Can MRI Be Used as a Sensor to Record Neural Activity?

Magnetic resonance provides exquisite anatomical images and functional MRI monitors physiological activity by recording blood oxygenation. This review attempts to answer the following question: Can MRI be used as a sensor to directly record neural behavior? It considers MRI sensing of electrical activity in the heart and in peripheral nerves before turning to the central topic: recording of brain activity. The primary hypothesis is that bioelectric current produced by a nerve or muscle creates a magnetic field that influences the magnetic resonance signal, although other mechanisms for detection are also considered. Recent studies have provided evidence that using MRI to sense neural activity is possible under ideal conditions. Whether it can be used routinely to provide functional information about brain processes in people remains an open question. The review concludes with a survey of artificial intelligence techniques that have been applied to functional MRI and may be appropriate for MRI sensing of neural activity.

Biomagnetism: The First Sixty Years.

Biomagnetism is the measurement of the weak magnetic fields produced by nerves and muscle. The magnetic field of the heart-the magnetocardiogram (MCG)-is the largest biomagnetic signal generated by the body and was the first measured. Magnetic fields have been detected from isolated tissue, such as a peripheral nerve or cardiac muscle, and these studies have provided insights into the fundamental properties of biomagnetism. The magnetic field of the brain-the magnetoencephalogram (MEG)-has generated much interest and has potential clinical applications to epilepsy, migraine, and psychiatric disorders. The biomagnetic inverse problem, calculating the electrical sources inside the brain from magnetic field recordings made outside the head, is difficult, but several techniques have been introduced to solve it. Traditionally, biomagnetic fields are recorded using superconducting quantum interference device (SQUID) magnetometers, but recently, new sensors have been developed that allow magnetic measurements without the cryogenic technology required for SQUIDs.

The mechanical bidomain model of cardiac muscle with curving fibers.

In the heart, cardiac muscle fibers curve creating zones of membrane forces resulting in regions of mechanotransduction. This study uses the finite difference method to solve the mechanical bidomain equations numerically for a complex fiber geometry. The magnitude of the active tension T is constant but its direction makes an angle with the x-axis that varies with position. Differences between the intracellular and extracellular displacements result from the bidomain behavior of the tissue that gives rise to forces on the integrin proteins in the membrane. The long-term goal is to use the mechanical bidomain model to suggest experiments and make predictions about growth and remodeling in the heart.

Heterogeneous anisotropic magnetic susceptibility of the myelin-water layers causes local magnetic field perturbations in axons.

One goal of MRI is to determine the myelin water fraction in neural tissue. One approach is to measure the reduction in T2 * arising from microscopic perturbations in the magnetic field caused by heterogeneities in the magnetic susceptibility of myelin. In this paper, analytic expressions for the induced magnetic field distribution are derived within and around an axon, assuming that the myelin susceptibility is anisotropic. Previous models considered the susceptibility to be piecewise continuous, whereas this model considers a sinusoidally varying susceptibility. Many conclusions are common in both models. When the magnetic field is applied perpendicular to the axon, the magnetic field in the intraaxonal space is uniformly perturbed, the magnetic field in the myelin sheath oscillates between the lipid and water layers, and the magnetic field in the extracellular space just outside the myelin sheath is heterogeneous. These field heterogeneities cause the spins to dephase, shortening T2 *. When the magnetic field is applied along the axon, the field is homogeneous within water-filled regions, including between lipid layers. Therefore the spins do not dephase and the magnetic susceptibility has no effect on T2 *. Generally, the response of an axon is given as the superposition of these two contributions. The sinusoidal model uses a different set of approximations compared with the piecewise model, so their common predictions indicate that the models are not too sensitive to the details of the myelin-water distribution. Other predictions, such as the sensitivity to water diffusion between myelin and water layers, may highlight differences between the two approaches. Copyright © 2016 John Wiley & Sons, Ltd.

Modeling bipolar stimulation of cardiac tissue.

