Improvement of multisource localization of magnetic particles in an animal

Chin-Wei Lin,Chia-Hao Su,Shu-Hsien Liao,Li-Min Wang,Han-Sheng Huang,Kuen-Lin Chen,Jyh-Horng Chen

doi:10.1038/s41598-021-88847-8

Abstract

In this simulation work, the linearized Bregman iterative algorithm was applied to solve the magnetic source distribution problem of a magnetic particle imaging (MPI) system for small animals. MPI system can apply an excitation magnetic field, and the induced magnetic field from the magnetic nanoparticles (MNPs) can be detected by the sensors of MPI system. With a gaussian distribution source at the upper side of the mouse brain, sensors set above the mouse brain and the constant excitation magnetic field, the average deviation of the calculated source distribution from the multiplane scanning along the axis away from the mouse brain and the closest plane scanning are 2.78 × 10–3 and 2.84 × 10–3 respectively. The simulated result showed that combination of multiplane scanning hardly improves the accuracy of the source localization. In addition, a gradient scan method was developed that uses gradient magnetic field to scan the mouse brain. The position of the maximum of the lead field matrix will be controlled by the gradient field. With a set up gaussian distribution source at the bottom of the mouse brain, the average deviation of the calculated source distribution from the gradient scan method and the constant field are 4.42 × 10–2 and 5.05 × 10–2. The location error from the two method are 2.24 × 10–1 cm and 3.61 × 10–1 cm. The simulation showed that this method can improve the accuracy compared to constant field when the source is away from the sensor and having a potential for application.

Highlights

Term is being applied to make the ill-posed problem solvable
When the sensors moved away from the position of 2.5–6.0 cm, the location error increased from 0 to 2.449 × 10−1 and the average deviation increased from 2.78 × 1 0−3 to 5.88 × 10−3
The combination of all the scanning planes with different Z positions, a multiple scan, had almost the same inverse source distribution with the closest plane scanning at 2.5 cm

Summary

Introduction

Term is being applied to make the ill-posed problem solvable. Constraints according to our experimental condition are applied to help in obtaining the most possible s olution[3]. The localization accuracy becomes worse when the magnetic source is far from the sensors. We simulated the detection of the MPI system. A “gradient scan method” was established to improve the accuracy of the calculated magnetic particle distribution, especially for the “deeper sources,” was developed

Methods

Results

Conclusion