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

BACKGROUND: 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. METHODS: Participants completed three standardized BSGM tests using Gastric Alimetry® (Alimetry, New Zealand). The test encompassed a fasting baseline (30 min), a 482 kCal standard meal, and a 4 h postprandial recording. The first two tests were >6 months apart and the last occurred ~1 week after the second test, to evaluate long and short-term reproducibility. RESULTS: Fourteen patients with upper gastrointestinal symptoms and 14 healthy controls were recruited. There were no significant differences in any BSGM metrics between the tests at short and long term (all p > 0.180). Lin's concordance correlation coefficients (CCC) for the primary metrics were high, ranging from 0.58 to 0.96, with intra-individual coefficients of variance (CVintra) ranging from 0.2% to 1.9%. Reproducibility was higher, and intra-individual variation lower, than in previous studies of GES (CCC = 0.54-0.83, CVintra = 3%-77%), GEBT (CVintra = 8%-11%), and EGG (CVintra = 3%-78%). CONCLUSIONS: BSGM spectral metrics demonstrate high reproducibility and low intra-individual variation at both short and long term, with superior results to comparable tests. The high reproducibility of Gastric Alimetry supports its role as a diagnostic aid for gastric dysfunction and a reliable tool for evaluating treatment outcomes and disease progression over time.

Effect of Menstrual Cycle and Menopause on Human Gastric Electrophysiology.

Chronic gastroduodenal symptoms disproportionately affect females of childbearing age; however, the effect of menstrual cycling on gastric electrophysiology is poorly defined. To establish the effect of the menstrual cycle on gastric electrophysiology, healthy subjects underwent non-invasive Body Surface Gastric Mapping (BSGM; 8x8 array), with validated symptom logging App (Gastric AlimetryⓇ, New Zealand). Participants were premenopausal females in follicular (n=26) and luteal phases (n=18). Postmenopausal females (n=30) and males (n=51) were controls. Principal gastric frequency (PGF), BMI-adjusted amplitude, Gastric Alimetry Rhythm Index (GA-RI), fasted-fed amplitude ratio (ff-AR), meal response curves, and symptom burden were analysed. Menstrual cycle-related electrophysiological changes were then transferred to an established anatomically-accurate computational gastric fluid dynamics model (meal viscosity 0.1 Pas), to predict the impact on gastric mixing and emptying. PGF was significantly higher in the luteal vs. follicular phase (mean 3.21 cpm, SD (0.17) vs. 2.94 cpm, SD (0.17), p<0.001) and vs. males (3.01 cpm, SD (0.2), p<0.001). In the computational model, this translated to 8.1% higher gastric mixing strength and 5.3% faster gastric emptying for luteal versus follicular phases. Postmenopausal females also exhibited higher PGF than females in the follicular phase (3.10 cpm, SD (0.24) vs. 2.94 cpm, SD (0.17), p=0.01), and higher BMI-adjusted amplitude (40.7 µV (33.02-52.58) vs. 29.6 µV (26.15-39.65), p<0.001), GA-RI (0.60 (0.48-0.73) vs. 0.43 (0.30-0.60), p=0.005), and ff-AR (2.51 (1.79-3.47) vs. 1.48 (1.21-2.17), p=0.001) than males. There were no differences in symptoms. These results define variations in gastric electrophysiology with regard to human menstrual cycling and menopause.

Meal effects on gastric bioelectrical activity utilizing body surface gastric mapping in healthy subjects.

