Short‐ and long‐term reproducibility of body surface gastric mapping using the Gastric Alimetry® system

Many diagnostic tests for gastroduodenal symptoms, such as gastric emptying scintigraphy (GES), gastric emptying breath tests (GEBT), and electrogastrography (EGG) show variable intra‐individual reproducibility over time. This study investigated the short‐ and long‐term reproducibility of body surface gastric mapping (BSGM), a non‐invasive test for assessing gastric function, in controls and patients with chronic gastroduodenal disorders.


| INTRODUC TI ON
Chronic gastroduodenal symptoms are highly prevalent, affecting at least 20% of the global population. 1,2These symptoms encompass nausea, vomiting, early satiation, epigastric pain and burning, and excessive postprandial fullness, which are often clinically recognized as functional dyspepsia (FD), chronic nausea and vomiting syndrome (CNVS), and gastroparesis.These conditions impart a high burden on patients and the healthcare system 3,4 and, due to their heterogeneous presentations and overlapping symptomology, are difficult to diagnose and manage. 5This highlights the need for better diagnostics which can reliably and objectively identify specific patient subgroups to inform personalized care.
Reproducibility was often dependent on the measures assessed, with many EGG metrics showing unsatisfactory reproducibility. 8,19,21 addition, diagnoses based on GES tests show low reproducibility, with 30%-42% discordance in patient diagnoses between tests repeated 2 24 and 48 weeks apart, 5 which were unrelated to changes in symptomatology.
7][28] Gastric Alimetry enables the high-resolution recording of gastric myoelectrical activity through a non-invasive high-density electrode array of 64 electrodes.This test offers several novel biomarkers of gastric function capable of differentiating patients with overlapping gastroduodenal symptom profiles. 289][30][31] It has shown substantially superior diagnostic profiling capabilities compared to EGG 32 and substantially expands patient phenotyping when used with GES, 33 as well as the ability to detect gastric motility changes as a result of auricular vagus neuromodulation. 34However, the intra-individual reproducibility of the results from a Gastric Alimetry test is still unknown.
The lack of consistent evidence for the reproducibility of GEBT, GES, and EGG tests highlights the importance of testing the reproducibility of Gastric Alimetry to ensure it does not have the same pitfalls as these other diagnostic motility tests.The establishment of both short-and long-term reproducibility of Gastric Alimetry is critical to ensure that test-related variability does not significantly affect the results that an individual receives and that test outputs remain relatively stable over time, assuming no change in the underlying disease status.Therefore, this study aimed to investigate the reproducibility and inter/intra-individual variability of Gastric Alimetry BSGM metrics in patients and healthy controls, over the short term (1 week) and long term (>6 months).

| ME THODS
As shown in Figure 1, healthy controls and patients with chronic gastroduodenal symptoms were recruited for reproducibility profiling 6+ months after first completing a standard Gastric Alimetry BSGM test, as part of a multi-national consortium database study, on behalf of the BSGM Consortium (https:// www.bsmco nsort ium.com/ ).Patients were defined as meeting the Rome IV Criteria for FD or CNVS, while healthy controls were defined as those without chronic gastrointestinal symptoms, including not meeting the Rome IV criteria. 35Standard exclusion criteria for a Gastric Alimetry test were applied (breastfeeding, pregnancy, previous major gastric surgery, adhesive allergies, or damaged epigastric skin). 36Additionally, participants were only included if their first test showed good technical quality, as outlined previously (e.g., <50% artifact detected, >50% of the meal consumed). 36,37rticipants were excluded if they had significant physical or medication changes since their original test (e.g., weight loss/gain of more than 10 kg, change in medications that affect the stomach, or significant medical procedures or surgeries).Recruitment was performed consecutively, based on meeting eligibility criteria and participant availability.

