design of implantable antennas for medical telemetry: dependence

16 International Journal of Monitoring and Surveillance Technologies Research, 1(1), 16-33, January-March 2013

Copyright © 2013, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

ABSTRACTImplantable Medical Devices (IMDs) with wireless telemetry functionalities in the radio-frequency (RF) range are recently attracting significant scientific interest for medical prevention, diagnosis, and therapy. One of the most crucial challenges for IMDs is the design of the integrated implantable antenna which enables bidirectional wireless communication between the IMD and exterior monitoring/control equipment. In this paper, a parametric model of a miniature implantable antenna is initially proposed, which can be adjusted to suit any antenna design and implantation scenario requirements in hand. Dependence of the resonance, radiation, and safety performance of implantable antennas upon (a) operation frequency, (b) tissue anatomy and dielectric properties, and (c) implantation site is further studied. Simulations are carried out: (a) at 402, 433, 868 and 915 MHz considering a 13-tissue anatomical head model, (b) at 402 MHz considering five head models (3- and 5-layer spherical, 6-, 10- and 13-tissue anatomical) and seven dielectric parameter scenarios (variations ±20% in the reference permittivity and conductivity values), and (c) at 402 MHz considering 3-layer canonical models of the human head, arm, and trunk. The study provides valuable insight into the design of implantable antennas. Finite Element and Finite Difference Time Domain numerical solvers are used.

Design of Implantable Antennas for Medical Telemetry:

Dependence upon Operation Frequency, Tissue Anatomy, and Implantation Site

Asimina Kiourti, School of Electrical and Computer Engineering,National Technical University of Athens, Athens, Greece

Konstantina S. Nikita, School of Electrical and Computer Engineering,National Technical University of Athens, Athens, Greece

Keywords: Implantable Antenna, Industrial Scientific and Medical (ISM) Band, Medical Implant Communications Service (MICS) Band, Medical Telemetry, Miniaturization, Specific Absorption Rate (SAR)

INTRODUCTION

Implantable Medical Devices (IMDs) with wireless telemetry functionalities in the radio-frequency (RF) range are nowadays used to perform an expanding variety of diagnostic

and therapeutic functions. Example applica-tions include temperature monitors (Scanlon, 1997), pacemakers and cardioverter defibril-lators (Wessels, 2002), functional electrical stimulators (FES) (Guillory & Normann, 1999), blood-glucose sensors (Shults, 1994), cochlear (Buchegger, 2005), gastric and bladder control-lers (Sani, 2009), glucose monitors (Karacolak,

DOI: 10.4018/ijmstr.2013010102


International Journal of Monitoring and Surveillance Technologies Research, 1(1), 16-33, January-March 2013 17

2008), retinal implants (Gosalia, 2004) etc. As technology continues to evolve, new implant-able medical devices are being developed, and their use is expected to rapidly increase from an already large base.

A key and critical component of RF-enabled IMDs is the integrated implantable antenna, which enables bidirectional wireless communication between the IMD and exte-rior monitoring/control equipment. Designers of implantable antennas need to deal with a number of challenges, including miniaturiza-tion, biocompatibility, impedance matching, radiation performance, and compliance with international safety guidelines for the specific absorption rate (SAR). Patch designs are usually preferred, because of their flexibility in design, shape, and conformability. Furthermore, patch antennas lend themselves easily to a number of miniaturization techniques, including use of high-permittivity dielectric materials, lengthen-ing of the current flow path excited on the radiat-ing patch, addition of shorting pins between the ground and patch planes, and vertical stacking of multiple patches (Kiourti & Nikita, 2012a).

Medical implant communications most commonly take place in the medical implant communications service (MICS) band (402.0–405.0 MHz), which is regulated by the United States Federal Communications Commission (FCC, 1999) and the European Radiocom-munications Committee (ERC, 1997). The 433.1-434.8, 868.0-868.6 and 902.8-928.0 MHz industrial, scientific and medical (ISM) bands are additionally suggested for biotelemetry in some countries (ITU-R). However, focus is on the MICS band, because of its advantages to be available worldwide and feasible with low power and low cost circuits, reliably support high data rate transmissions, fall within a rela-tively low noise portion of the spectrum, and acceptably propagate through human tissue.