BACKGROUND: Gastric sensorimotor disorders are prevalent. While gastric emptying measurements are commonly used, they may not fully capture the underlying pathophysiology. Body surface gastric mapping (BSGM) recently emerged to assess gastric sensorimotor dysfunction. This study assessed varying meal size on BSGM responses to inform test use in a wider variety of contexts. METHODS: Data from multiple healthy cohorts receiving BSGM were pooled, using four different test meals. A standard BSGM protocol was employed: 30-min fasting, 4-h post-prandial, using Gastric Alimetry® (Alimetry, New Zealand). Meals comprised: (i) nutrient drink + oatmeal bar (482 kcal; 'standard meal'); (ii) oatmeal bar alone; egg and toast meal, and pancake (all ~250 kcal). Gastric Alimetry metrics included BMI-adjusted Amplitude, Principal Gastric Frequency, Gastric Alimetry Rhythm Index (GA-RI) and Fed:Fasted Amplitude Ratio (ff-AR). KEY RESULTS: 238 participants (59.2% female) were included. All meals significantly increased amplitude and frequency during the first postprandial hour (p < 0.05). There were no differences in postprandial frequency across meals (p > 0.05). The amplitude and GA-RI of the standard meal (n = 110) were significantly higher than the energy bar alone (n = 45) and egg meal (n = 65) (all p < 0.05). All BSGM metrics were comparable across the three smaller meals (p > 0.05). A higher symptom burden was found in the oatmeal bar group versus the standard meal and pancake meal (p = 0.01, 0.003, respectively). CONCLUSIONS & INFERENCES: The consumption of lower calorie meals elicited different postprandial responses, when compared to the standard Gastric Alimetry meal. These data will guide interpretations of BSGM when applied with lower calorie meals.

Transcutaneous Auricular Vagus Nerve Stimulation Normalizes Induced Gastric Myoelectrical Dysrhythmias in Controls Assessed by Body-Surface Gastric Mapping.

OBJECTIVES: Transcutaneous auricular vagus nerve stimulation (TaVNS) is a supplementary treatment for gastric symptoms resulting from dysrhythmias. The main objective of this study was to quantify the effects of 10, 40, and 80 Hz TaVNS and sham in healthy individuals in response to a 5-minute water-load (WL5) test. MATERIALS AND METHODS: Eighteen healthy volunteers aged between 21 and 55 years (body mass index: 27.1 ± 3.2) were recruited. Each subject fasted for up to eight hours and participated in four 95-minute sessions, which consisted of 30 fasted baseline, 30 minutes TaVNS, WL5, and 30 minutes post-WL5. Heart rate variability was assessed using sternal electrocardiogram. Body-surface gastric mapping and bloating (/10) were recorded. One-way ANOVA with post hoc Tukey test was performed to test the difference between TaVNS protocols in terms of frequency, amplitude, bloating scores, root mean square of the successive differences (RMSSD), and stress index (SI). RESULTS: On average, the subjects consumed 526 ± 160 mL of water, with volume ingested correlated to bloating (mean score 4.1 ± 1.8; r = 0.36, p = 0.029). In general, the reduction in frequency and rhythm stability during the post-WL5 period in sham was normalized by all three TaVNS protocols. Both 40- and 80-Hz protocols also caused increases in amplitude during the stim-only and/or post-WL5 periods. RMSSD increased during the 40-Hz protocol. SI increased during the 10-Hz protocol but decreased during the 40- and 80-Hz protocols. CONCLUSION: TaVNS proved effective in normalizing gastric dysrhythmias by WL5 in healthy subjects by altering both parasympathetic and sympathetic pathways.

Gastric Alimetry Expands Patient Phenotyping in Gastroduodenal Disorders Compared with Gastric Emptying Scintigraphy.

INTRODUCTION: Gastric emptying testing (GET) assesses gastric motility, however, is nonspecific and insensitive for neuromuscular disorders. Gastric Alimetry (GA) is a new medical device combining noninvasive gastric electrophysiological mapping and validated symptom profiling. This study assessed patient-specific phenotyping using GA compared with GET. METHODS: Patients with chronic gastroduodenal symptoms underwent simultaneous GET and GA, comprising a 30-minute baseline, 99m TC-labelled egg meal, and 4-hour postprandial recording. Results were referenced to normative ranges. Symptoms were profiled in the validated GA App and phenotyped using rule-based criteria based on their relationships to the meal and gastric activity: (i) sensorimotor, (ii) continuous, and (iii) other. RESULTS: Seventy-five patients were assessed, 77% female. Motility abnormality detection rates were as follows: GET 22.7% (14 delayed, 3 rapid), GA spectral analysis 33.3% (14 low rhythm stability/low amplitude, 5 high amplitude, and 6 abnormal frequency), and combined yield 42.7%. In patients with normal spectral analysis, GA symptom phenotypes included sensorimotor 17% (where symptoms strongly paired with gastric amplitude, median r = 0.61), continuous 30%, and other 53%. GA phenotypes showed superior correlations with Gastroparesis Cardinal Symptom Index, Patient Assessment of Upper Gastrointestinal Symptom Severity Index, and anxiety scales, whereas Rome IV Criteria did not correlate with psychometric scores ( P > 0.05). Delayed emptying was not predictive of specific GA phenotypes. DISCUSSION: GA improves patient phenotyping in chronic gastroduodenal disorders in the presence and absence of motility abnormalities with increased correlation with symptoms and psychometrics compared with gastric emptying status and Rome IV criteria. These findings have implications for the diagnostic profiling and personalized management of gastroduodenal disorders.

