International Workshop on Magnetic Particle Imaging

Emerging Applications of Color Magnetic Particle Imaging

Color Magnetic Particle Imaging (MPI) is a rapidly developing area within MPI, promising numerous important applications such as identifying the properties of the MNP environment and catheter tracking during cardiovascular interventions. Different MNPs are expected to induce different MPI signals based on factors such as their magnetic material and core diameter. Furthermore, the relaxation behaviors of MNPs are affected by the differences in their environmental conditions, such as temperature or viscosity. This talk will discuss recent techniques in color MPI that leverage these differences to distinguish different MNP types and/or their environmental conditions. Specifically, the presentation will concentrate on calibration-free x-space-based relaxation mapping for color MPI.

Control and Detection Strategies for Magnetic Microrobots Inspired by Magnetic Particle Imaging

In addition to underpinning several minimally invasive biomedical imaging modalities, magnetic stimuli offer a compelling means to control and power medical microrobots deployed within the body. The variety of forms that these magnetically responsive microrobots can take is continuously expanding, ranging from fabricated structures designed to efficiently locomote in response to applied magnetic fields, to naturally magnetic organisms repurposed as biohybrid microrobots. In the past, the magnetic stimuli most frequently applied to these microrobots have been directable uniform magnetic fields that can steer their intrinsic motion or gradient fields that apply forces to pull them. However, these methods entail serious practical challenges, since control often requires a complimentary form of live imaging, and the feasibility of producing adequate gradients in patients diminishes as the microrobots shrink in size. Rotating magnetic fields (RMFs), in which magnitude remains constant while direction is swept around one or more planes of rotation, offer a promising alternative for actuation. Not only do RMFs deliver mechanical energy through magnetic torque-based actuation schemes that more readily scale to patients, but our work has also shown that concepts inspired by magnetic particle imaging present unique possibilities for control and feedback when adapted to RMFs. Actuating microrobots with RMFs permits simultaneous actuation and inductive detection, which we demonstrate with setups adapted for low frequencies (Hz to 10s of Hz) and physical phase cancellation. Moreover, we have shown that a superimposed magnetostatic selection field can spatially restrict torque-based actuation to a single point. By combining these ideas, and employing signal processing techniques focused on phase decomposition, we show that spatially selective inductive signal acquisition is possible in a gating field and that the resolution is set by the relative magnitude of the magnetostatic field and RMF. These principles build toward improved methods for drug targeting with live feedback and show how concepts from MPI can be fruitfully extended to magnetic microrobotic control.

Strategies and tips for the development of prototype instrumentation in an academic lab

Machine Learning for MPI Reconstruction

Magnetic particle imaging (MPI) promises an unparalleled combination of contrast and resolution for tracing magnetic nanoparticles. Yet, formation of images from acquired data is a heavily ill-posed problem given limits on the imaging speed and signal-to-noise ratio efficiency in MPI. Classical approaches to reconstruction of imaging data rely on hand-constructed priors that can fail to address these limitations effectively. In this talk, I will share an overview of recent efforts on devising deep learning techniques that instead adopt data-driven priors to surpass fundamental barriers. I will showcase neural network architectures and learning strategies that empower performance leaps in system matrix recovery, image reconstruction and processing.

Particle magnetization models in MPI

Finding sufficiently accurate models for the concentration-to-voltage mapping in MPI is still a challenging problem. One crucial aspect is the magnetization behavior of magnetic nanoparticles (MNPs) in the applied magnetic field. After a short general introduction to modeling aspects of magnetic particle imaging, physical models for the dynamic behavior of the MNP's magnetic moment (Brownian/Néel rotation) and their influence on the MPI signal are considered in this tutorial. The tutorial further provides an introduction to simulation techniques for solving the Fokker-Planck equation resulting from the stochastic ODEs of Brownian and Néel rotation of the particle's magnetic moment. 

How to Write a Successful S10 NIH Shared Instrumentation Grant: Sharing is Caring

Not everyone is lucky enough to have extra funds available for purchasing a new MPI machine. An alternative option is to submit a shared instrumentation grant (SIG) grant application to NIH, capped at $600,000.- (low-end instrumentation) or $2,000,000.-  (high-end instrumentation). A successful application needs to address a proper justification of need, identify about 10 NIH-funded major users, describe the technical expertise that exists to operate the machine, provide an effective management/administration plan, a detailed siting/housing plan with or without major renovation of existing space, and a financial/business plan for the first 5 years including institutional commitment. Examples of a successful application will be shown from the Kennedy Krieger Institute together with Johns Hopkins University, along with an outline of things to say and not to say.