Unipolar stimulation of cardiac tissue is often used in the design of cardiac pacemakers because of the low current required to depolarize the surrounding tissue at rest. However, the advantages of unipolar over bipolar stimulation are not obvious at shorter coupling intervals when the tissue near the pacing electrode is relatively refractory. Therefore, this paper analyzes bipolar stimulation of cardiac tissue. The strength-interval relationship for bipolar stimulation is calculated using the bidomain model and a recently developed parsimonious ionic current model. The strength-interval curves obtained using different electrode separations and arrangements (electrodes placed parallel to the fibers versus perpendicular to the fibers) indicate that bipolar stimulation results in more complex activation patterns compared to unipolar stimulation. An unusually low threshold stimulus current is observed when the electrodes are close to each other (a separation of 1 mm) because of break excitation. Unlike for unipolar stimulation, anode make excitation is not present during bipolar stimulation, and an abrupt switch from anode break to cathode make excitation can cause dramatic changes in threshold with very small changes in the interval. These results could impact the design of implantable pacemakers and defibrillators.

The Mechanism of Reflection Type Reentry: A Simulation Study.

INTRODUCTION: Reflection is a special type of reentry in which an electrical wave front travels in a forward direction through tissue that is then re-excited by a wave front that propagates backward. This type of reentry has been studied computationally in 1-dimensional fibers and verified experimentally. Different hypotheses explaining reflected reentry have been proposed based on the structure and heterogeneity of the tissue properties, but the mechanism remains uncertain. METHODS AND RESULTS: We used the bidomain model to represent cardiac tissue and the Luo-Rudy model to describe the active membrane properties. We consider an ischemic region in a volume of ventricular myocardium. Our results show that a slow depolarization in the ischemic border zone caused by electrotonic coupling to depolarized tissue in the normal region creates a delay between proximal and distal regions that produces enough electrotonic current in the distal region to re-excite the proximal region. CONCLUSION: Our simulation shows that an early afterdepolarization (EAD) is not the source of the reflection. It depends on the pacing interval and stimulus strength necessary to maintain enough time delay between proximal and distal regions.

Intracellular calcium and the mechanism of the dip in the anodal strength-interval curve in cardiac tissue.

BACKGROUND: The strength-interval (SI) curve is an important measure of refractoriness in cardiac tissue. The anodal SI curve contains a "dip" in which the S2 threshold increases with interval. Two explanations exist for this dip: (1) electrotonic interaction between regions of depolarization and hyperpolarization; and (2) the sodium-calcium exchange (NCX) current. The goal of this study is to use mathematical modeling to determine which explanation is correct. METHODS AND RESULTS: The bidomain model represents cardiac tissue and the Luo-Rudy model describes the active membrane. The SI curve is determined by applying a threshold stimulus at different time intervals after a previous action potential. During space-clamped and equal-anisotropy-ratios simulations, anodal excitation does not occur. During unequal-anisotropy-ratios simulations, electrotonic currents, not membrane currents, are present during the few milliseconds before excitation. The dip disappears with no NCX current, but is present with 50% or 75% reduction of it. The calcium-induced-calcium-release (CICR) current has little effect on the dip. CONCLUSIONS: These results indicate that neither the NCX nor the CICR current is responsible for the dip in the anodal SI curve. It is caused by the electrotonic interaction between regions of depolarization and hyperpolarization following the S2 stimulus.

The magnetocardiogram.

The magnetic field produced by the heart's electrical activity is called the magnetocardiogram (MCG). The first 20 years of MCG research established most of the concepts, instrumentation, and computational algorithms in the field. Additional insights into fundamental mechanisms of biomagnetism were gained by studying isolated hearts or even isolated pieces of cardiac tissue. Much effort has gone into calculating the MCG using computer models, including solving the inverse problem of deducing the bioelectric sources from biomagnetic measurements. Recently, most magnetocardiographic research has focused on clinical applications, driven in part by new technologies to measure weak biomagnetic fields.

Depth-Dependent Strain Model (1D) for Anisotropic Fibrils in Articular Cartilage.

