Ij

Electromagnetic Simulation Aids Development

Of Unique EEG Cap

By Remcom, Inc.

Researchers at the Athinoula A. Martinos Center for Biomedical Imaging at

Massachusetts General Hospital (MGH), Charlestown, Massachusetts, are developing a

special electroencephalography (EEG) cap for use with a research technique that

integrates EEG with magnetic resonance imaging (MRI). This integrated technique offers

the potential to improve investigations of brain activity. But the RF generated by MRI

induces currents in the EEG electrodes. Concerns have arisen that this current could

generate temperature increases in sensitive brain tissues. The researchers recognized that

they would have to evaluate many different electrode types and arrangements in order to

optimize the design of the cap. These experiments obviously cannot be performed on

human volunteers. It would be expensive to build a physical model with properties close

to a human head. On the other hand, virtual models can be developed with properties

very similar to a human head and used to predict the RF fields and tissue exposure to RF.

But virtual models cannot be used with confidence to predict the performance of various

types of electrodes and leads until they have been verified by physical experiments. It is

very expensive to build a matching physical model to verify the accuracy of such a

virtual model.

The MGH researchers overcame this challenge by developing a relatively inexpensive

physical model called a phantom that approximates the behavior of a human head in the

vicinity of an electromagnetic field. They tested this physical model using EEG

electrodes and a bird cage coil that produces a radiofrequency field similar to an MRI

instrument. They simulated the effects of using different types of EEG electrodes and

leads on the phantom using electromagnetic simulation software and the simulation

results were compared to the physical testing results. "A number of different electrodes

were evaluated on both the physical and virtual models to confirm that the virtual model

can predict differences in their performance," said Leonardo Angelone, one of the MGH

researchers. "Now the virtual model is being modified to more closely match human

anatomy and reconfigured to evaluate many different types of arrangements of electrodes.

This capability is aiding the process of designing the EEG cap by enabling the

researchers to quickly evaluate the performance of design alternatives at a lower cost and

higher accuracy than physical experiments alone."

As seen in the August 28, 2008 edition of the RF Globalnet (www.rfglobalnet.com) newsletter.

Feature Article

Promising New Research Technique

The integration of EEG and MRI has the potential to improve brain research. MRI is a

non invasive imaging technique that provides a spatial resolution of millimeters. But MRI

primarily measures a hemodynamic response with a time constant on the order of

seconds. Hemodynamic response refers to the way in which blood flow is regulated in the

brain to give active neural assemblies more energy than inactive assemblies. This means

that MRI cannot reveal the precise mechanics of information exchange within the brain,

which occurs in milliseconds. On the other hand, EEG can deliver the required

millisecond range temporal accuracy but delivers spatial accuracy only in the range of

centimeters. Combining these two methods has the potential to enable major advances in

brain research but there are some safety issues that need to be resolved. There is a

concern that the high magnetic fields used in MRI could generate electrical currents in

the EEG electrodes and produce heating in tissue near the tip of the electrode.

MGH researchers are working to develop a cap that will enable EEG and MRI to be

performed simultaneously while ensuring that specific absorption rate (SAR), which is

used to measure tissue exposure to RF, is kept at safe levels. The new EEG cap is

intended to ensure patient safety at magnetic fields of 7 Tesla. SAR is a measure of the

energy absorbed by biological tissue. A key part of this research is identifying the EEG

electrodes and leads that will provide the lowest exposure levels. The MGH researchers

recently evaluated three different approaches to reducing SAR: the use of electrodes with

different resistivity, putting an RF resistor on the lead near the EEG electrode, and use of

carbon fiber electrodes.

Developing A Virtual Model

Figure 1: Virtual model

As seen in the August 28, 2008 edition of the RF Globalnet (www.rfglobalnet.com) newsletter.

