Bayesian inverse methods for spatiotemporal characterization of gastric electrical activity from cutaneous multi-electrode recordings

Artificial Intelligence Tools for Improving Manometric Diagnosis of Esophageal Dysmotility

PURPOSE OF REVIEW

Artificial intelligence (AI) is a broad term that pertains to a computer's ability to mimic and sometimes surpass human intelligence in interpretation of large datasets. The adoption of AI in gastrointestinal motility has been slower compared to other areas such as polyp detection and interpretation of histopathology.

RECENT FINDINGS

Within esophageal physiologic testing, AI can automate interpretation of image-based tests, especially high resolution manometry (HRM) and functional luminal imaging probe (FLIP) studies. Basic tasks such as identification of landmarks, determining adequacy of the HRM study and identification from achalasia from non-achalasia patterns are achieved with good accuracy. However, existing AI systems compare AI interpretation to expert analysis rather than to clinical outcome from management based on AI diagnosis. The use of AI methods is much less advanced within the field of ambulatory reflux monitoring, where challenges exist in assimilation of data from multiple impedance and pH channels. There remains potential for replication of the AI successes within esophageal physiologic testing to HRM of the anorectum, and to innovative and novel methods of evaluating gastric electrical activity and motor function. The use of AI has tremendous potential to improve detection of dysmotility within the esophagus using esophageal physiologic testing, as well as in other regions of the gastrointestinal tract. Eventually, integration of patient presentation, demographics and alternate test results to individual motility test interpretation will improve diagnostic precision and prognostication using AI tools.

Unsupervised learning of stationary and switching dynamical system models from Poisson observations

Objective. Investigating neural population dynamics underlying behavior requires learning accurate models of the recorded spiking activity, which can be modeled with a Poisson observation distribution. Switching dynamical system models can offer both explanatory power and interpretability by piecing together successive regimes of simpler dynamics to capture more complex ones. However, in many cases, reliable regime labels are not available, thus demanding accurate unsupervised learning methods for Poisson observations. Existing learning methods, however, rely on inference of latent states in neural activity using the Laplace approximation, which may not capture the broader properties of densities and may lead to inaccurate learning. Thus, there is a need for new inference methods that can enable accurate model learning.Approach. To achieve accurate model learning, we derive a novel inference method based on deterministic sampling for Poisson observations called the Poisson Cubature Filter (PCF) and embed it in an unsupervised learning framework. This method takes a minimum mean squared error approach to estimation. Terms that are difficult to find analytically for Poisson observations are approximated in a novel way with deterministic sampling based on numerical integration and cubature rules.Main results. PCF enabled accurate unsupervised learning in both stationary and switching dynamical systems and largely outperformed prior Laplace approximation-based learning methods in both simulations and motor cortical spiking data recorded during a reaching task. These improvements were larger for smaller data sizes, showing that PCF-based learning was more data efficient and enabled more reliable regime identification. In experimental data and unsupervised with respect to behavior, PCF-based learning uncovered interpretable behavior-relevant regimes unlike prior learning methods.Significance. The developed unsupervised learning methods for switching dynamical systems can accurately uncover latent regimes and states in population spiking activity, with important applications in both basic neuroscience and neurotechnology.

Electroanatomical mapping of the stomach with simultaneous biomagnetic measurements

Gastric motility is coordinated by bioelectric slow waves (SWs) and dysrhythmic SW activity has been linked with motility disorders. Magnetogastrography (MGG) is the non-invasive measurement of the biomagnetic fields generated by SWs. Dysrhythmia identification using MGG is currently challenging because source models are not well developed and the impact of anatomical variation is not well understood. A novel method for the quantitative spatial co-registration of serosal SW potentials, MGG, and geometric models of anatomical structures was developed and performed on two anesthetized pigs to verify feasibility. Electrode arrays were localized using electromagnetic transmitting coils. Coil localization error for the volume where the stomach is normally located under the sensor array was assessed in a benchtop experiment, and mean error was 4.2±2.3mm and 3.6±3.3° for a coil orientation parallel to the sensor array and 6.2±5.7mm and 4.5±7.0° for a perpendicular coil orientation. Stomach geometries were reconstructed by fitting a generic stomach to up to 19 localization coils, and SW activation maps were mapped onto the reconstructed geometries using the registered positions of 128 electrodes. Normal proximal-to-distal and ectopic SW propagation patterns were recorded from the serosa and compared against the simultaneous MGG measurements. Correlations between the center-of-gravity of normalized MGG and the mean position of SW activity on the serosa were 0.36 and 0.85 for the ectopic and normal propagation patterns along the proximal-distal stomach axis, respectively. This study presents the first feasible method for the spatial co-registration of MGG, serosal SW measurements, and subject-specific anatomy. This is a significant advancement because these data enable the development and validation of novel non-invasive gastric source characterization methods.

