISF EEG-NF training in individuals with chronic low back pain and major depressive disorders [50,147]. These studies successfully targeted the dACC, showing clinical improvements and variability following the training. In addition, a recent study using a similar methodology successfully targeted the insular cortex, specifically training the left posterior insula in healthy volunteers. This study explored the relationship between the duration of successful neurofeedback training and clinical outcomes, aiming to understand the impact of EEG-NF training on gastric slow-wave activity [80]. Thus, this study adopts a similar methodological framework to develop the ISF EEG-NF program [75,120], using established and reliable methods that have been successfully implemented in previous research [43,52,72,80,89]. The ROIs (RIns and dACC) are created with all the nearby voxels with a localization of 5 mm [41,52,120]. Moreover, no studies have attempted ISF EEG-NF to manage NP-like qualities in individuals with chronic MSK pain, highlighting the novelty of this project and its potential for future clinical translation.

This study will also introduce a novel approach to measuring EEG learning effects by assessing the duration of successful neurofeedback within each session, a method previously used in only one study [80]. The secondary aims will allow for exploratory analysis to better understand data variability and trends in EEG measures and clinical pain outcomes in individuals with NP-like qualities in chronic MSK conditions.

During this study, several challenges may be encountered while assessing feasibility. One of the primary issues may be due to recruitment of participants. This could be attributed to the challenges faced by patients coming to the study location to receive treatment, as their pain may be debilitating, as highlighted in previous literature emphasizing activity limitations [148]. This study is also targeting a demographic that may be spread throughout the wider local community and thus may not encounter the flyers placed throughout the community. We plan to mitigate this by also including the study advertisement in newspapers, which may be more accessible for recruiting patients who are further away from the surrounding area. Since PainDETECT is being used as a screening tool, we are excluding patients with scores below 19, including those falling within the range of 14-18 that indicates a possible neuropathic component. This produces a challenge as it reduces recruitment and potentially excludes patients who may have a neuropathic component to their pain. However, only recruiting patients who score 19 or above strengthens the validity of the patients having a neuropathic component, as painDETECT has a high sensitivity and specificity [149]. That said, we acknowledge that this can impact patient recruitment and enrollment rates.

The data and findings of this study will aid in designing future fully powered clinical trials to further evaluate the efficacy of this novel ISF EEG-NF training for NP management. Additionally, the foundational data from this pilot open-label feasibility trial will provide the necessary information to enable future studies to perform power calculations for designing upcoming trials.