| Procedure
Patients and controls were invited a minimum of 6 months after their original test to complete two more identical tests, conducted approximately 1 week apart.The three Gastric Alimetry tests were conducted using standardized methodology according to a recent consensus technical review. 26,28,30,36All tests were conducted in the morning after an overnight fast of at least 8 h.Per standard test methodology, medications known to affect gastrointestinal motility were withheld for 48 h prior to each test, and caffeine and nicotine were avoided on the day of the test. 30The Gastric Alimetry system employs a stretchable array of 8 × 8 electrodes on an adhesive patch, coupled with a wearable reader, which is placed over the epigastrium.The abdominal skin was shaved and prepared using NuPrep (NuPrep, Weaver & Co, CO, USA) prior to array placement, with impedance checks before commencing recordings.
As per standardized procedures, 36 data were recorded over a 30-min fasting baseline, followed by a 482 kCal test meal, consumed over 10 min, and a 4-h continuous postprandial recording.The standard meal consisted of an oatmeal energy bar (250 kcal, 5 g fat, 45 g carbohydrate, 10 g protein, 7 g fiber; Clif Bar & Company, CA, USA) and Ensure (232 kcal, 250 mL; Abbott Nutrition, IL, USA), or an appropriate calorie-matched diabetic or gluten-free meal substitute.The researchers ensured that the morning start time of the three tests was as close as possible, to control for potential variation induced by an individual's typical meal schedule and circadian variations.Participants were asked to sit in a reclined position and limit their movement during the test.
Motion artifacts were automatically corrected or rejected using validated algorithms. 26,38

| Gastric Alimetry metrics
Gastric Alimetry provides a single spectrogram for each test, summated from the top-ranked channels of the electrode array, which describes the frequency and amplitude content of the recorded signals as a function of time. 26These spectrograms were used to compute four spectral metrics for analyses, described in detail elsewhere 36,37,39  • BMI-Adjusted Amplitude (μV): characterizes the strength of the recorded gastric activity as an average of the whole-test amplitude, adjusted for BMI.• Gastric Alimetry Rhythm Index™ (GA-RI): a measure of the concentration of amplitude in the gastric frequency band over time (between 0 and 1), which characterizes the rhythmic stability of the recorded gastric activity; also adjusted for BMI.
• Principal Gastric Frequency (cpm): characterizes the frequency associated with stable, persistent gastric activity as defined by GA-RI.
• Fed: Fasted Amplitude Ratio (ff-AR): characterizes the responsiveness to a meal stimulus, which is calculated as a ratio of the maximal 1-h average postprandial amplitude to the amplitude in the preprandial period.
Principal Gastric Frequency, BMI-Adjusted Amplitude, and GA-RI are considered the primary BSGM metrics as they can be used to classify patients into phenotypes based on normative metric reference intervals. 36,40,41

| Gastrointestinal symptoms
Participants logged their symptoms throughout each test using the validated Gastric Alimetry App. 27Nausea, bloating, upper gut pain, heartburn, stomach burn, and excessive fullness were measured at minimum 15-min intervals on a scale from 0 (none) to 10 (most severe imaginable).Early satiation was measured once, at the end of the meal, on the same scale.An average of each individual symptom during the test was calculated, along with a total symptom burden score, which was calculated by taking the sum of the postprandial averages of each continuous symptom plus the reported earlysatiation score. 27Symptom events (reflux, belching, vomiting) are also recorded in the app.

| Secondary measures
Secondary measures of long-term symptoms, quality of life, and depression were measured using validated questionnaires.

| Statistical analysis
Data were analyzed using IBM SPSS Statistics v29.All analyses were performed separately for patients and controls to allow for the assessment of reproducibility in both groups independently.
Statistical significance was set to p < 0.05.Independent sample t-tests were conducted to assess the differences in the Gastric Alimetry metrics between controls and patients at each of the three tests.Given the bimodal distribution of Principal Gastric Frequency, these t-tests were conducted on Principal Gastric Frequency deviation (the absolute difference of Principal Gastric Frequency from 3 cpm, which is the midpoint of the reference interval 36 ).
One-way repeated measures ANOVAs with Greenhouse Geisser adjustments were conducted to define the differences in the average metrics and secondary outcomes between the three tests.
Reproducibility of the Gastric Alimetry metrics was assessed for both the long term (between Test 1 and Test 2) and short term (between Test 2 and Test 3).Concordance between the metrics was assessed using Lin's concordance correlation coefficient (CCC) 46 and Bland-Altman statistics. 47The inter-individual coefficients of variation (CV inter ) were calculated as: where standard deviation (SD) and mean (M) are computed from the data from the two different tests. 16The intra-individual coefficient of variation (CV intra ) for each participant was calculated as: where SD d is the SD of the differences between the two tests and M is the mean of the two tests. 16Both CV Inter and CV intra are expressed as percentages.
The ranges of the CCC and CV intra for all metrics at both short and long terms were visually compared against the ranges from previous studies of GEBT, GES, and EGG which used the same statistical tests.
The participants were independently categorized into phenotypes for each of the three tests based on the normative intervals of the Gastric Alimetry metrics. 36,37Gastric dysfunction was categorized as low rhythm stability (GA-RI <0.25), low amplitude (BMI-Adjusted Amplitude <22 μV), high amplitude (>70 μV), high frequency (>3.35 cpm), and low frequency (<2.65 cpm). 36anges in phenotypes and Rome categories between the tests at both short and long terms were reported as the percentage of patients and controls who were categorized differently between the tests.Unweighted kappas (κ) were also calculated to formally assess the reproducibility of the phenotypes and Rome diagnoses at both short and long terms.
There were minor variations in meal start times (median difference = 10 min, Range = 0-86 min) due to participant arrival time and test set-up time.Due to participant requests, three participants switched between standard, gluten-free, and diabetic meals.
These meals were calorie-matched, with only minor differences in nutritional content.As shown in Figure 1, the Gastric Alimetry metrics for one of the patients at Test 3 were removed from the analyses due to a technical problem with a device.Another patient did not complete the third test due to an unrelated medical complication.

