This dissertation introduces two fundamental aspects (hardware and software) of a novel bio-magnetic sensing system that transcends the current state-of-the-art specifications in terms of size, weight, cost, detection environment, signal bandwidth, sensitivity, and performance. In vitro and in vivo experiments were undertaken to validate the performance, repeatability, and accuracy. The clinical application of the system is explored, and preliminary results will inform the evolution of our concept.
Monitoring the magnetic fields that are naturally emanated by the human body can aid in both the diagnosis and personalized treatment of various health conditions related to the heart, the nervous system, and other physiological systems. Notably, magnetic fields overcome limitations associated with the electrode-based detection of bio-signals that are characterized by (1) low accuracy; (2) intrusiveness; (3) poor signal localization; and (4) associated discomfort. One central challenge associated with bio-magnetic sensing is the extremely low field strength that is weaker than the Earth’s magnetic field. Existing practices involve the use of devices such as Superconducting Quantum Interference Devices (SQUIDs) and Atomic Magnetometers (AM) to record these bio-magnetic signals. Unfortunately, the (1) high operational costs; (2) large space needed to install supplemental equipment, such as heaters, coolers, lasers, and shielding; (3) difficulties in fabrication, and (4) the bulkiness and heavy weight of the accompanied structures, make such devices undesirable for application outside of constrained environments (hospital, lab etc.). Although some other bio-magnetic sensors have been developed that require no shielding, they are bulky, heavy, cumbersome to use, and characterized by a poor detection sensitivity.
To overcome those limitations, a novel miniaturized, low-cost, and high-sensitivity sensing system capable of measuring bio-magnetic signals in non-shielded environments has been developed, which is the subject of this dissertation. The hardware is composed of four (4) primary components: (1) induction coil sensors; (2) amplifier boards; (3) an analog to digital converter (ADC); and (4) a computer for digital signal processing (DSP). The coil sensors capture and channel the bio-magnetic signal to the amplifier boards so that the extremely low bio-signal can be picked up by the ADC. The ADC then translates the analog signal into a digital signal which is later processed in a computer using DSP. The software component, viz DSP, is needed as the raw bio-magnetic signal is extremely weak and obscured under the noise floor: our novel approach filters the noise so that the real signal is thus exposed and seen. Four (4) DSP methods, namely, (1) notch filtering, (2) Ensemble Empirical Mode Decomposition (EEMD) filters, (3) bandpass filtering, and (4) averaging (multiple cycle averaging and multiple sensors averaging), are employed to reduce environmental and power line noise. Depending on the targeted signal and experimental conditions, not all of these methods may be eventually needed and/or not to the same extent.
In vitro and in vivo experiments were conducted to validate the performance of the sensing system. Results of the in vitro experiment confirmed the hypothesis that the sensing system can indeed detect bio-magnetic signals from an emulated single frequency bio- signal, cardiac signals, and wideband neuron action potentials in a non-shielded environment. Results of the in vivo experiments, advancing this concept, were conducted on healthy subjects and confirmed the feasibility of our system to detect real-time MCG in earth ambient noise with adequate RF interference. Using the heart rate variability (HRV) index calculated from real-time MCG, to quantify cognitive workload was an example application that was pursued in this research.
Future research with gathered preliminary results based upon this concept entails (1) detect different types of bio-signals with various coil array configurations, (2) establish guidelines for the translation and integration of this technology into everyday e-textile clothing products, (3) achieve lower detection level and higher sensitivity with the additional noise coil subtraction methods, (4) extend the detection potential of the senor to even weaker signals in real life and with shorter averages in a clinical environment with partially shielded environment.
Overall, this technology establishes new opportunities in bio-magnetic signal sensing and medical diagnosis in primarily cardiac and neural applications with the ultimate goal of making the sensor portable for integration into wearable garments that can detect bio-magnetic signals dynamically.