Value of computational modeling in understanding human pulmonary defensive reflexes

Speaker: Donald Bolser

Abstract:

Value of Computational Modeling in Understanding Human Pulmonary Defensive Reflexes.

Donald C. Bolser, PhD. Dept. of Physiological Sciences, University of Florida College of Veterinary Medicine. Gainesville, FL.

Airway defense is the prevention and/or correction of aspiration. During swallowing, aspiration is prevented by vocal fold adduction, changes in breathing, and appropriate movement of the laryngeal-epiglottic complex. If aspiration occurs, cough is elicited to produce high-velocity airflows to dislodge and eject materials. Most neuromuscular diseases result in cough (dystussia) and/or swallow (dysphagia) impairment. We have shown that objective analysis of coughing in patients with neuromuscular disease can have high sensitivity and specificity in predicting risk of aspiration. Currently, there are no multiscale computational models that are being used to facilitate clinical decisions regarding risk of aspiration. One of the goals of our work has been to develop a predictive model of the brainstem neural circuit that produces and controls coughing and swallowing behaviors. We propose that such a model will predict complex phenotypes of airway defensive behaviors in different patients, which could become part of a non-invasive diagnostic protocol that facilitates identification of patients at risk of aspiration. The functioning of this model circuit encompasses multiple scales, including the behavior of single neurons, cooperative activity of a neural circuit, regulation of multiple behaviors by that circuit, and the activity of respiratory muscles that drive pulmonary and chest wall mechanics. We have developed the first neuromechanical model of the respiratory control system. Our neuromechanical model does predict several temporal and mechanical aspects of single human coughs. One of the main obstacles to application of our neuromechanical model to human airway defensive behaviors is that the mechanics of human repetitive coughing differ from that of animal models. In particular, humans engage in repetitive laryngeal adduction and expulsion without intervening inspiratory efforts during coughing. Our attempts to simulate human repetitive coughing (which does not require an inspiratory effort for every cough expulsion) have led us to devolve the computational neural network model. Our simulations of a less complex version of the parent model support the existence of a dual oscillator-based neural network. Further, this devolved version of the computational model can produce motor patterns that are similar to phenotypes of repetitive coughing in patients with Parkinson’s Disease. Barriers to adoption of this model include: a) informing it with more specific data from animal models that address the dual-oscillator hypothesis, b) education of speech pathologists of the value of modeling in diagnostic protocols, and c) delivery of a clinician-centric platform that can be utilized in a hospital environment. Supported by HL 103415.

 

About the Speaker:

http://www.vetmed.ufl.edu/about-the-college/faculty-directory/don-bolser/

 

Return to the 2015 MSM Consortium Agenda

Event Date:

Wednesday, September 9, 2015

Table sorting checkbox
Off