Biomedical engineering

Deep learning strategies combined with wearable sensors have advanced the capabilities of monitoring systems in biomedical applications, offering precise and efficient solutions for diagnosing and managing diseases. However, applying these systems faces several challenges. One of the challenges is the diminishing performance when these systems encounter new data with more complex patterns than those seen before. Another challenge is the limited availability of labeled data, on which deep learning-based systems depend highly. Additionally, obtaining high-quality labeled data to train deep learning models is often expensive, requiring significant time and resources. Another significant challenge is ensuring the practicality, accessibility, and convenience of the monitoring systems.
This dissertation proposes an innovative deep learning framework to overcome these challenges and improve system generalization performance in classification and regression tasks, specifically monitoring patients with neurological disorders like Parkinson’s.

Malaria is an ancient lethargic disease that remains a global burden. It has been difficult to end the scourge of P. falciparum malaria because of the parasites’ drug resistance so early diagnosis of malaria is crucial. Microscopy remains the gold standard but has limited reliability in detecting malaria parasites. This study proffered a method towards detection of low parasitemia P. falciparum infected RBCs (Pf-RBCs) based on dielectrophoresis (DEP). A microfluidic device was designed for label-free cell sorting of Pf-RBCs from other whole blood in a continuous manner, based on the intrinsic electrical signatures of the cells. The design was validated by a finite element simulation using COMSOL Multiphysics. Simulations show the feasibility of the separation in a 9-mm long microfluidic channel under laminar flow conditions, using a low voltage supply of +/-10 V at 50 kHz.

Cerebrospinal fluid (CSF) has a role, in keeping the brain and spinal cord safe and nourished within the nervous system (CNS). This clear and colorless fluid is produced in the ventricles of the brain. Surrounds these structures acting as a protective cushion. CSF plays a role in maintaining nervous system health and ensuring optimal functioning. CSF accomplishes four objectives.
Protection: The brain and spinal cord are shielded from harm due to CSFs natural shock absorbing properties. This effectively safeguards these structures, from injuries caused by impacts or collisions.
Nutrition It ensures a favorable environment for neural cells to perform at their peak by supplying essential nutrients and removing waste products from the brain and spinal cord.

Time-series data in biomedical applications are gaining an increased interest to detect and predict underlying diseases and estimate their severity, such as Parkinson’s disease (PD) and cardiovascular diseases. This interest is driven by advances in wearable sensors and deep learning models to a large extent. In the literature, less attention has been paid to regression models for continuous outcomes in these applications, especially when dealing with limited data. Training deep learning models on raw limited data results in overfitted models, which is the main technical challenge we address in this dissertation. An example of limited and\or imbalanced time-series data is PD’s motion signals that are needed for the continuous severity estimation of Parkinson’s disease (PD). The significance of this continuous estimation is providing a tool for longitudinal monitoring of daily motor and non-motor fluctuations and managing PD medications.
The dissertation objective is to train generalizable deep learning models for biomedical regression problems when dealing with limited training time-series data. The goal is designing, developing, and validating an automatic assessment system based on wearable sensors that can measure the severity of PD complications in the home-living environment while patients with PD perform their activities of daily living (ADL). We first propose using a combination of domain-specific feature engineering, transfer learning, and an ensemble of multiple modalities. Second, we utilize generative adversarial networks (GAN) and propose a new formulation of conditional GAN (cGAN) as a generative model for regression to handle an imbalanced training dataset. Next, we propose a dual-channel auxiliary regressor GAN (AR-GAN) trained using Wasserstein-MSE-correlation loss. The proposed AR-GAN is used as a data augmentation method in regression problems.

Mass transport is important for all biological functions to protect the cell’s environment and to keep its balance of nutrients, proteins and keep the organism alive. We are motivated to study two different types of mass transport, glucose and oxygen that are critical in human system. Specifically, this study focused on mass and oxygen transport in human placenta and oxygen transport in transfusion of artificial oxygen carriers. Studying these processes in vivo or ex vivo are difficult due to ethical or technical challenges.
In this dissertation, Organ-on-a-chip devices were used to simulate placental barrier and blood vessels. In first device, 3D placenta–on-a-chip device consists of a polycarbonate membrane and two Poly dimethylsiloxane microchannels was used. Human umbilical vein endothelial cells were cultured in microfluidic devices and mass transport was measured. In the second device, 3-lane OrganoPlate was used to develop the placental barrier model. The human umbilical vein endothelial cells and trophoblast cells cultured in two microchannels compartmented by polycarbonate membrane (first device) and extracellular matrix gel (second device) to mimic the placental barrier in vitro. Finally, the glucose transfer across the placental barrier affected by malaria parasite was investigated. The results of this study can be used for better understanding of placental malaria pathology and drug efficacy testing.

Chronic Diseases are the major cause of mortality around the world, accounting for 7 out of 10 deaths each year in the United States. Because of its adverse effect on the quality of life, it has become a major problem globally. Health care costs involved in managing these diseases are also very high. In this thesis, we will focus on two major chronic diseases Asthma and Diabetes, which are among the leading causes of mortality around the globe. It involves design and development of a predictive analytics based decision support system which uses five supervised machine learning algorithm to predict the occurrence of Asthma and Diabetes. This system helps in controlling the disease well in advance by selecting its best indicators and providing necessary feedback.

Fidgeting and otherwise constant movements in individuals can be beneficial in those who suffer from Attention Deficit/Hyperactivity Disorder or Generalized Anxiety Disorder as well as others. However this constant movement can also be a distraction to others as well as protrude an air of no self confidence. It is the drawbacks from these actions that we wish to address. By developing an intelligent system that can detect these motions and alert the user, it will allow the wearer of the device to self correct. It is in this self control that one may exhibit more confidence or simply reduce the level of irritation experienced by those in the immediate vicinity. We have designed and built a low cost, mobile, lightweight, untethered, wearable prototype device. It will detect these actions and deliver user selectable biofeedback through a light emitting diode, buzzer, vibromotor or an electric shock to allow for self control.