| Differences between patients and controls
There were significant differences in the GA-RI and BMI-Adjusted Amplitude between healthy controls and patients at all three tests, with patients showing significantly lower GA-RI and BMI-Adjusted Amplitude than controls (all p < 0.05).ff-AR showed similar significant differences between healthy controls and patients at tests 1 and 3 (p < 0.05) and a trending difference at test 2 (p = 0.083), with patients showing lower ff-AR than controls.In contrast, there were no significant differences in Principal Gastric Frequency deviation between healthy controls and patients for any of the three tests (all p > 0.05).

| Metric reproducibility
Averaged summary spectrograms and symptom profiles from the Gastric Alimetry test, for both patients and controls, are shown in Figure 2, demonstrating qualitatively high similarity across the three tests.Controls showed a stable 3 cpm dominant band of Principal Gastric Frequency across all three tests, while patients showed a consistently greater degree of spectral instability, indicating the presence of dysrhythmic activity occurring in some patients within the cohort.Both patients and controls showed a similar pronounced meal response curve on averaged data across all three tests.Total symptom burden was higher for patients than controls, across the duration of the test.Further details on how to interpret the spectrograms and symptom profiles can be found in Foong et al. 37 and Varghese et al. 41 Repeated measures ANOVAs demonstrated no significant differences in the mean Gastric Alimetry metrics between the three tests for both patients and controls (see Figure 3), with individual participant results shown in Figure 4; Figure S1.The lack of significant differences in the mean Gastric Alimetry metrics remained when controlling for sex (all p > 0.276), age (all p > 0.234), and BMI (all p > 0.106).As shown in Figure 4, half of the patients fell within the normal intervals for all of the BSGM metrics, while the other half showed a spectral abnormality for at least one of the metrics measured.
F I G U R E 2 Averaged spectrograms, median (IQR shaded) BMI-Adjusted Amplitude curves, and median (IQR shaded) symptom burden curves for patients and controls for each of the three repeated tests.
The average and range of change scores for each Gastric Alimetry metric for both short and long terms are shown in Table 1.The Principal Gastric Frequency, BMI-Adjusted Amplitude and GA-RI metrics showed Lin's CCCs ranging from 0.58 to 0.96 indicating good concordance between the tests at both short and long terms in both patients and controls, especially for Principal Gastric Frequency (see Table 2).However, ff-AR in patients at both the short and long term demonstrated poor concordance, indicating lower reproducibility.The CV intras for all metrics were under 3%, indicating markedly low variation between the three tests within each individual, despite wide variation in the CV inters (Table 2).Similar results were found even when the patient group was further separated by those with FD vs. CNVS diagnosis and epigastric pain syndrome and postprandial distress symptom diagnosis based on the Rome IV criteria.The metrics generally showed lower reproducibility pre-meal and during the first hour postprandially, with greater reproducibility observed from the second to fourth postprandial hours.Full results are shown in Table S2.

