Elif S. Bayrak

Multivariable adaptive closed-loop control of an artificial pancreas without meal and activity announcement.

BACKGROUND Accurate closed-loop control is essential for developing artificial pancreas (AP) systems that adjust insulin infusion rates from insulin pumps. Glucose concentration information from continuous glucose monitoring (CGM) systems is the most important information for the control system. Additional physiological measurements can provide valuable information that can enhance the accuracy of the control system. Proportional-integral-derivative control and model predictive control have been popular in AP development. Their implementations to date rely on meal announcements (e.g., bolus insulin dose based on insulin:carbohydrate ratios) by the user. Adaptive control techniques provide a powerful alternative that do not necessitate any meal or activity announcements. MATERIALS AND METHODS Adaptive control systems based on the generalized predictive control framework are developed by extending the recursive modeling techniques. Physiological signals such as energy expenditure and galvanic skin response are used along with glucose measurements to generate a multiple-input-single-output model for predicting future glucose concentrations used by the controller. Insulin-on-board (IOB) is also estimated and used in control decisions. The controllers were tested with clinical studies that include seven cases with three different patients with type 1 diabetes for 32 or 60 h without any meal or activity announcements. RESULTS The adaptive control system kept glucose concentration in the normal preprandial and postprandial range (70-180 mg/dL) without any meal or activity announcements during the test period. After IOB estimation was added to the control system, mild hypoglycemic episodes were observed only in one of the four experiments. This was reflected in a plasma glucose value of 56 mg/dL (YSI 2300 STAT; Yellow Springs Instrument, Yellow Springs, OH) and a CGM value of 63 mg/dL). CONCLUSIONS Regulation of blood glucose concentration with an AP using adaptive control techniques was successful in clinical studies, even without any meal and physical activity announcement.

Three-dimensional modeling of angiogenesis in porous biomaterial scaffolds

Vascularization of biomaterial scaffolds is essential for the successful clinical application of engineered tissues. Experimental studies are often performed to investigate the role of scaffold architecture on vascularized tissue formation. However, experiments are expensive and time-consuming and synthesis protocols often do not allow for independent investigation of specific scaffold properties. Computational models allow for rapid screening of potential material designs with control over scaffold properties that is difficult in laboratory settings. We have developed and tested a three-dimensional agent-based framework for investigating the effect of scaffold pore architecture on angiogenesis. Software agents represent endothelial cells, interacting together and with their micro-environment, leading to the invasion of blood vessels into the scaffold. A rule base, driven by experimental findings, governs the behavior of individual agents. 3D scaffold models with well-defined homogeneous and heterogeneous pore architectures were simulated to investigate the impact of various design parameters. Simulation results indicate that pores of larger size with higher interconnectivity and porosity support rapid and extensive angiogenesis. The developed framework can be used to screen biomaterial scaffold designs for optimal vascularization and investigate complex interactions among invading blood vessels and their micro-environment.

Hypoglycemia Early Alarm Systems Based On Multivariable Models.

Hypoglycemia is a major challenge of artificial pancreas systems and a source of concern for potential users and parents of young children with Type 1 diabetes (T1D). Early alarms to warn the potential of hypoglycemia are essential and should provide enough time to take action to avoid hypoglycemia. Many alarm systems proposed in the literature are based on interpretation of recent trends in glucose values. In the present study, subject-specific recursive linear time series models are introduced as a better alternative to capture glucose variations and predict future blood glucose concentrations. These models are then used in hypoglycemia early alarm systems that notify patients to take action to prevent hypoglycemia before it happens. The models developed and the hypoglycemia alarm system are tested retrospectively using T1D subject data. A Savitzky-Golay filter and a Kalman filter are used to reduce noise in patient data. The hypoglycemia alarm algorithm is developed by using predictions of future glucose concentrations from recursive models. The modeling algorithm enables the dynamic adaptation of models to inter-/intra-subject variation and glycemic disturbances and provides satisfactory glucose concentration prediction with relatively small error. The alarm systems demonstrate good performance in prediction of hypoglycemia and ultimately in prevention of its occurrence.