Longitudinal outcome monitoring in patients with chronic gastroduodenal symptoms investigated using the Gastric Alimetry system: study protocol.

INTRODUCTION: The Gastric Alimetry platform offers a multimodal assessment of gastric function through body surface gastric mapping (BSGM) and concurrent symptom-tracking via a validated App. We aim to perform a longitudinal cohort study to examine the impact of Gastric Alimetry, and changes in clinical management on patient symptoms, quality of life and psychological health. METHODS AND ANALYSIS: This is a prospective multicentre longitudinal observational cohort study of participants with chronic gastroduodenal symptoms. Consecutive participants undergoing Gastric Alimetry will be invited to participate. Quality of life will be assessed via EuroQol-5D and the Patient Assessment of Upper Gastrointestinal Disorders-Quality of Life score. Gastrointestinal symptoms will be assessed via the Patient Assessment of Upper Gastrointestinal Symptom Severity index, and the Gastroparesis Cardinal Symptom Index. Psychometrics will be assessed, including anxiety via the General Anxiety Disorder-7, perceived stress using the Perceived Stress Scale 4, and depression via the Patient Health Questionnaire 9. Clinical parameters including diagnoses, investigations and treatments (medication and procedures) will also be captured. Assessments will be made the week after the BSGM test, at 30 days, 90 days, 180 days and 360 days thereafter. The primary outcome is feasibility of longitudinal follow-up of a cohort that have undergone Gastric Alimetry testing; from which patients' continuum of care can be characterised. Secondary outcomes include changes in patient-reported symptoms, quality of life and psychometrics (anxiety, stress and depression). Inferential causal analyses will be performed at the within patient level to explore causal associations between treatment changes and clinical outcomes. The impact of Gastric Alimetry on clinical management will also be captured. ETHICS AND DISSEMINATION: The protocol has been approved in Aotearoa New Zealand by the Auckland Health Research Ethics Committee. Results will be submitted for conference presentation and peer-reviewed publication.

Gastric Alimetry® Test Interpretation in Gastroduodenal Disorders: Review and Recommendations.

Chronic gastroduodenal symptoms are prevalent worldwide, and there is a need for new diagnostic and treatment approaches. Several overlapping processes may contribute to these symptoms, including gastric dysmotility, hypersensitivity, gut-brain axis disorders, gastric outflow resistance, and duodenal inflammation. Gastric Alimetry® (Alimetry, New Zealand) is a non-invasive test for evaluating gastric function that combines body surface gastric mapping (high-resolution electrophysiology) with validated symptom profiling. Together, these complementary data streams enable important new clinical insights into gastric disorders and their symptom correlations, with emerging therapeutic implications. A comprehensive database has been established, currently comprising > 2000 Gastric Alimetry tests, including both controls and patients with various gastroduodenal disorders. From studies employing this database, this paper presents a systematic methodology for Gastric Alimetry test interpretation, together with an extensive supporting literature review. Reporting is grouped into four sections: Test Quality, Spectral Analysis, Symptoms, and Conclusions. This review compiles, assesses, and evaluates each of these aspects of test assessment, with discussion of relevant evidence, example cases, limitations, and areas for future work. The resultant interpretation methodology is recommended for use in clinical practice and research to assist clinicians in their use of Gastric Alimetry as a diagnostic aid and is expected to continue to evolve with further development.