The mechanical response of articular cartilage (AC) under compression is anisotropic and depth-dependent. AC is osmotically active, and its intrinsic osmotic swelling pressure is balanced by its collagen fibril network. This mechanism requires the collagen fibers to be under a state of tensile pre-strain. A simple mathematical model is used to explain the depth-dependent strain calculations observed in articular cartilage under 1D axial compression (perpendicular to the articular surface). The collagen fibers are under pre-strain, influenced by proteoglycan concentration (fixed charged density, FCD) and collagen stiffness against swelling stress. The stiffness is introduced in our model as an anisotropic modulus that varies with fibril orientation through tissue depth. The collagen fibers are stiffer to stretching parallel to their length than perpendicular to it; when combined with depth-varying FCD, the model successfully predicts how tissue strains decrease with depth during compression. In summary, this model highlights that the mechanical properties of cartilage depend not only on proteoglycan concentration but also on the intrinsic properties of the pre-strained collagen network. These properties are essential for the proper functioning of articular cartilage.

Associations Between Sodium-Glucose Co-transporter 2 Inhibitors and Urologic Diseases: Implications for Lower Urinary Tract Symptoms From a Multi-State Health System Analysis.

OBJECTIVE: To analyze the frequency of new urologic visits and urologic diagnoses in patients prescribed sodium-glucose co-transporter-2 inhibitors (SGLT-2is). MATERIAL AND METHODS: Records from a multi-state health system between 2014 and 2022 were reviewed to identify patients referred for outpatient urology evaluation within 2 years of diabetes medication prescription. Patients were stratified by the prescription of SGLT-2is or another diabetes medication. Frequency of urology visits within 1-year, urologic diagnoses, and prescriptions to treat lower urinary tract symptoms (LUTS) were compared. Patients were stratified by whether they had achieved HbA1c goal (≥7% or <7%) following treatment as well as by sex. Multivariable logistic regression was performed to determine if SGLT-2 use independently predicted outcomes of interest. RESULTS: 163,827 patients met inclusion criteria. Use of SGLT-2is was associated with a higher frequency of early urologic referral, balanitis/balanoposthitis, overactive bladder, urinary frequency, urgency, and need for LUTS medications in males with HbA1c ≥7%. Females on SGLT-2is with HbA1c ≥7% also had higher rates of urinary incontinence. In those with HbA1c <7%, only balanitis/balanoposthitis and urinary incontinence were higher in the SGLT-2i cohorts for males and females, respectively. Multivariable analysis found SGLT-2i use as predictive of early urology referral, balanitis/balanoposthitis, urinary urgency, frequency, overactive bladder, and need for LUTS medications in males. Multivariable analysis of females demonstrated similar results. CONCLUSION: SGLT-2is may lead to worse urologic outcomes and increased utilization of urologic care relative to other diabetic medications. Future studies are necessary to identify which patients are at highest risk of adverse urologic outcomes.

Transmembrane current imaging in the heart during pacing and fibrillation.

Recently, we described a method to quantify the time course of total transmembrane current (Im) and the relative role of its two components, a capacitive current (Ic) and a resistive current (Iion), corresponding to the cardiac action potential during stable propagation. That approach involved recording high-fidelity (200 kHz) transmembrane potential (Vm) signals with glass microelectrodes at one site using a spatiotemporal coordinate transformation via measured conduction velocity. Here we extend our method to compute these transmembrane currents during stable and unstable propagation from fluorescence signals of Vm at thousands of sites (3 kHz), thereby introducing transmembrane current imaging. In contrast to commonly used linear Laplacians of extracellular potential (Ve) to compute Im, we utilized nonlinear image processing to compute the required second spatial derivatives of Vm. We quantified the dynamic spatial patterns of current density of Im and Iion for both depolarization and repolarization during pacing (including nonplanar patterns) by calibrating data with the microelectrode signals. Compared to planar propagation, we found that the magnitude of Iion was significantly reduced at sites of wave collision during depolarization but not repolarization. Finally, we present uncalibrated dynamic patterns of Im during ventricular fibrillation and show that Im at singularity sites was monophasic and positive with a significant nonzero charge (Im integrated over 10 ms) in contrast with nonsingularity sites. Our approach should greatly enhance the understanding of the relative roles of functional (e.g., rate-dependent membrane dynamics and propagation patterns) and static spatial heterogeneities (e.g., spatial differences in tissue resistance) via recordings during normal and compromised propagation, including arrhythmias.

Boundary Layers and the Distribution of Membrane Forces Predicted by the Mechanical Bidomain Model.