Hypoglycemia Early Alarm Systems Based on Recursive Autoregressive Partial Least Squares Models

Background: Hypoglycemia caused by intensive insulin therapy is a major challenge for artificial pancreas systems. Early detection and prevention of potential hypoglycemia are essential for the acceptance of fully automated artificial pancreas systems. Many of the proposed alarm systems are based on interpretation of recent values or trends in glucose values. In the present study, subject-specific linear models are introduced to capture glucose variations and predict future blood glucose concentrations. These models can be used in early alarm systems of potential hypoglycemia. Method: A recursive autoregressive partial least squares (RARPLS) algorithm is used to model the continuous glucose monitoring sensor data and predict future glucose concentrations for use in hypoglycemia alarm systems. The partial least squares models constructed are updated recursively at each sampling step with a moving window. An early hypoglycemia alarm algorithm using these models is proposed and evaluated. Results: Glucose prediction models based on real-time filtered data has a root mean squared error of 7.79 and a sum of squares of glucose prediction error of 7.35% for six-step-ahead (30 min) glucose predictions. The early alarm systems based on RARPLS shows good performance. A sensitivity of 86% and a false alarm rate of 0.42 false positive/day are obtained for the early alarm system based on six-step-ahead predicted glucose values with an average early detection time of 25.25 min. Conclusions: The RARPLS models developed provide satisfactory glucose prediction with relatively smaller error than other proposed algorithms and are good candidates to forecast and warn about potential hypoglycemia unless preventive action is taken far in advance.

Pore Interconnectivity Influences Growth Factor-Mediated Vascularization in Sphere-Templated Hydrogels.

Rapid and controlled vascularization within biomaterials is essential for many applications in regenerative medicine. The extent of vascularization is influenced by a number of factors, including scaffold architecture. While properties such as pore size and total porosity have been studied extensively, the importance of controlling the interconnectivity of pores has received less attention. A sintering method was used to generate hydrogel scaffolds with controlled pore interconnectivity. Poly(methyl methacrylate) microspheres were used as a sacrificial agent to generate porous poly(ethylene glycol) diacrylate hydrogels with interconnectivity varying based on microsphere sintering conditions. Interconnectivity levels increased with sintering time and temperature with resultant hydrogel structure showing agreement with template structure. Porous hydrogels with a narrow pore size distribution (130-150 μm) and varying interconnectivity were investigated for their ability to influence vascularization in response to gradients of platelet-derived growth factor-BB (PDGF-BB). A rodent subcutaneous model was used to evaluate vascularized tissue formation in the hydrogels in vivo. Vascularized tissue invasion varied with interconnectivity. At week 3, higher interconnectivity hydrogels had completely vascularized with twice as much invasion. Interconnectivity also influenced PDGF-BB transport within the scaffolds. An agent-based model was used to explore the relative roles of steric and transport effects on the observed results. In conclusion, a technique for the preparation of hydrogels with controlled pore interconnectivity has been developed and evaluated. This method has been used to show that pore interconnectivity can independently influence vascularization of biomaterials.

Agent-based modeling of osteogenic differentiation of mesenchymal stem cells in porous biomaterials.

Mesenchymal stem cells (MSC) have shown promise in tissue engineering applications due to their potential for differentiating into mesenchymal tissues such as osteocytes, chondrocytes, and adipocytes and releasing proteins to promote tissue regeneration. One application involves seeding MSCs in biomaterial scaffolds to promote osteogenesis in the repair of bone defects following implantation. However, predicting in vivo survival and differentiation of MSCs in biomaterials is challenging. Rapid and stable vascularization of scaffolds is required to supply nutrients and oxygen that MSCs need to survive as well as to go through osteogenic differentiation. The objective of this study is to develop an agent-based model and simulator that can be used to investigate the effects of using gradient growth factors on survival and differentiation of MSCs seeded in scaffolds. An agent-based model is developed to simulate the MSC behavior. The effect of vascular endothelial growth factor (VEGF) and bone morphogenic protein-2 (BMP-2) on both survival and osteogenic differentiation is studied. Results showed that the survival ratio of MSCs can be enhanced by increasing VEGF concentration. BMP-2 caused a slight increase on survival ratio. Osteogenesis strongly depends on the VEGF concentration as well because of its effect on vascularization. BMP-2 increased the osteogenic differentiation of MSCs.