Comparison of Gastric Alimetry® body surface gastric mapping versus electrogastrography spectral analysis.

Electrogastrography (EGG) non-invasively evaluates gastric motility but is viewed as lacking clinical utility. Gastric Alimetry® is a new diagnostic test that combines high-resolution body surface gastric mapping (BSGM) with validated symptom profiling, with the goal of overcoming EGG's limitations. This study directly compared EGG and BSGM to define performance differences in spectral analysis. Comparisons between Gastric Alimetry BSGM and EGG were conducted by protocolized retrospective evaluation of 178 subjects [110 controls; 68 nausea and vomiting (NVS) and/or type 1 diabetes (T1D)]. Comparisons followed standard methodologies for each test (pre-processing, post-processing, analysis), with statistical evaluations for group-level differences, symptom correlations, and patient-level classifications. BSGM showed substantially tighter frequency ranges vs EGG in controls. Both tests detected rhythm instability in NVS, but EGG showed opposite frequency effects in T1D. BSGM showed an 8× increase in the number of significant correlations with symptoms. BSGM accuracy for patient-level classification was 0.78 for patients vs controls and 0.96 as compared to blinded consensus panel; EGG accuracy was 0.54 and 0.43. EGG detected group-level differences in patients, but lacked symptom correlations and showed poor accuracy for patient-level classification, explaining EGG's limited clinical utility. BSGM demonstrated substantial performance improvements across all domains.

Gastric Alimetry ® improves patient phenotyping in gastroduodenal disorders compared to gastric emptying scintigraphy alone.

Objectives: Gastric emptying testing (GET) assesses gastric motility, however is non-specific and insensitive for neuromuscular disorders. Gastric Alimetry® (GA) is a new medical device combining non-invasive gastric electrophysiological mapping and validated symptom profiling. This study assessed patient-specific phenotyping using GA compared to GET. Methods: Patients with chronic gastroduodenal symptoms underwent simultaneous GET and GA, comprising a 30-minute baseline, 99m TC-labelled egg meal, and 4-hour postprandial recording. Results were referenced to normative ranges. Symptoms were profiled in the validated GA App and phenotyped using rule-based criteria based on their relationships to the meal and gastric activity: i) sensorimotor; ii) continuous; and iii) other. Results: 75 patients were assessed; 77% female. Motility abnormality detection rates were: GET 22.7% (14 delayed, 3 rapid); GA spectral analysis 33.3% (14 low rhythm stability / low amplitude; 5 high amplitude; 6 abnormal frequency); combined yield 42.7%. In patients with normal spectral analysis, GA symptom phenotypes included: sensorimotor 17% (where symptoms strongly paired with gastric amplitude; median r=0.61); continuous 30%; other 53%. GA phenotypes showed superior correlations with GCSI, PAGI-SYM, and anxiety scales, whereas Rome IV Criteria did not correlate with psychometric scores (p>0.05). Delayed emptying was not predictive of specific GA phenotypes. Conclusions: GA improves patient phenotyping in chronic gastroduodenal disorders in the presence and absence of motility abnormalities with improved correlation with symptoms and psychometrics compared to gastric emptying status and Rome IV criteria. These findings have implications for the diagnostic profiling and personalized management of gastroduodenal disorders. Study Highlights: 1) WHAT IS KNOWN Chronic gastroduodenal symptoms are common, costly and greatly impact on quality of lifeThere is a poor correlation between gastric emptying testing (GET) and symptomsGastric Alimetry® is a new medical device combining non-invasive gastric electrophysiological mapping and validated symptom profiling 2) WHAT IS NEW HERE Gastric Alimetry generates a 1.5x higher yield for motility abnormalities than GETWith symptom profiling, Gastric Alimetry identified 2.7x more specific patient categories than GETGastric Alimetry improves clinical phenotyping, with improved correlation with symptoms and psychometrics compared to GET.

Principles and clinical methods of body surface gastric mapping: Technical review.