The mechanical bidomain model is a mathematical description of the elastic properties of cardiac tissue. The unique feature of the bidomain model is that it is a macroscopic continuum representation of tissue that nevertheless accounts for the intracellular and extracellular spaces individually, thereby focusing on mechanical forces arising across the cell membrane. In this paper, the mechanical bidomain model describes a two-dimensional sheet of cardiac tissue undergoing a uniform active tension. At the boundary, the tissue sheet is free to move. Analytical solutions are found for the intracellular and extracellular displacements and pressures. The model predicts that membrane forces, which may be responsible for phenomena such as mechanotransduction and remodeling, are large near the tissue boundary, and fall off rapidly with distance from the boundary.

The Electric Field Induced by a Microcoil During Magnetic Stimulation.

OBJECTIVE: The purpose of this study is to calculate the electric field produced by an implanted microcoil during magnetic stimulation of the brain. METHODS: The electric field of a microcoil was calculated numerically. RESULTS: The maximum value of the induced electric field is approximately 0.000026 V/m for a 1 mA, 3 kHz current passed through the coil. CONCLUSION: This electric field value is too small to cause neural stimulation. SIGNIFICANCE: Previous studies reporting magnetic stimulation using a microcoil must have been exciting neurons by some other mechanism.

Bidomain modeling of electrical and mechanical properties of cardiac tissue.

Throughout the history of cardiac research, there has been a clear need to establish mathematical models to complement experimental studies. In an effort to create a more complete picture of cardiac phenomena, the bidomain model was established in the late 1970s to better understand pacing and defibrillation in the heart. This mathematical model has seen ongoing use in cardiac research, offering mechanistic insight that could not be obtained from experimental pursuits. Introduced from a historical perspective, the origins of the bidomain model are reviewed to provide a foundation for researchers new to the field and those conducting interdisciplinary research. The interplay of theory and experiment with the bidomain model is explored, and the contributions of this model to cardiac biophysics are critically evaluated. Also discussed is the mechanical bidomain model, which is employed to describe mechanotransduction. Current challenges and outstanding questions in the use of the bidomain model are addressed to give a forward-facing perspective of the model in future studies.

Measurements of transmembrane potential and magnetic field at the apex of the heart.

We studied the transmembrane potential and magnetic fields from electrical activity at the apex of the isolated rabbit heart experimentally using optical mapping and superconducting quantum interference device microscopy, and theoretically using monodomain and bidomain models. The cardiac apex has a complex spiral fiber architecture that plays an important role in the development and propagation of action currents during stimulation at the apex. This spiral fiber orientation contains both radial electric currents that contribute to the electrocardiogram and electrically silent circular currents that cannot be detected by the electrocardiogram but are detectable by their magnetic field, B(z). In our experiments, the transmembrane potential, V(m), was first measured optically and then B(z) was measured with a superconducting quantum interference device microscope. Based on a simple model of the spiral structure of the apex, V(m) was expected to exhibit circular wave front patterns and B(z) to reflect the circular component of the action currents. Although the circular V(m) wave fronts were detected, the B(z) maps were not as simple as expected. However, we observed a pattern consistent with a tilted axis for the apex spiral fiber geometry. We were able to simulate similar patterns in both a monodomain model of a tilted stack of rings of dipole current and a bidomain model of a tilted stack of spiraled cardiac tissue that was stimulated at the apex. The fact that the spatial pattern of the magnetic data was more complex than the simple circles observed for V(m) suggests that the magnetic data contain information that cannot be found electrically.

Mechanical model of neural tissue displacement during Lorentz effect imaging.

Allen Song and coworkers recently proposed a method for MRI detection of biocurrents in nerves called "Lorentz effect imaging." When exposed to a magnetic field, neural currents are subjected to a Lorentz force that moves the nerve. If the displacement is large enough, an artifact is predicted in the MR signal. In this work, the displacement of a nerve of radius a in a surrounding tissue of radius b and shear modulus mu is analyzed. The nerve carries a current density J and lies in a magnetic field B. The solution to the resulting elasticity problem indicates that the nerve moves a distance BJ/4mu a2ln(b/a). Using realistic parameters for a human median nerve in a 4T field, this calculated displacement is 0.013 microm or less. The distribution of displacement is widespread throughout the tissue, and is not localized near the nerve. This displacement is orders of magnitude too small to be detected by conventional MRI methods.