Biomagnetic field sensor systems and methods for diagnostic evaluation of cardiac conditions
The present disclosure provides a biomagnetic field sensor system for diagnostic evaluation of a cardiac condition of an individual. The biomagnetic field sensor system may comprise an array of biomagnetic field sensors configured to sense an electromagnetic field associated with a heart of the individual and generate electromagnetic field data therefrom; a computer processor coupled to the array of biomagnetic field sensors; a memory configured to store the electromagnetic field data generated by the array of biomagnetic field sensors; and a non-transitory computer-readable medium encoded with a computer program including instructions that, when executed by the computer processor, cause the computer processor to receive the electromagnetic field data, and generate a diagnostic evaluation of a cardiac condition of the individual based at least in part on an analysis of the electromagnetic field data.
1 . A biomagnetic field sensor system for diagnostic evaluation of an individual, comprising:
an array of biomagnetic field sensors configured to acquire electromagnetic field (EMF) signals associated with a tissue, an organ, or a body part of the individual and generate magnetic field measurements therefrom;
a computer processor coupled to the array of biomagnetic field sensors; and
a memory coupled to the computer processor, the memory storing instructions that are configured to be executed by the computer processor, the instructions, when executed by the computer processor, cause the biomagnetic field sensor system to:
receive the EMF signals acquired by the array of biomagnetic field sensors;
filter the EMF signals to obtain filtered data;
execute a machine learning algorithm to process the filtered data, wherein the machine learning algorithm is trained to construct a hypothesis function according to training data comprising EMF measurements and corresponding abnormality data, to determine presence of one or more abnormalities in human subjects, the machine learning algorithm including a plurality of blocks of dilated convolutional neural networks, a respective block comprising a respective number of layers and a respective set of dilation rates, and wherein executing the machine learning algorithm includes:
constructing, by the machine learning model, the hypothesis function based on the filtered data;
determining, by the machine learning algorithm, whether an outcome of the hypothesis function is achieved; and
determining, by the machine learning algorithm based on the outcome of the hypothesis function, whether an abnormality is present in the individual; and
receive, as output via the machine learning algorithm, an analysis result indicating a presence or an absence of an abnormality of the tissue, the organ, or the body part of the individual, wherein the presence of the abnormality includes localization of the abnormality to a particular region of the tissue, organ, or body part.
2 . The biomagnetic field sensor system of claim 1 , further comprising a shield configured to shield at least a portion of a body of the individual from an environmental magnetic field.
3 . The biomagnetic field sensor system of claim 2 , wherein the shield is configured to at least partially enclose the at least the portion of the body of the individual, including at least a portion of a chest of the individual.
4 . The biomagnetic field sensor system of claim 2 , wherein the shield comprises a permalloy or a mumetal.
5 . The biomagnetic field sensor system of claim 1 , further comprising a movable base unit, and an arm having a proximal end and a distal end, the proximal end being coupled to the movable base unit by a first joint, the first joint configured so that the arm moves relative to the movable base unit with at least one degree of freedom.
6 . The biomagnetic field sensor system of claim 5 , wherein the arm comprises a proximal segment and a distal segment, and wherein a second joint is positioned between the proximal segment and the distal segment and is configured so that the distal segment articulates relative to the proximal segment.
7 . The biomagnetic field sensor system of claim 5 , wherein the array of biomagnetic field sensors is movably coupled to the distal end of the arm so that the array moves relative to the arm with at least one degree of freedom.
8 . The biomagnetic field sensor system of claim 1 , wherein the array of biomagnetic field sensors is arranged to match a generalized contour of a portion of a body of the individual.
9 . The biomagnetic field sensor system of claim 1 , wherein the instructions, when executed by the computer processor, cause the computer processor to further generate a visual representation of the magnetic field measurements comprising a waveform.
10 . The biomagnetic field sensor system of claim 9 , wherein the visual representation comprises a magnetocardiogram (MCG).
11 . The biomagnetic field sensor system of claim 9 , wherein the visual representation comprises an electric current map.
12 . The biomagnetic field sensor system of claim 1 , wherein the analysis result is indicative of a location of the abnormality, a type of the abnormality, or present conditions associated with the abnormality.
13 . The biomagnetic field sensor system of claim 1 , wherein the instructions, when executed by the computer processor, cause the computer processor to further localize an anatomical region of the body, organ, or tissue associated with the abnormality.
14 . The biomagnetic field sensor system of claim 1 , wherein the diagnostic evaluation comprises a diagnosis of the individual.
15 . The biomagnetic field sensor system of claim 1 , wherein the diagnostic evaluation comprises a prognosis of the individual.
16 . The biomagnetic field sensor system of claim 1 , wherein the array of biomagnetic field sensors comprises optically pumped magnetometer (OPM) sensors, magnetic induction sensors, magneto-resistive sensors, superconducting quantum interference device (SQUID) sensors, or a combination thereof.
17 . The biomagnetic field sensor system of claim 1 , wherein the array of biomagnetic field sensors is configured to be positioned outside of a body of the individual and not in contact with the body of the individual, in order to non-invasively acquire the EMF signals.
18 . The biomagnetic field sensor system of claim 1 , wherein the tissue, the organ, or the body part of the individual is a heart of the individual.
19 . The biomagnetic field sensor system of claim 18 , wherein the abnormality comprises ischemia or coronary artery disease (CAD).
20 . The biomagnetic field sensor system of claim 1 , wherein the tissue, the organ, or the body part of the individual is a brain of the individual.
21 . The biomagnetic field sensor system of claim 20 , wherein the abnormality is associated with stroke, traumatic brain injury, traumatic spine injury, encephalitis, meningitis, brain tumor, Alzheimer's disease, Parkinson's disease, ataxia, or a psychiatric disorder.
22 . The biomagnetic field sensor system of claim 1 , wherein the tissue, the organ, or the body part of the individual is a liver or a pancreas of the individual.
23 . The biomagnetic field sensor system of claim 1 , wherein the tissue, the organ, or the body part of the individual is a gastrointestinal tract of the individual.
24 . The biomagnetic field sensor system of claim 23 , wherein the abnormality is associated with a gastrointestinal cancer, Crolm's disease, ulcerative colitis, irritable bowel disease, dismotility disorder, gall stone, colitis, cholangitis, liver failure, pancreatitis, or a gastrointestinal infection.
25 . The biomagnetic field sensor system of claim 1 , further comprising executing, by the machine learning algorithm, automatic statistical analysis of the filtered data to determine which one or more features to extract and analyze.
26 . The biomagnetic field sensor system of claim 1 , wherein the machine learning algorithm is trained by repeatedly presenting the machine learning algorithm with EMF data together with abnormality data concerning an existence or non-existence of an abnormality in an organ, tissue, body, or portion thereof, the abnormality data further including information concerning an absence of an irregular health condition in a subject.
27 . The biomagnetic field sensor system of claim 1 , wherein the machine learning algorithm is trained on a single EMF measurement with accompanying abnormality data, the single EMF measurement having multiple, continuous segments of EMF signals.
28 . The biomagnetic field sensor system of claim 1 , wherein the instructions, when executed by the computer processor, cause the biomagnetic field sensor system to:
after receiving the EMF signals acquired by the array of biomagnetic field sensors, encode the received EMF signals to audio signals; and
filter the audio signals to obtain the filtered data.
29 . The biomagnetic field sensor system of claim 1 , wherein filtering the EMF signals to obtain filtered data includes:
converting the received EMF signals into one or more signals having a morphology of an audio signal wave; and
generating one or more coefficients for a short term power spectrum of a sound.