BACKGROUND AND PURPOSE: Chronic gastric symptoms are common, however differentiating specific contributing mechanisms in individual patients remains challenging. Abnormal gastric motility is present in a significant subgroup, but reliable methods for assessing gastric motor function in clinical practice are lacking. Body surface gastric mapping (BSGM) is a new diagnostic aid, employs multi-electrode arrays to measure and map gastric myoelectrical activity non-invasively in high resolution. Clinical adoption of BSGM is currently expanding following studies demonstrating the ability to achieve specific patient subgrouping, and subsequent regulatory clearances. An international working group was formed in order to standardize clinical BSGM methods, encompassing a technical group developing BSGM methods and a clinical advisory group. The working group performed a technical literature review and synthesis focusing on the rationale, principles, methods, and clinical applications of BSGM, with secondary review by the clinical group. The principles and validation of BSGM were evaluated, including key advances achieved over legacy electrogastrography (EGG). Methods for BSGM were reviewed, including device design considerations, patient preparation, test conduct, and data processing steps. Recent advances in BSGM test metrics and reference intervals are discussed, including four novel metrics, being the 'principal gastric frequency', BMI-adjusted amplitude, Gastric Alimetry Rhythm Index™, and fed: fasted amplitude ratio. An additional essential element of BSGM has been the introduction of validated digital tools for standardized symptom profiling, performed simultaneously during testing. Specific phenotypes identifiable by BSGM and the associated symptom profiles were codified with reference to pathophysiology. Finally, knowledge gaps and priority areas for future BSGM research were also identified by the working group.

Normative Values for Body Surface Gastric Mapping Evaluations of Gastric Motility Using Gastric Alimetry: Spectral Analysis.

INTRODUCTION: Body surface gastric mapping (BSGM) is a new noninvasive test of gastric function. BSGM offers several novel and improved biomarkers of gastric function capable of differentiating patients with overlapping symptom profiles. The aim of this study was to define normative reference intervals for BSGM spectral metrics in a population of healthy controls. METHODS: BSGM was performed in healthy controls using Gastric Alimetry (Alimetry, New Zealand) comprising a stretchable high-resolution array (8 × 8 electrodes; 196 cm 2 ), wearable Reader, and validated symptom-logging App. The evaluation encompassed a fasting baseline (30 minutes), 482 kCal meal, and 4-hour postprandial recording. Normative reference intervals were calculated for BSGM metrics including the Principal Gastric Frequency, Gastric Alimetry Rhythm Index (a measure of the concentration of power in the gastric frequency band over time), body mass index (BMI)-adjusted amplitude (µV), and fed:fasted amplitude ratio. Data were reported as median and reference interval (5th and/or 95th percentiles). RESULTS: A total of 110 subjects (55% female, median age 32 years [interquartile range 24-50], median BMI 23.8 kg/m 2 [interquartile range 21.4-26.9]) were included. The median Principal Gastric Frequency was 3.04 cycles per minute; reference interval: 2.65-3.35 cycles per minute. The median Gastric Alimetry Rhythm Index was 0.50; reference interval: ≥0.25. The median BMI-adjusted amplitude was 37.6 µV; reference interval: 20-70 µV. The median fed:fasted amplitude ratio was 1.85; reference interval ≥1.08. A higher BMI was associated with a shorter meal-response duration ( P = 0.014). DISCUSSION: This study provides normative reference intervals for BSGM spectral data to inform diagnostic interpretations of abnormal gastric function.

Revised spectral metrics for body surface measurements of gastric electrophysiology.