| Comparisons to other gastric motility tests
Figure 5 compares the ranges of CCC and CV intra for Principal Gastric Frequency, BMI-Adjusted Amplitude, and GA-RI at both the short and long term against the ranges from previous studies.The primary Gastric Alimetry metrics showed a higher CCC upper limit than the one other study that has measured CCC in GES 24 (Figure 5A).The primary Gastric Alimetry metrics also showed a considerably lower range of CV intra than studies of GES, 12,13,16,24 GEBT, 9 EGG, 8 and SmartPill gastric emptying time, and small bowel, colonic, and whole gut transit time 22 (Figure 5B).
All of the patients and controls that changed phenotypes between the tests were considered borderline cases (i.e., their values on the metrics for the test that had a different phenotype were only just above/below the normative reference intervals used for phenotyping).For example, two controls and one patient were categorized as having a normal Principal Gastric Frequency for two of the three tests

| DISCUSS ION
This study is the first to evaluate the reproducibility of BSGM using the Gastric Alimetry system.The results demonstrated that the Gastric Alimetry BSGM metrics show high reproducibility and low intra-individual variation at both short (1 week) and long term  (>6 months) in healthy controls and patients with chronic upper gastrointestinal symptoms.These data establish that the results from a Gastric Alimetry test are unlikely to be influenced by test-related variability and remain consistent over time.
The three primary Gastric Alimetry metrics also showed superior reproducibility and substantially lower intra-individual variation when compared to other gastric motility tests, comprising GEBT, GES, EGG, and SmartPill. 8,12,13,16,22,24However, only one previous study of GES has measured Lin's CCC, 24 with comparisons again showing more desirable CCC ranges for the Gastric Alimetry metrics.In particular, Principal Gastric Frequency showed excellent reproducibility, while BMI-Adjusted Amplitude and GA-RI showed good reproducibility.In contrast, ff-AR showed good reproducibility in controls, but poor reproducibility in patients, despite the low intra-individual variability.This finding was not unexpected, as ff-AR has a wide reference interval, reflecting variability of the baseline gastric amplitude when fasting, which is thought to be due to the occurrence of migrating motor complexes and circadian variations. 36This finding therefore reinforces the current view that the ff-AR metric should not be considered a reliable indicator of gastric dysfunction in isolation, and should only be used as a supporting metric for an abnormal test in combination with other metrics. 37gacy EGG research has shown that reproducibility is variable according to the metrics studied. 19,21For example, the frequency of slow waves and slow wave coupling has been shown to have good to moderate reproducibility on EGG; whereas parameters related to the power of gastric slow waves and maximum dominant frequency show poorer reproducibility. 19,21However, in the current study, Principal Gastric Frequency demonstrated excellent reproducibility, with all Lin's CCCs above 0.89, demonstrating its superiority over traditional EGG metrics.In addition, the BMIadjusted amplitude also showed good reproducibility, compared to the poor reproducibility of EGG power, which likely reflects the superior signal-to-noise ratio achieved by the high-resolution array in combination with advanced signal processing and artifact rejection algorithms. 30,38ile the primary metrics showed overall high reproducibility, both the pre-meal and in the first hour postprandial exhibited lower reproducibility than other time periods.This was not unexpected given the stomach's preprandial quiescence, whereby electrical activities are often decoupled from mechanical contractions, thereby affecting metrics driven by signal strength (such as BMI-Adjusted Amplitude and GA-RI). 39The transition from a fasted to fed state is typically completed within the initial 30 min post-ingestion and can exhibit variability, especially in pathologies like autonomic neuropathy and myopathic disorders. 48Variabilities have also been observed in GES studies, with a well-documented "lag phase" following meal ingestion which can range up to an hour depending on the measurement method, and with the clinical implications incompletely understood. 49Given these inherent variances, particularly during transitional phases, there is a consensus advocating for extended study durations for gastric function testing beyond 1 h postprandially.Recent comparative analyses between BSGM and EGG further Comparison of Gastric Alimetry reproducibility statistics for Principal Gastric Frequency, BMI-Adjusted Amplitude, and GA-RI with similar gastric motility tests analyzed in previous literature using: (A) Lin's concordance correlation coefficient (CCC), and (B) intra-individual variation (CV intra ).Bars represent the range in reproducibility statistics for the primary metrics measured within each study.Comparative manuscripts were: Desai et al. 24 ; Lartigue et al. 16 ; Roland et al. 13 ; Cremonini et al. 12 ; Camilleri et al. 25 ; Horner et al. 9 ; DiBaise et al. 8 ; Diaz Tartera et al. 22  SmartPill [22]   EGG [8]   GEBT [9]   GES [25]   GES [12]   GES [13]   GES [16]   GES [24]   Gastric Alimetry® Intra-individual VariaƟon (%) emphasize the merits of a prolonged observational period, which contributes to the improved sensitivity and specificity of the superior Gastric Alimetry metrics. 