Since implantable antennas are intended to operate inside human tissue, their performance strongly depends on the surrounding tissue environment. This includes the anatomical features of the individual, the dielectric param-eters (permittivity, εr, and conductivity, σ) of

the biological tissues, and the part of the body where the antenna is to be implanted, known as the implantation site. For example, smaller size (female and low body mass index male) anatomical models have been found to exhibit higher radiated power levels and far-field gain values (Sani, 2009). Uncertainties and inter-subject variability in dielectric parameters need also to be taken into account. Maximum standard deviations of 16% have been reported in the dielectric parameter values of rat brain tissue (Bao, 1997), while a decrease in per-mittivity and conductivity values by 4% and 10% has been recorded for pig tissue within 4 h after death, respectively (Schmid, 2003). Dependency of dielectric parameters on the age of the subject has also been highlighted (Conil, 2008). Finally, the intended implantation site determines the types and structure of the tissues surrounding the antenna, and, therefore, its di-electric loading. For example, skin-implantable antennas might operate in a different way while implanted within the skin-tissue of different parts of the body.

In this study, a parametric model of a min-iature implantable antenna is proposed, which can be adjusted to suit any antenna design (e.g. size, material etc) and tissue model (e.g. implantation site) requirements in hand. Depen-dence of the resonance, radiation, and safety performance of implantable antennas upon (a) operation frequency, (b) tissue anatomy and dielectric properties, and (c) implantation site is further studied. Simulations are carried out: (a) at 402, 433, 868 and 915 MHz considering a 13-tissue anatomical head model (Kiourti & Nikita, 2012b), (b) at 402 MHz considering five head models (3- and 5-layer spherical, 6-, 10- and 13-tissue anatomical) and seven dielectric parameter scenarios (variations ±20% in the reference permittivity and conductivity values) (Kiourti & Nikita, 2012c), and (c) at 402 MHz considering 3-layer canonical models of the human head, arm, and trunk (Kiourti & Nikita, 2012d, 2012e). Finite Element (Ansoft HFSS (HFSS, 2008)) and Finite Difference Time Domain (Remcom XFDTD (XFDTD, 2005)) numerical solvers are used. The utmost goal is



to provide a valuable insight into the design of implantable antennas, while at the same time addressing issues related to operation frequency selection, suitability of canonical against ana-tomical tissue models for antenna design and performance evaluation, and uncertainties inherent to the tissue anatomy and dielectric parameters and implantation site.

MODELS AND METHODS

Antenna Models

Free-space wavelength at the frequency bands allocated for medical implants can be computed as 74.6, 69.3, 34.6, and 32.8 cm at 402, 433, 868, and 915 MHz, respectively. Therefore, dimensions of the traditional half–wavelength or quarter–wavelength antennas occur to be inadequately high for implantable medical ap-plications. Human tissues in which implantable antennas are intended to operate exhibit high permittivity (εr) values, which, in turn, increase the effective dielectric constant of the antenna, and work to advantageously reduce its physical size. However, a number of additional minia-turization techniques need to be applied in order to render the size of the designed antenna suitable for implantation. In this study, a para-metric model of an implantable patch antenna is proposed, which includes a superstrate layer for biocompatibility purposes, and combines the following miniaturization techniques in order to reduce size:

• Use of high-permittivity and low-thickness dielectric materials. High-permittivity dielectrics are selected for implantable patch antennas (e.g. Rogers 3210, εr = 10.2, ceramic alumina, εr = 9.4), because they shorten the effective wavelength and result in low resonance frequencies. However, even with such high-permittivity dielec-trics, the superstrate layer still insulates the antenna structure from the higher-permittivity surrounding tissue material.