BACKGROUND: Electrogastrography (EGG) non-invasively evaluates gastric function but has not achieved common clinical adoption due to several technical limitations. Body Surface Gastric Mapping (BSGM) has been introduced to overcome these limitations, but pitfalls in traditional metrics used to analyze spectral data remain unaddressed. This study critically evaluates five traditional EGG metrics and introduces improved BSGM spectral metrics, with validation in a large cohort. METHODS: Pitfalls in five EGG metrics were assessed (dominant frequency, percentage time normogastria, amplitude, power ratio, and instability coefficient), leading to four revised BSGM spectral metrics. Traditional and revised metrics were compared to validate performance using a standardized 100-subject database of BSGM tests (30 min baseline; 4-h postprandial) recorded using Gastric Alimetry® (Alimetry). KEY RESULTS: BMI and amplitude were highly correlated (r = -0.57, p < 0.001). We applied a conservative BMI correction to obtain a BMI-adjusted amplitude metric (r = -0.21, p = 0.037). Instability coefficient was highly correlated with both dominant frequency (r = -0.44, p < 0.001), and percent bradygastria (r = 0.85, p < 0.001), in part due to misclassification of low frequency transients as gastric activity. This was corrected by introducing distinct gastric frequency and stability metrics (Principal Gastric Frequency and Gastric Alimetry Rhythm Index (GA-RI)TM ) that were uncorrelated (r = 0.14, p = 0.314). Only 28% of subjects showed a maximal averaged amplitude within the first postprandial hour. Calculating Fed:Fasted Amplitude Ratio over a 4-h postprandial window yielded a median increase of 0.31 (IQR 0-0.64) above the traditional ratio. CONCLUSIONS & INFERENCES: The revised metrics resolve critical pitfalls impairing the performance of traditional EGG, and should be applied in future BSGM spectral analyses.

Gastric dysfunction in patients with chronic nausea and vomiting syndromes defined by a noninvasive gastric mapping device.

Chronic nausea and vomiting syndromes (NVSs) are prevalent and debilitating disorders. Putative mechanisms include gastric neuromuscular disease and dysregulation of brain-gut interaction, but clinical tests for objectively defining gastric motor function are lacking. A medical device enabling noninvasive body surface gastric mapping (BSGM) was developed and applied to evaluate NVS pathophysiology. BSGM was performed in 43 patients with NVS and 43 matched controls using Gastric Alimetry (Alimetry), a conformable high-resolution array (8 × 8 electrodes; 20-mm interelectrode spacing), wearable reader, and validated symptom-logging app. Continuous measurement encompassed a fasting baseline (30 minutes), 482-kilocalorie meal, and 4-hour postprandial recording, followed by spectral and spatial biomarker analyses. Meal responses were impaired in NVS, with reduced amplitudes compared to controls (median, 23.3 microvolts versus 38.0 microvolts, P < 0.001), impaired fed-fasting power ratios (1.1 versus 1.6, P = 0.02), and disorganized slow waves (spatial frequency stability, 13.6 versus 49.5; P < 0.001). Two distinct NVS subgroups were evident with indistinguishable symptoms (all P > 0.05). Most patients (62%) had normal BSGM studies with increased psychological comorbidities (43.5% versus 7.7%; P = 0.03) and anxiety scores (median, 16.5 versus 13.0; P = 0.035). A smaller subgroup (31%) had markedly abnormal BSGM, with biomarkers correlating with symptoms (nausea, pain, excessive fullness, early satiety, and bloating; all r > 0.35, P < 0.05). Patients with NVS share overlapping symptoms but comprise distinct underlying phenotypes as revealed by a BSGM device. These phenotypes correlate with symptoms, which should inform clinical management and therapeutic trial design.

Modeling relationships between rhythmic processes and neuronal spike timing.

Neurons are embedded in complex networks, where they participate in repetitive, coordinated interactions with other neurons. Neuronal spike timing is thus predictably constrained by a range of ionic currents that shape activity at both short (milliseconds) and longer (tens to hundreds of milliseconds) timescales, but we lack analytical tools to rigorously identify these relationships. Here, we innovate a modeling approach to test the relationship between oscillations in the local field potential (LFP) and neuronal spike timing. We use kernel density estimation to relate single neuron spike timing and the phase of LFP rhythms (in simulated and hippocampal CA1 neuronal spike trains). We then combine phase and short (3 ms) spike history information within a logistic regression framework ("phaseSH models"), and show that models that leverage refractory constraints and oscillatory phase information can effectively test whether-and the degree to which-rhythmic currents (as measured from the LFP) reliably explain variance in neuronal spike trains. This approach allows researchers to systematically test the relationship between oscillatory activity and neuronal spiking dynamics as they unfold over time and as they shift to adapt to distinct behavioral conditions.NEW & NOTEWORTHY Statistical models that incorporate neural spiking history and relationships to the phase of ongoing oscillations in the local field potential robustly capture and predict neuronal engagement in rhythmic processes. These models constitute a powerful tool to systematically test explicit hypotheses regarding the specific rhythmic currents that constrain neural spiking activity over time and during different behaviors.