Adaptive multivariable closed-loop control of blood glucose concentration in patients with Type 1 Diabetes

Closed-loop control of blood glucose concentration is an alternative treatment for Diabetes patients by automatically controlling their blood glucose level. Recursive time-series models are found to be powerful tool for prediction and control of glucose concentration. However, the stability of such models cannot be guaranteed by regular identification methods. In this paper, we use a constrained optimization method for stability of models. Generalized predictive controller (GPC) is used to close the loop. Physiological signals are used and a multivariable model is created that also includes insulin on-board information to prevent hypoglycemia events. Seven experiments with 3 different patients are done. Results show GPC with constrained identification method and physiological signals can regulate blood glucose concentration successfully.

Agent-based modeling of porous scaffold degradation and vascularization: Optimal scaffold design based on architecture and degradation dynamics

UNLABELLED A multi-layer agent-based model (ABM) of biomaterial scaffold vascularization is extended to consider the effects of scaffold degradation kinetics on blood vessel formation. A degradation model describing the bulk disintegration of porous hydrogels is incorporated into the ABM. The combined degradation-angiogenesis model is used to investigate growing blood vessel networks in the presence of a degradable scaffold structure. Simulation results indicate that higher porosity, larger mean pore size, and rapid degradation allow faster vascularization when not considering the structural support of the scaffold. However, premature loss of structural support results in failure for the material. A strategy using multi-layer scaffold with different degradation rates in each layer was investigated as a way to address this issue. Vascularization was improved with the multi-layered scaffold model compared to the single-layer model. The ABM developed provides insight into the characteristics that influence the selection of optimal geometric parameters and degradation behavior of scaffolds, and enables easy refinement of the model as new knowledge about the underlying biological phenomena becomes available. STATEMENT OF SIGNIFICANCE This paper proposes a multi-layer agent-based model (ABM) of biomaterial scaffold vascularization integrated with a structural-kinetic model describing bulk degradation of porous hydrogels to consider the effects of scaffold degradation kinetics on blood vessel formation. This enables the assessment of scaffold characteristics and in particular the disintegration characteristics of the scaffold on angiogenesis. Simulation results indicate that higher porosity, larger mean pore size, and rapid degradation allow faster vascularization when not considering the structural support of the scaffold. However, premature loss of structural support by scaffold disintegration results in failure of the material and disruption of angiogenesis. A strategy using multi-layer scaffold with different degradation rates in each layer was investigated as away to address this issue. Vascularization was improved with the multi-layered scaffold model compared to the single-layer model. The ABM developed provides insight into the characteristics that influence the selection of optimal geometric and degradation characteristics of tissue engineering scaffolds.

Simulating 3-D bone tissue growth using repast HPC: initial simulation design and performance results

Bone is one of the most implanted tissues worldwide. Bone tissue engineering deals with the replacement and regeneration of bone tissue; outcomes are determined by complex biological interactions, making it difficult to design an optimal tissue growth environment. Agent-Based Modeling (ABM) is a powerful tool to simulate such a system. We present a simulation of engineered bone tissue growth using the Repast HPC toolkit, an ABM tool for high-performance computing environments. We use this example to provide preliminary performance tests on new features (not yet publicly released) of Repast HPC that accommodate operations common to biological modeling: 3-Dimensional parallelized spatial simulation and diffusion in 3 dimensions. Repast HPC is a general ABM toolkit, and the performance documented here should be representative of performance on other simulations. Using the baseline Repast HPC tools provides flexibility for continued model development and improvement.

Active inference for dynamic Bayesian networks with an application to tissue engineering

In temporal domains, agents need to actively gather information to make more informed decisions about both the present and the future. When such a domain is modeled as a temporal graphical model, what the agent observes can be incorporated into the model by setting the respective random variables as evidence. Motivated by a tissue engineering application where the experimenter needs to decide how early a laboratory experiment can be stopped so that its possible future outcomes can be predicted within an acceptable uncertainty, we first present a dynamic Bayesian network (DBN) model of vascularization in engineered tissues and compare it with both real-world experimental data and agent-based simulations. We then formulate the question of “how early an experiment can be stopped to guarantee an acceptable uncertainty about the final expected outcome” as an active inference problem for DBNs and empirically and analytically evaluate several search algorithms that aim to find the ideal time to stop a tissue engineering laboratory experiment.