Therefore, dielectric materials with high permittivity values and thin superstrate layers are solicited;

• Lengthening of the current flow path on the patch surface. Longer effective current flow paths excited on the radiating patch of the antenna can reduce its resonance frequency, and achieve a more compact size. Meandering techniques of the patch surface area are hereafter considered, while spiral (Kiourti, 2011a), waffle-type (So-ontornpipit, 2005), and hook-slotted (Liu, 2008) patches have also been suggested in the literature for this purpose;

• Addition of a shorting pin. Inserting a shorting pin between the ground and patch planes increases the effective size of the antenna, and, in turn, reduces the required physical dimensions. The technique works in much the same way that a ground plane doubles the height of a monopole antenna, i.e., it typically produces a planar inverted-F antenna (PIFA) with the same resonance performance as a double-sized antenna without the shorting pin;

• Patch-stacking. Vertically stacking two radiating patches reduces antenna size by increasing (nearly doubling) the length of the current flow path.

The proposed parametric antenna model is shown in Figure 1. The model consists of a ground plane (radius of Rg) and two verti-cally stacked patches (radius of Rp) printed on dielectric substrates (thicknesses of h1 and h2, respectively). A dielectric superstrate (thickness of h3) covers the structure in order to provide separation between the copper of the upper patch and the body, thus ensuring biocompatibility and robustness for the antenna. The dielectric material is considered to be the same for the substrates and superstrate, but selectable accord-ing to the intended antenna design. Meanders of variable lengths (Li, i = 1–5, 1’–6’) and widths (w) are inserted into the patches in order to increase the effective length of the current flow



(inter-meander distance of 1 mm). A shorting pin (S: sx, sy) connects the ground plane with the lower patch to increase the apparent length of the antenna, while a 50 Ohm coaxial cable excites both radiating patches (F: (fx, fy)). An-tenna design parameters can be appropriately selected for a good 50 Ohm impedance match at the desired operation frequency, as dictated by the antenna design (e.g. size, material etc) and tissue model (e.g. implantation scenario) requirements in hand. Throughout this paper, the origin of the Cartesian coordinate system is considered to be located at the center of the antenna’s ground plane, as shown in Figure 1. Circular shape has been chosen for the proposed antenna structure in order to avoid sharp edges which might harm the surrounding biological tissues.

Examples of miniature implantable anten-nas which have been designed based on the proposed parametric model, and which are used in this study are given in Table 1. Design of all antennas has been carried out while implanted inside the skin-tissue of the 3-layer spherical head model of Figure 3(a). To validate the pro-posed parametric antenna model, a prototype of

antenna model #(1) has been fabricated (Figure 2(a)) and experimentally tested inside a liquid emulating skin-tissue properties at 402 MHz (εr = 46.7, σ = 0.69 S/m (Gabriel, 1996)) (Ki-ourti & Nikita, 2012b). Good agreement exists between the simulated and measured reflection coefficient frequency response, as shown in Figure 2(b). Slight discrepancies are observed which are within the uncertainty range imposed by the in-house antenna fabrication procedure.

Tissue Models

Numerical tissue models of the human head, arm and trunk are used, as shown in Figure 3. Antenna performance is evaluated for implan-tation inside the skin-tissue of these models and applications such as intra-cranial pressure (head), blood pressure (arm), and glucose (trunk) monitoring. More specifically, the numerical tissue models of this study include:

• A 3-layer spherical head model consisting of skin (thickness of 5 mm), cortical bone (thickness of 5 mm), and gray matter tissues (Figure 3(a)) (Kiourti & Nikita, 2012b);

Figure 1. Proposed parametric antenna model: (a) ground plane, (b) lower patch, (c) upper patch, and (d) side view



• A 5-layer spherical head model consisting of skin (thickness of 5 mm), cortical bone (thickness of 4 mm), cerebrospinal fluid (thickness of 2 mm), gray matter (thick-ness of 14 mm), and white matter tissues (Figure 3(b));

• A 6-tissue anatomical head model con-sisting of skin (thickness of 9 mm on the z-axis), cortical bone (thickness of 9 mm on the z-axis), gray matter, vitreous humor, cartilage, and muscle tissues (Figure 3(c));

• A 10-tissue anatomical head model con-sisting of skin (thickness of 4.7 mm on the z-axis), cortical bone (thickness of 8.4 mm on the z-axis), cerebrospinal fluid, gray

matter, white matter, eye sclera, muscle, bone marrow, blood, and fat tissues (Figure 3(d)) (Wiart, 2008);

