Microwave sensing system for portable and wearable neurodegenerative disease monitoring

Abstract

Neurodegenerative Disease (ND) is an incurable and progressive condition characterized by the gradual degeneration and death of nerve cells. One of the most prevalent forms of ND is Alzheimer’s disease (AD), characterized by symptoms such as memory loss, confusion with time and place, and difficulties in speaking and writing, which severely disrupt daily life. With the ageing of the ’baby boom’ generation, the number of people affected by the disease is projected to rise dramatically in the next 40 years unless preventive measures are implemented. This drastic increase in people suffering from the disease has become one of the great global healthcare challenges affecting patients, families, caregivers, and society. Typical pathological changes associated with AD include cerebral atrophy and lateral ventricular enlargement, which start in the early stage of AD and deteriorate with time. Therefore, the extent of atrophy and lateral ventricular enlargement could be used to assess disease progression.

Imaging has played a key role in assessing and understanding the pathological changes of AD. However, existing modalities such as Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and Positron Emission Tomography (PET) are expensive, bulky, and require medical supervision, limiting their accessibility for long-term monitoring in care homes or clinic scenarios. Furthermore, CT scans involve ionizing radiation. MRI might not be suitable for patients with claustrophobia, and PET involves injecting radioactive liquid into the body. The limitations of existing modalities necessitate the need for alternative approaches that are convenient, effective, and non-invasive for ND monitoring.

Microwave sensing and Imaging (MSI) are emerging as a non-intrusive and noninvasive way of examining functional and pathological tissue conditions. MSI systems can be designed to be portable and wearable, which could potentially enable in-time diagnosis in acute or life-threatening situations. They can also facilitate long-term monitoring for non-hospital scenarios such as care homes and community clinics. In addition, the monitored data can be stored in clouds for long-term health tracking and disease diagnosis, allowing doctors and patients to access the health data more easily. This thesis presents microwave sensing systems that are designed to be portable and wearable for ND monitoring. The primary contributions include the design of wideband antennas, as well as the development of a fixed array configured and dynamic array configured prototype. Numerical modelling and experimental validation have been conducted to verify the effectiveness of the designed systems. The data collected is processed by a modified radar-based algorithm for brain abnormality localization.

In the first stage, customized wideband monopole antennas that adopted a rectangular topology were designed. The antennas were fabricated on rigid and flexible substrates, including FR-4 and polyimide materials. The performance of the antennas was first evaluated in terms of transmission coefficients, gain, and radiation pattern in free space and subsequently evaluated when placed near human head models.

In the second stage, two antenna array configurations were implemented, namely fixed array configuration and dynamic array configuration. The fixed array configuration involved the integration of RF switching circuits into an antenna array. Specifically, this fixed array configuration utilized an antenna pair, an RF switching circuit, a Bluetooth module, an Arduino microcontroller, and a mobile app for commanding the antenna elements. As a proof of concept, a One Pole Four Throw (1P4T) switching circuit was utilized to switch between two antennas placed near brain phantoms. The S-parameter data collection was then automated and remotely controlled using the microcontroller, the Bluetooth module, and the mobile app. In contrast, the dynamic array configuration involved a rotating platform and a flexible antenna to create a virtual antenna array. This configuration employed the Circular Synthetic Aperture Radar (CSAR) concept, creating a virtual antenna array by rotating the antenna around brain phantoms in a circular trajectory. Using the two configurations, S-parameter data were collected from a normal brain phantom and a brain phantom with lateral ventricle enlargement (LVE).

In the final stage, the efficacy and sensitivity of the two configurations were evaluated and compared. The data collected previously were processed with a radar-based Microwave Imaging via Space-Time (MIST) algorithm to create images indicating brain abnormalities. Both configurations proved the feasibility and effectiveness of detecting brain abnormalities associated with ND. The proposed wearable and portable systems could potentially be deployed in care homes or community clinics, providing a comfortable, efficient, and timely assessment for patients with ND.