Postprandial effect of gastrointestinal hormones and gastric activity in patients with irritable bowel syndrome

Altered gut regulation, including motor and secretory mechanisms, is characteristic of irritable bowel syndrome (IBS). The severity of postprandial symptoms in IBS patients is associated with discomfort and pain; gas-related symptoms such as bloating and abdominal distension; and abnormal colonic motility. The aim of this study was to assess the postprandial response, i.e., gut peptide secretion and gastric myoelectric activity, in patients with constipation-predominant IBS. The study was conducted on 42 IBS patients (14 males, 28 females, mean age 45.1 ± 15.3 years) and 42 healthy participants (16 males, 26 females, mean age 41.1 ± 8.7 years). The study assessed plasma gut peptide levels (gastrin, CCK-Cholecystokinin, VIP-Vasoactive Intestinal Peptide, ghrelin, insulin) and gastric myoelectric activity obtained from electrogastrography (EGG) in the preprandial and postprandial period (meal-oral nutritional supplement 300 kcal/300 ml). Mean preprandial gastrin and insulin levels were significantly elevated in IBS patients compared to the control group (gastrin: 72.27 ± 26.89 vs. 12.27 ± 4.91 pg/ml; p < 0.00001 and insulin: 15.31 ± 12.92 vs. 8.04 ± 3.21 IU/ml; p = 0.0001), while VIP and ghrelin levels were decreased in IBS patients (VIP: 6.69 ± 4.68 vs. 27.26 ± 21.51 ng/ml; p = 0.0001 and ghrelin: 176.01 ± 88.47 vs. 250.24 ± 84.55 pg/ml; p < 0.0001). A nonsignificant change in the CCK level was observed. IBS patients showed significant changes in postprandial hormone levels compared to the preprandial state-specifically, there were increases in gastrin (p = 0.000), CCK (p < 0.0001), VIP (p < 0.0001), ghrelin (p = 0.000) and insulin (p < 0.0001). Patients with IBS showed reduced preprandial and postprandial normogastria (59.8 ± 22.0 vs. 66.3 ± 20.2%) compared to control values (83.19 ± 16.7%; p < 0.0001 vs. 86.1 ± 9.4%; p < 0.0001). In response to the meal, we did not observe an increase in the percentage of normogastria or the average percentage slow-wave coupling (APSWC) in IBS patients. The postprandial to preprandial power ratio (PR) indicates alterations in gastric contractions; in controls, PR = 2.7, whereas in IBS patients, PR = 1.7, which was significantly lower (p = 0.00009). This ratio reflects a decrease in gastric contractility. Disturbances in the postprandial concentration of gut peptides (gastrin, insulin and ghrelin) in plasma may contribute to abnormal gastric function and consequently intestinal motility, which are manifested in the intensification of clinical symptoms, such as visceral hypersensitivity or irregular bowel movements in IBS patients.

Stomach Geometry Reconstruction Using Serosal Transmitting Coils and Magnetic Source Localization

OBJECTIVE

Bioelectric slow waves (SWs) are a key regulator of gastrointestinal motility, and disordered SW activity has been linked to motility disorders. There is currently a lack of practical options for the acquisition of the 3D stomach geometry during research studies when medical imaging is challenging. Accurately recording the geometry of the stomach and co-registering electrode and sensor positions would provide context for in-vivo studies and aid the development of non-invasive methods of gastric SW assessment.

METHODS

A stomach geometry reconstruction method based on the localization of transmitting coils placed on the gastric serosa was developed. The positions and orientations of the coils, which represented boundary points and surface-normal vectors, were estimated using a magnetic source localization algorithm. Coil localization results were then used to generate surface models. The reconstruction method was evaluated against four 3D-printed anatomically realistic human stomach models and applied in a proof of concept in-vivo pig study.

RESULTS

Over ten repeated reconstructions, average Hausdorff distance and average surface-normal vector error values were 4.7 ±0.2 mm and 18.7 ±0.7° for the whole stomach, and 3.6 ±0.2 mm and 14.6 ±0.6° for the corpus. Furthermore, mean intra-array localization error was 1.4 ±1.1 mm for the benchtop experiment and 1.7 ±1.6 mm in-vivo.

CONCLUSION AND SIGNIFICANCE

Results demonstrated that the proposed reconstruction method is accurate and feasible. The stomach models generated by this method, when co-registered with electrode and sensor positions, could enable the investigation and validation of novel inverse analysis techniques.