32,39assuringly, similar statistical reporting was observed across both short term and long term in both patients and controls.To be included in this study, all patients and controls had to have no significant changes in their clinical disease status and symptomatology since their first test.Therefore, reassurance can be provided that in the absence of disease progression or status change, the Gastric Differences between healthy controls and patients were observed for GA-RI, BMI-Adjusted Amplitude and ff-AR, with half of the patients studied having a spectral abnormality.These findings are consistent with other recent studies showing the capability of the Gastric Alimetry System to diagnose specific subgroups of patients with gastric neuromuscular disorders. 28,30,31In contrast, no differences in Principal Gastric Frequency deviation were observed between patients and controls.However, this finding is expected, as differences in the frequency of healthy controls and patients are not normally anticipated except in specific patient populations such as long-term diabetics and post-fundoplication patients in the case of high frequencies 31,50 and post-surgical patients in the case of low frequencies. 51rthermore, the Gastric Alimetry phenotypes remained consistent in 88% of participants short term and 79% long term.This is higher than for previous reproducibility research on GES, the current gold standard of gastric motility testing, which only shows consistent diagnoses in 58%-70% of patients. 5,24Additionally, all participants who were categorized as different phenotypes across the tests were considered borderline cases, as their relevant metrics were on the edge of the normative reference intervals.This further highlights the consistency of the metrics, while emphasizing the importance of clinical discretion when phenotyping patients based on normative reference intervals and using these for clinical decisionmaking, particularly when patients lie at the edges of the interval.
Of interest, the Rome criteria demonstrated lower reproducibility than the Gastric Alimetry phenotypes over the course of the study, as has previously been reported in irritable bowel syndrome, 52 indicating that chronic gastroduodenal disorders inherently wax and wane in intensity over time in a proportion of patients.This is particularly concerning short term, as the Rome criteria recalls symptoms over the last 6 months and therefore is not expected to vary week to week.Previous research has commented on the issues with the Rome criteria including insufficient specificity, 53 poor clinical utility, 53,54 and long recall period. 55The objective nature of the Gastric Alimetry spectral phenotypes may help to remove this recall bias and add more specificity to guide clinical management.
The overall consistency of these phenotypes indicates their use in aiding the development of tailored and integrated management plans.For example, a recent paper by Gharibans et al. 28 demonstrated that patients with chronic nausea and vomiting could be effectively subgrouped using Gastric Alimetry into those with neuromuscular disorders and those with normal gastric electrophysiology and higher psychological comorbidity, indicating different treatment pathways.In another study, Wang et al. 33 recently showed that Gastric Alimetry expands the phenotyping and detection of motility disorders in patients with gastroduodenal disorders, in comparison to gastric emptying testing.Therefore, the phenotypes derived from a Gastric Alimetry test could have implications for the diagnostic profiling and personalized management of gastroduodenal disorders.
A strength of this study was that it independently analyzed the reproducibility of the Gastric Alimetry metrics for patients and healthy controls.Many reproducibility studies for other gastric motility tests only recruited healthy controls, which limits the generalizability of their results to the target populations of the tests.This also does not allow for the consideration of the effect of cyclical symptomatology and disease progression on the test results.
However, there are limitations to this study that need to be noted.First, as there are multiple spectral phenotypes for patients with gastroduodenal disorders, not all of these subgroups may have been adequately evaluated within the current sample, which may have underestimated the true variability of the Gastric Alimetry metrics in certain groups.For example, only one patient had a low enough GA-RI score to be categorized as having low rhythm stability.Therefore, future research is needed to definitively explore the reproducibility of this phenotype over time.Although the tests were kept as standardized as possible to control for any variability, there was slight variation in the test timings and three participants switched to alternative meals of equal caloric content.However, research has shown no significant differences in the results of EGG 56 or BSGM 57 for calorie-matched meals with different compositions.
In addition, we did not account for menstrual cycle variability in premenopausal female subjects, with elevations in Principal Gastric Frequency recently shown to occur during the luteal phase. 58

| 3 of 13 LAW
et al.Data collection were conducted in Auckland (New Zealand), Calgary (Canada), and Western Sydney (Australia).Ethical approval was obtained for each data collection site via: The Auckland Health Research Ethics Committee (AHREC; AH1130), The University of Calgary Conjoint Health Research Ethics Board (REB19-1925), and the Human Research Ethics Committee at Western Sydney (H15157).All participants provided written informed consent.

