S.N. Erne

Biomagnetic systems for clinical use

Abstract We present two multichannel systems based on a superconducting quantum interference device (SQUID) for biomagnetic measurements, installed at the University of Chieti. Both systems have been designed for clinical and routine use and have been developed owing to an international cooperation. The main issues in the instrument implementation were field sensitivity and spatial resolution, as well as flexibility and stability during operation. The first system is a planar system and is devised for magnetocardiographic measurements. This system is composed of 74 dc SQUID integrated magnetometers contained in a low-noise dewar: 55 sensors are measurement channels and 21 are placed far from the subject and are used as reference channels to create software gradiometers. The second system is a helmet system and consists of 165 dc SQUID integrated magnetometers to perform magnetoencephalographic recordings; 153 channels are distributed over a surface covering the whole scalp and 12 channels are used as references. The field noise of the SQUID magnetometers is about 5fTHz−1/2. Each system is placed in a magnetically shielded room for eddy current shielding and magnetic shielding. The magnetic field is recorded with sampling frequencies up to 10 kHz. The analogue-to-digital converted data are processed on line by means of an array of digital signal processors, allowing bandpass filtering, decimation and noise compensation.

Magnetocardiography and exercise testing

Twenty healthy male subjects (age range, 15-25 years; median, 21 years) underwent magnetocardiography during physical exercise. Significant ST-segment displacements of the magnetic signal were found during exercise at a heart rate of 120 beats/min compared to the magnetic signal at rest (P < .001). Since no significant ST-segment changes were found in the electrocardiogram recorded simultaneously with the magnetocardiogram, it is concluded that the magnetocardiogram shows junctional ST-T segment changes earlier than the electrocardiogram.

The study of steady magnetic fields associated with primary and secondary ST shift in ischaemic rabbit hearts

The study of injury potentials associated with DC currents that generate the primary or secondary ST shifts during cardiac ischaemia is possible only through the invasive technique of the DC electrogram. Clinical surface ECG recordings are AC coupled and cannot be used. This paper reports the use of non-invasive and unshielded magnetocardiographic measurements to evaluate the DC injury currents associated with ST shifts during coronary artery occlusions in the isolated rabbit heart. The effect on the magnetic ST shift is studied under different ischaemic conditions including regional ischaemia, global ischaemia, global ischaemia following long periods of regional ischaemia, regional ischaemia after repeated episodes of reversible global ischaemia, and bilateral regional ischaemia. Recording of DC magnetic fields allows the characterization of primary and secondary ST displacement for each induced ischaemic condition. Our measurements show that the ST shift starts earlier when inducing ischaemia in hearts previously subjected to ischaemic episodes than in hearts where the ischaemia was produced for the first time.

Multiple Modality Biomagnetic Analysis System

As an emerging new modality, biomagnetic analysis presents exciting opportunities while raising some difficult issues. During the past four years researchers at General Electric, The University of Ulm and The University of Chieti have created a state of the art biomagnetic analysis system, and have addressed many of these difficulties. One of the critical aspects of a biomagnetic analysis system is how it handles different types of data. Many of the clinical requirements for a successful system require fast, efficient and flexible data management. This paper describes what those requirements are and how we addressed them. We start by providing a high level overview of our system. We then list some of the pertinent clinical requirements, present our solutions and describe how well they have worked. Finally we discuss future directions.

Magnetocardiography and exercise testing: data acquisition and data processing

Sixteen healthy male subjects underwent magnetocardiography during physical exercise. A specific signal averaging technique was developed because the quality of the magnetic signals exceeded that of the electrocardiographic signals, particularly regarding baseline shifting. Significant ST segment displacements of the magnetic signal were found during exercise at a heart rate of 120 bpm (p<0.001). No significant ST segment changes were found in the electrocardiographic map, which consisted of an array of 16 precordial electrodes, simultaneously recorded. The hypothesis that junctional ST segment displacements can be detected earlier with the magnetocardiogram is discussed.<<ETX>>

Template Analysis on Interictal Neuromagnetic Data from Cases of Focal and Generalized Epilepsy