An automated artifact detection and rejection system for body surface gastric mapping.

BACKGROUND: Body surface gastric mapping (BSGM) is a new clinical tool for gastric motility diagnostics, providing high-resolution data on gastric myoelectrical activity. Artifact contamination was a key challenge to reliable test interpretation in traditional electrogastrography. This study aimed to introduce and validate an automated artifact detection and rejection system for clinical BSGM applications. METHODS: Ten patients with chronic gastric symptoms generated a variety of artifacts according to a standardized protocol (176 recordings) using a commercial BSGM system (Alimetry, New Zealand). An automated artifact detection and rejection algorithm was developed, and its performance was compared with a reference standard comprising consensus labeling by 3 analysis experts, followed by comparison with 6 clinicians (3 untrained and 3 trained in artifact detection). Inter-rater reliability was calculated using Fleiss' kappa. KEY RESULTS: Inter-rater reliability was 0.84 (95% CI:0.77-0.90) among experts, 0.76 (95% CI:0.68-0.83) among untrained clinicians, and 0.71 (95% CI:0.62-0.79) among trained clinicians. The sensitivity and specificity of the algorithm against experts was 96% (95% CI:91%-100%) and 95% (95% CI:90%-99%), respectively, vs 77% (95% CI:68%-85%) and 99% (95% CI:96%-100%) against untrained clinicians, and 97% (95% CI:92%-100%) and 88% (95% CI:82%-94%) against trained clinicians. CONCLUSIONS & INFERENCES: An automated artifact detection and rejection algorithm was developed showing >95% sensitivity and specificity vs expert markers. This algorithm overcomes an important challenge in the clinical translation of BSGM and is now being routinely implemented in patient test interpretations.

Continuous action deep reinforcement learning for propofol dosing during general anesthesia.

PURPOSE: Anesthesiologists simultaneously manage several aspects of patient care during general anesthesia. Automating administration of hypnotic agents could enable more precise control of a patient's level of unconsciousness and enable anesthesiologists to focus on the most critical aspects of patient care. Reinforcement learning (RL) algorithms can be used to fit a mapping from patient state to a medication regimen. These algorithms can learn complex control policies that, when paired with modern techniques for promoting model interpretability, offer a promising approach for developing a clinically viable system for automated anesthestic drug delivery. METHODS: We expand on our prior work applying deep RL to automated anesthetic dosing by now using a continuous-action model based on the actor-critic RL paradigm. The proposed RL agent is composed of a policy network that maps observed anesthetic states to a continuous probability density over propofol-infusion rates and a value network that estimates the favorability of observed states. We train and test three versions of the RL agent using varied reward functions. The agent is trained using simulated pharmacokinetic/pharmacodynamic models with randomized parameters to ensure robustness to patient variability. The model is tested on simulations and retrospectively on nine general anesthesia cases collected in the operating room. We utilize Shapley additive explanations to gain an understanding of the factors with the greatest influence over the agent's decision-making. RESULTS: The deep RL agent significantly outperformed a proportional-integral-derivative controller (median episode median absolute performance error 1.9% ± 1.8 and 3.1% ± 1.1). The model that was rewarded for minimizing total doses performed the best across simulated patient demographics (median episode median performance error 1.1% ± 0.5). When run on real-world clinical datasets, the agent recommended doses that were consistent with those administered by the anesthesiologist. CONCLUSIONS: The proposed approach marks the first fully continuous deep RL algorithm for automating anesthestic drug dosing. The reward function used by the RL training algorithm can be flexibly designed for desirable practices (e.g. use less anesthetic) and bolstered performances. Through careful analysis of the learned policies, techniques for interpreting dosing decisions, and testing on clinical data, we confirm that the agent's anesthetic dosing is consistent with our understanding of best-practices in anesthesia care.

Perioperative Multimodal General Anesthesia Focusing on Specific CNS Targets in Patients Undergoing Cardiac Surgeries: The Pathfinder Feasibility Trial.

