Division ofHealth AI

Vagus nerve stimulation (VNS) is emerging as potential treatment for several chronic diseases. However, limited control of fiber activation, e.g., to promote desired effects over side effects, restricts clinical translation. Towards that goal, we describe a VNS method consisting of intermittent, interferential sinusoidal current stimulation (i2CS) through multi-contact epineural cuffs. In experiments in anesthetized swine, i2CS elicits nerve potentials and organ responses, from lungs and laryngeal muscles, that are distinct from equivalent non-interferential sinusoidal stimulation. Resection and micro-CT imaging of a previously stimulated nerve, to resolve anatomical trajectories of nerve fascicles, demonstrate that i2CS responses are explained by activation of organ-specific fascicles rather than the entire nerve. Physiological responses in swine and activity of single fibers in anatomically realistic, physiologically validated biophysical vagus nerve models indicate that i2CS reduces fiber activation at the interference focus. Experimental and modeling results demonstrate that current steering and beat and repetition frequencies predictably shape the spatiotemporal pattern of fiber activation, allowing tunable and precise control of nerve and organ responses. When compared to equivalent sinusoidal stimulation in the same animals, i2CS produces reduced levels of a side-effect by larger laryngeal fibers, while attaining similar levels of a desired effect by smaller bronchopulmonary fibers.

Introduction: Obstructive sleep apnea (OSA) can cause severe complications if left untreated. Several challenges hinder OSA identification in females, resulting in underdiagnosis and undertreatment in this population. This study aimed to develop a machine learning (ML) approach specifically tailored to screen for moderate-to-severe OSA in women. Methods: A retrospective study using clinical records of 1210 women who underwent polysomnography at our institution was conducted. Collected data included demographics, body metrics, nocturnal oxygen saturation levels, medical conditions, medications, laboratory measurements, and polysomnography results. Four ML algorithms to classify participants into moderate-to-severe and none-to-mild OSA groups were employed. Results: Due to the high missingness of laboratory values in the whole cohort, two sets of models were developed: one that considered all subjects but excluded lab tests and another that only used a subgroup of 383 participants and additionally incorporated hemoglobin and lipid profile levels alongside the other features. Without laboratory measurements, the best-performing model was adaptive boosting, which achieved an area under the receiver operating characteristic curve and accuracy of 0.811 and 76.03%, respectively. When lab tests were included, gradient boosting machine outperformed its competitors, with the above metrics reaching 0.872 and 84.42%, respectively. Conclusion: The promising performance of our approach underlines the potential of artificial intelligence in refining screening strategies for OSA in women. Nadir oxygen saturation during sleep emerged as a particularly strong predictor, reinforcing the central role of nocturnal hypoxemia in OSA risk stratification. Future research should focus on incorporating broader clinical inputs and using larger, diverse datasets to deploy a highly accurate and robust model that meets clinical standards and is suitable for real-world implementation.