The investigation of partial (focal) epilepsy by neuromagnetic measurements has provided significant results on the localization of epileptogenic foci related to intercritical activity (Barth et al., 1982; Chapman et al., 1983; Barth et al., 1984; Ricci et al., 1985). It has been demonstrated, however, that quite often the morphology of epileptic signals is not unique (Rose et al., 1987). This may be interpreted in terms of various neural groups simultaneously active but underlying different firing process. If we do not consider quasi-rhythmic activity, for which the best analytical approach remains the Relative Covariance Method (RCM) (Chapman et al., 1983), it would be particularly useful to identify a procedure that permits to reliably select epileptic signals (spikes, sharp waves, spike-and-wave complexes), and to eventually carry out a source localization procedure from the magnetic field distribution over the scalp relative to each of the selected epileptiform signals, henceforth referred to as epileptic complexes. An automatic selection procedure would be of great help also in the investigation of more complex diseases like, for instance, generalized epilepsy. In these pathologies the lack of simultaneity, which so far characterizes all neuromagnetic measurements, even those carried out with the largest systems today available, is one crucial drawback, the other being the variety of different signals that are usually recorded also during intercritical periods. For these reasons, the neuromagnetic study of cases of generalized epilepsy has been so far limited to simple morphological approaches, with no attempt to carry out source localization. with the unique exception of a recently reported study (Ricci et al., 1988) of photoconvulsive response, which may be referred to as one possible manifestation of generalized epilepsy (Jeavons and Harding, 1975).

The clinical magnetocardiographer in Ulm

We present a magnetocardiographic system developed in the framework of an international cooperation and installed at the University of Ulm. The system design is conceived for clinical and routine use. The main sensor consists of a hexagonal 55 channels planar system with dc SQUID magnetometers contained in a low noise dewar. Nineteen reference channels are used to form software gradiometers. The patient handling setup and the dewar support are designed to easily allow the positioning procedure and the interaction with the patient. The sensor is operated in a high quality magnetically shielded room made of three soft magnetic metal layers and an external aluminum layer. The low frequency performances can be enhanced by means of an active compensation technique. A set of biopreamplifiers for 128 channels BSPM measurements is included. The data acquisition system allows to perform MCG and BSPM measurement simultaneously with sampling frequencies up to 10 kHz. All basic preprocessing, bandpass filtering, decimation and noise compensation are performed on the fly using an array of DSPs. The data analysis session provides both time domain approach and source identification approach. Direct integration of the results with CT or MRI data is also available.

SQUID sensor with additional compensation module for operation in an AC applied field

A possible implementation of an in-vivo SQUID susceptometer able to estimate the liver iron concentration of humans uses a low frequency applied field together with a lock-in detection. The room-temperature magnetising coils and the detection coils are designed to minimize their mutual coupling. Nevertheless, deviation from ideal behaviour causes a residual signal in the detection coil, with an amplitude significantly larger than the patients. In addition low frequency noise is added by any relative displacement of the magnetising and sensing coils. Thus, we designed a SQUID sensor using a compact compensating module to be used in a multichannel SQUID susceptometer. The sensor consists of two second order axial gradiometers, wounded one inside the other on the same support. The sensing channel is larger than the compensation channel which is only sensitive to the residual signal. Each gradiometer is coupled to a dc SQUID with parallel washer configuration. The output of the compensation channel is A/D converted and is processed by an adaptive algorithm running on a real time unit. The compensation signal is coupled to the sensing channel by an additional feedback loop. The performances of a prototype module will be presented.

An AC magnetizing field biosusceptometer using a SQUID based sensor with additional compensation module

We present a new SQUID based susceptometer for liver iron concentration (LIC) assessment. The instrument is operated with an inhomogeneous AC magnetizing field. The susceptometer is based on 7 channels, each consisting of a sensing unit and of an additional unit for the compensation of the applied field residual and its variation. Each unit is a second order gradiometer and is coupled to a dc SQUID with a parallel washer configuration. The compensation module output is digitally processed and coupled to the sensing gradiometer by means of an additional feedback path. The estimation of LIC is based on a numerical model of the subjects torso obtained from the segmentation of ultrasound images.

Automatic detection of Migrating Motor Complexes using Neural Networks on magnetic recordings of gastric activity

main frequency component of the gastric activity. The Output neuron can assume a value 0 or l for the case of absence or presence, respectively of MMC in every observed pattern. Altered electrical Signals associated with gastric activities can be considered äs an indicator of frequcnt gastric diseases. Migrating Motor Complexes (MMC) can be used to monitor the contractile activity of the stomach during the interdigestive phase. For this reason an automatic method able to detect the MMC directly from raw magnetic data, obtained in a fülly non invasive way, would be veiy helpful in clinical diagnosis.