Multimodal general anesthesia (MMGA) is a strategy that utilizes the well-known neuroanatomy and neurophysiology of nociception and arousal control in designing a rational and clinical practical paradigm to regulate the levels of unconsciousness and antinociception during general anesthesia while mitigating side effects of any individual anesthetic. We sought to test the feasibility of implementing MMGA for seniors undergoing cardiac surgery, a high-risk cohort for hemodynamic instability, delirium, and post-operative cognitive dysfunction. Twenty patients aged 60 or older undergoing on-pump coronary artery bypass graft (CABG) surgery or combined CABG/valve surgeries were enrolled in this non-randomized prospective observational feasibility trial, wherein we developed MMGA specifically for cardiac surgeries. Antinociception was achieved by a combination of intravenous remifentanil, ketamine, dexmedetomidine, and magnesium together with bupivacaine administered as a pecto-intercostal fascial block. Unconsciousness was achieved by using electroencephalogram (EEG)-guided administration of propofol along with the sedative effects of the antinociceptive agents. EEG-guided MMGA anesthesia was safe and feasible for cardiac surgeries, and exploratory analyses found hemodynamic stability and vasopressor usage comparable to a previously collected cohort. Intraoperative EEG suppression events and postoperative delirium were found to be rare. We report successful use of a total intravenous anesthesia (TIVA)-based MMGA strategy for cardiac surgery and establish safety and feasibility for studying MMGA in a full clinical trial. Clinical Trial Number: www.clinicaltrials.gov; identifier NCT04016740 (https://clinicaltrials.gov/ct2/show/NCT04016740).

Etiology of Burst Suppression EEG Patterns.

Burst-suppression electroencephalography (EEG) patterns of electrical activity, characterized by intermittent high-power broad-spectrum oscillations alternating with isoelectricity, have long been observed in the human brain during general anesthesia, hypothermia, coma and early infantile encephalopathy. Recently, commonalities between conditions associated with burst-suppression patterns have led to new insights into the origin of burst-suppression EEG patterns, their effects on the brain, and their use as a therapeutic tool for protection against deleterious neural states. These insights have been further supported by advances in mechanistic modeling of burst suppression. In this Perspective, we review the origins of burst-suppression patterns and use recent insights to weigh evidence in the controversy regarding the extent to which burst-suppression patterns observed during profound anesthetic-induced brain inactivation are associated with adverse clinical outcomes. Whether the clinical intent is to avoid or maintain the brain in a state producing burst-suppression patterns, monitoring and controlling neural activity presents a technical challenge. We discuss recent advances that enable monitoring and control of burst suppression.

Machine learning of EEG spectra classifies unconsciousness during GABAergic anesthesia.

In current anesthesiology practice, anesthesiologists infer the state of unconsciousness without directly monitoring the brain. Drug- and patient-specific electroencephalographic (EEG) signatures of anesthesia-induced unconsciousness have been identified previously. We applied machine learning approaches to construct classification models for real-time tracking of unconscious state during anesthesia-induced unconsciousness. We used cross-validation to select and train the best performing models using 33,159 2s segments of EEG data recorded from 7 healthy volunteers who received increasing infusions of propofol while responding to stimuli to directly assess unconsciousness. Cross-validated models of unconsciousness performed very well when tested on 13,929 2s EEG segments from 3 left-out volunteers collected under the same conditions (median volunteer AUCs 0.99-0.99). Models showed strong generalization when tested on a cohort of 27 surgical patients receiving solely propofol collected in a separate clinical dataset under different circumstances and using different hardware (median patient AUCs 0.95-0.98), with model predictions corresponding with actions taken by the anesthesiologist during the cases. Performance was also strong for 17 patients receiving sevoflurane (alone or in addition to propofol) (median AUCs 0.88-0.92). These results indicate that EEG spectral features can predict unconsciousness, even when tested on a different anesthetic that acts with a similar neural mechanism. With high performance predictions of unconsciousness, we can accurately monitor anesthetic state, and this approach may be used to engineer infusion pumps to intelligibly respond to patients' neural activity.