Validation of noninvasive body-surface gastric mapping for detecting gastric slow-wave spatiotemporal features by simultaneous serosal mapping in porcine

Gastric disorders are increasingly prevalent, but reliable noninvasive tools to objectively assess gastric function are lacking. Body-surface gastric mapping (BSGM) is a noninvasive method for the detection of gastric electrophysiological features, which are correlated with symptoms in patients with gastroparesis and functional dyspepsia. Previous studies have validated the relationship between serosal and cutaneous recordings from limited number of channels. This study aimed to comprehensively evaluate the basis of BSGM from 64 cutaneous channels and reliably identify spatial biomarkers associated with slow-wave dysrhythmias. High-resolution electrode arrays were placed to simultaneously capture slow waves from the gastric serosa (32 × 6 electrodes at 4 mm spacing) and epigastrium (8 × 8 electrodes at 20 mm spacing) in 14 porcine subjects. BSGM signals were processed based on a combination of wavelet and phase information analyses. A total of 1,185 individual cycles of slow waves were assessed, out of which 897 (76%) were classified as normal antegrade waves, occurring in 10 (71%) subjects studied. BSGM accurately detected the underlying slow wave in terms of frequency (r = 0.99, P = 0.43) as well as the direction of propagation (P = 0.41, F-measure: 0.92). In addition, the cycle-by-cycle match between BSGM and transitions of gastric slow wave dysrhythmias was demonstrated. These results validate BSGM as a suitable method for noninvasively and accurately detecting gastric slow-wave spatiotemporal profiles from the body surface.NEW & NOTEWORTHY Gastric dysfunctions are associated with abnormalities in the gastric bioelectrical slow waves. Noninvasive detection of gastric slow waves from the body surface can be achieved through multichannel, high-resolution, body-surface gastric mapping (BSGM). BSGM matched the spatiotemporal characteristics of gastric slow waves recorded directly and simultaneously from the serosal surface of the stomach. Abnormal gastric slow waves, such as retrograde propagation, ectopic pacemaker, and colliding wavefronts can be detected by changes in the phase of BSGM.

Anatomically Constrained Gastric Slow Wave Localization using Biomagnetic Data

Detection of dysrhythmic gastric slow wave (SW) activity could have significant clinical utility because dysrhyth-mias have been linked to gastric motility disorders. The elec-trogastrogram (EGG) and magnetogastrogram (MGG) enable the non-invasive assessment of SW activity, but most analysis methods can only resolve frequency and velocity. Improved characterization of dysrhythmic propagation patterns from non-invasive measurements is important for the diagnosis of motility disorders and could allow early treatment stratification. In this study, we demonstrate the use of a penalized linear regression framework to localize SW events on the longitudinal stomach axis using simulated MGG data. Priors relating to spatial sparsity, the organization of wavefronts into complete circumferential rings, and the local distribution of depolar-ization and repolarization phases were used to constrain the inverse solution. This method was applied to MGG computed for a single wavefront case and a multiple wavefront case that were constructed from simulated 3 cycle-per-minute normal SW activity. Propagation patterns along the longitudinal stomach axis were identifiable from reconstructed SW activity for both cases. Localization error was 5.7 ± 0.1 mm and 7.7 ± 0.1 mm for each respective case within the distal stomach when the signal-to-noise ratio was 10 dB. Results indicate that penalized linear regression can successfully localize SW events provided the 3D geometry of the stomach and torso were acquired. Clinical Relevance- This method could help to improve the efficiency and accuracy of diagnosing gastric motility disorders from non-invasive measurements.

Miniaturized wireless gastric pacing via inductive power transfer with non-invasive monitoring using cutaneous Electrogastrography

BACKGROUND

Gastroparesis is a debilitating disease that is often refractory to pharmacotherapy. While gastric electrical stimulation has been studied as a potential treatment, current devices are limited by surgical complications and an incomplete understanding of the mechanism by which electrical stimulation affects physiology.

METHODS

A leadless inductively-powered pacemaker was implanted on the gastric serosa in an anesthetized pig. Wireless pacing was performed at transmitter-to-receiver distances up to 20 mm, frequency of 0.05 Hz, and pulse width of 400 ms. Electrogastrogram (EGG) recordings using cutaneous and serosal electrode arrays were analyzed to compute spectral and spatial statistical parameters associated with the slow wave.

RESULTS

Our data demonstrated evident change in EGG signal patterns upon initiation of pacing. A buffer period was noted before a pattern of entrainment appeared with consistent and low variability in slow wave direction. A spectral power increase in the EGG frequency band during entrainment also suggested that pacing increased strength of the slow wave.

CONCLUSION

Our preliminary in vivo study using wireless pacing and concurrent EGG recording established the foundations for a minimally invasive approach to understand and optimize the effect of pacing on gastric motor activity as a means to treat conditions of gastric dysmotility.

Electromyometrial imaging dataset of electromyograms and isochrone maps under deformation/electrical noise contaminations

The dataset presented in this paper is related to the recent work "Accuracy of electromyometrial imaging of uterine contractions in clinical environment" [1]. The dataset including body-uterus geometry obtained from magnetic resonance imaging (MRI), uterine electrograms and isochrone maps reconstructed using Electromyometrial imaging (EMMI) under various levels of deformations and electrical noise contamination in a translational sheep model are reported. The dataset make it possible for detailed evaluation and further improvement of EMMI. In addition, the researchers working on other types of electrophysiology imaging techniques, such as electrocardiographic imaging (ECGI), and Electrogastrography imaging (EGGI) could also adopt our method [1] and employ the dataset to evaluate and improve their imaging techniques.