F I G U R E 1
Overview of the study procedure and sample retention.Recruitment via mulƟ-naƟonal consorƟum database (n= 229)

| 7 of 13 LAW
(3.19/3.33 cpm, 3.31/3.33cpm, and 3.20/3.20cpm), but a high frequency for the remaining test (3.38,3.37, and 3.38 cpm respectively) despite all measured frequencies F I G U R E 3 Boxplots showing the differences in the four Gastric Alimetry metrics between the three tests, split by controls (top) and patients (bottom).NS represents non-significance using repeated measures ANOVAs.et al. being within 0.15 cpm of the upper limit of the normative interval (3.35 cpm).Therefore, despite changes in phenotypes seen in a minority of participants, these would not be considered clinically significant changes.
Four patients (29%) changed Rome IV diagnosis between Tests 1 and 2 (long term) and three patients (23%) changed Rome diagnosis between Test 2 and 3 (short term).Cohen's κ showed moderate reproducibility for Rome IV diagnosis in patients at both the short (κ = 0.56, p = 0.023) and long terms (κ = 0.48, p = 0.033).All controls indicated no Rome diagnoses at all three tests.

F I G U R E 4 36 TA B L E 1
Abbreviations:Δ , mean change score; ff-AR, Fed: Fasted Amplitude Ratio; GA-RI, Gastric Alimetry Rhythm Index™; range, range of change scores; SD, standard deviation of the change scores.
Alimetry metrics remain consistent over time and are not influenced by test-related variability.This highlights the ability of Gastric Alimetry to monitor changes over time attributable to disease progression or the impact of therapeutic interventions.Consequently, Gastric Alimetry could serve as a valuable tool for quantifying the effects of medications or other treatments on stomach electrophysiology within clinical and research contexts.
Futureresearch is needed to investigate the effect of the menstrual cycle on the reproducibility of the Gastric Alimetry metrics.Nevertheless, despite slight variations in meal type and test timings, this study consistently showed high reproducibility of the Gastric Alimetry metrics, enhancing the ecological validity of these findings and underscoring the stability of these metrics over time.Future research, with larger sample sizes, could be conducted to further explore the reproducibility of the Gastric Alimetry metrics within individual patient groups.For example, further research could focus on different subgroupings of gastroduodenal DGBIs, such as FD and CNVS, or within these subgroupings, such as investigating the postprandial distress syndrome and epigastric pain syndrome groupings within FD.In summary, this study demonstrates the high reproducibility and low intra-individual variation of the Gastric Alimetry BSGM metrics, both short (1 week) and long terms (>6 months) in patients and healthy controls.The superior reproducibility to other gastric motility tests supports the role of Gastric Alimetry as a reliable diagnostic aid for gastric dysfunction in clinical care and supporting test applications in the analyses of disease progression and treatment responsiveness.Phenotyping based on these metrics also showed consistency over time, with changes only observed in those participants who were on the borderline for each phenotype, highlighting the importance of clinical discretion when categorizing patients based on data at the edges of normative reference intervals.AUTH O R CO NTR I B UTI O N S ML, GS, GS, AG, CNA, GOG, and SC designed the research study.ML, DF, and IF performed the research.ML, GS (Gabriel Schamberg), and SC analyzed the data.ML wrote the paper.All authors reviewed and edited the paper.CNA, GOG, and SC provided supervision, resources, and funding.ACK N OWLED G M ENT Open access publishing facilitated by The University of Auckland, as part of the Wiley -The University of Auckland agreement via the Council of Australian University Librarians.FU N D I N G I N FO R M ATI O N This study is funded by a New Zealand Health Research Council Programme Grant 3715588.CO N FLI C T O F I NTER E S T S TATEM ENT GOG, AG and PD hold intellectual property and grants in gastric electrophysiology and are Directors of University of Auckland spinout companies (GOG: Alimetry Ltd., Insides Company; AG: Alimetry Ltd.; PD: FlexiMap).ML, GS, GS, DF, CD, CNA, and SC are members of Alimetry Ltd.The remaining authors have no relevant conflicts to declare.