Feature Article

The MGH researchers developed a virtual model of the integrated MRI and EEG

technique being performed on a human head. The electromagnetic fields and SAR

distributions generated by the MRI RF transmit coil were evaluated using XFdtd®

software from Remcom Inc., based in State College, Pennsylvania. XFdtd incorporates a

full-wave three-dimensional solver based on the finite difference time domain (FDTD)

method. According to the Federal Communications Commission of the United States as

stated in OET Bulletin 65, Evaluating Compliance with FCC Guidelines for Human

Exposure to Radiofrequency Electromagnetic Fields, Supplement C: "Currently, the

finite difference time domain (FDTD) algorithm is the most widely accepted

computational method for SAR modeling." Furthermore, in FCC Part 95 Section 603(f) it

is stated: "Applications for equipment authorization of devices operating under this

section must contain a finite difference time domain (FDTD) computational modeling

report showing compliance with these provisions for fundamental emissions."

The head model was developed using MRI data from a male volunteer. One single tissue

was used for the head model with a conductivity of σ=1 S m-1 and permittivity of εr = 100

which are both similar to the properties of the phantom used in the real measurements.

Three different types of EEG electrodes and leads were modeled and co-registered with

the head model. The first set of 32 electrodes was modeled as cylinders with five values

of lead resistivity ranging from 1.67X10-8 to 1 . The second set of electrodes was

created by placing a 10 k resistor between the electrodes and leads of each of the first

set of electrodes. The third set consisted of a 32-electrode EEG cap composed of carbon

fiber leads and metallic AgCl electrodes. A bird cage coil was modeled with 16 perfect

electrical conductor rods of 310 mm in length disposed on a 128 mm ring.

XFdtd was used to compute the electric and magnetic fields and the SAR with each type

of electrode. Simulations showed that the increase of local SAR depended upon the

resistivity used for the EEG leads. SAR values were the same with or without the use of

resistors between the EEG electrode and lead. At each electrode, the SAR value

calculated with electrodes and leads was divided by the SAR values calculated without

electrodes and leads to yield S. When leads with high levels of resistivity were used, S

was close to or even below 1, while for the leads with the lowest levels of resistivity S

increased to levels as high as 2,000. The temperature changes as a function of the SAR

were predicted by solving the heat equation.

As seen in the August 28, 2008 edition of the RF Globalnet (www.rfglobalnet.com) newsletter.

Feature Article

Building A Physical Model

Figure 2: Physical model

The numerical head model was used to build a physical model of the heading with a 3D

printer. Then the model was used to create an RTV silicone mold by pouring the liquid

silicone around the head. The mold was filled with a mixture of distilled water, agarose,

and salt. The salt concentration was selected to have a conductivity of σ = 0.6-1 S m-1

which is in the range of human head tissues. The agarose created a solid gel that allowed

placing the EEG electrodes and leads directly on the phantom surface. Three different

sets of EEG electrodes and leads were used for the measurements. These sets matched the

sets used in the simulation. An MRI coil matching the simulation was used to generate

the magnetic field. A fluoroptic thermometer with four fluorescent probes was used to

measure temperatures near the center of the head, under the frontal pole electrode, 5-15

mm deep under the central electrode, and at the EEG paste of the central electrode.

Figure 3: Simulation shows S decreases with increasing resistivity

As seen in the August 28, 2008 edition of the RF Globalnet (www.rfglobalnet.com) newsletter.

Feature Article

The researchers demonstrated they could accurately simulate the SAR of different types

of EEG electrodes leads. The distribution of temperatures described by the heat equation

matched that observed during the physical measurements. Both the simulations and the

experiments showed that carbon fiber leads performed better than copper leads. Both

simulation and experiments show that SAR is reduced as lead resistivity increases.

However, lead resistivity is limited to a value much smaller than the input resistance of

the EEG amplifier to avoid signal loss. Both simulations and experiments showed SAR

was not affected by the use of an RF resistor between the EEG electrode and the lead.

"This study showed an excellent correlation between simulation and physical

experiments," Angelone said. "Most important the simulation was able to accurately

predict the differences in performance between the various electrodes.

Figure 4: Spatial distribution of S as predicted by simulation

Angelone LM, Vasios CE, Wiggins G, Purdon PL, Bonmassar G.

"On the effect of resistive EEG electrodes and leads during 7 T MRI:

simulation and temperature measurement studies."

Magn Reson Imaging. 2006 Jul;24(6):801-12.

As seen in the August 28, 2008 edition of the RF Globalnet (www.rfglobalnet.com) newsletter.