• A 13-tissue anatomical head model con-sisting of skin (thickness of 5.2 mm on the z-axis), cortical bone (thickness of 7.7 mm on the z-axis), cerebrospinal fluid, gray matter, white matter, eye sclera, vitreous humor, lens, cartilage, muscle, dura, spinal cord, and cerebellum tissues (Figure 3(e)) (Kiourti & Nikita, 2012b);

• A 3-layer cylindrical arm model consisting of skin (thickness of 5 mm), muscle (thick-ness of 25 mm), and bone tissues (Figure 3(f)) (Wegmueller, 2006);

Table 1. Parameter values of implantable antennas used in the study, designed based on the parametric model of Figure 1 (in [mm])

Figure 2. Experimental validation of the proposed parametric antenna model: (a) fabricated prototype, and (b) simulated versus measured reflection coefficient frequency response



• A 3-layer ellipsoidal trunk model consisting of skin (thickness of 5 mm), fat (thickness of 10 mm), and muscle tissues (Figure 3(g)) (Shiba, 2008).

Tissue dielectric parameters (permittivity, εr, and conductivity, σ) used in this study are summarized in Table 2 (Gabriel, 1996), and approximated as constant inside a ±100 MHz frequency range around the intended operation frequency (402, 433, 868 or 915 MHz). Using this approximation, the maximum errors in permittivity (εr) and conductivity (σ) values at 402 MHz are given by 6.59% and 8.89%, re-spectively (Kiourti, 2011b). Accuracy is further improved at higher frequencies. Deviations are small enough, justifying the suitability of con-sidering frequency-independent tissue dielectric parameters inside a small frequency range. Mass densities of the tissues under consideration are also provided in Table 2.

Numerical Methods

Finite Element (FE) simulations are carried out within the 3-layer spherical head model (Figure 3(a)), the 3-layer cylindrical arm model (Figure 3(f)), and the 3-layer ellipsoidal trunk model (Figure 3(g)), using the commercial software Ansoft HFSS (HFSS, 2008). The solver uses a tetrahedron-shaped basic mesh element, which significantly accelerates solving of curved geometries, and has widely been used for implantable antenna design and performance evaluation (e.g. (Liu, 2008; Huang & Kishk, 2011; Kiourti & Nikita, 2012b)). Meshing is iteratively refined in an automatic way, with a perturbation of 30% between each iterative pass. The mesh refinement procedure stops when the maximum change in the reflection coefficient magnitude between two consecutive passes is less than 0.02, or when the number of passes exceeds 10.

Figure 3. Tissue models used in the study: (a) 3-layer spherical head model (Kiourti & Nikita, 2012b), (b) 5-layer spherical head model, (c) 6-tissue anatomical head model, (d) 10-tissue anatomical head model (Wiart, 2008), (e) 13-tissue anatomical head model (Kiourti & Nikita, 2012b), (f) 3-layer cylindrical arm model (Wegmueller, 2006), and (g) 3-layer ellipsoidal trunk model (Shiba, 2008)



Finite Difference Time Domain (FDTD) simulations are carried out within the spheri-cal and anatomical head models (Figure 3(a)-(e)) using the commercial software Remcom XFDTD (XFDTD, 2005). The FDTD method has long been used to assess bioelectromagnetic interactions (Sullivan, 1987), while the XFDTD solver enables efficient modeling of detailed anatomical human body parts. Tissue models are meshed in cubic cells with an edge of 3.5 mm (3- and 5-layer spherical), 3 mm (6-tissue anatomical), 1.3 mm (10-tissue anatomical), and 1.25 mm (13-tissue anatomical). Sub-gridding with a non-uniform grid of 0.1×0.1×0.2 mm3 is performed to accurately model the miniature implantable antenna. Meshing is adaptive to avoid abrupt transitions.

In order to take free-space radiation into account and extend radiation infinitely far, all simulation setups are surrounded by absorbing boundaries. These are set λ0/4 (λ0 is the free-space wavelength) away from the geometries under consideration in an attempt to guarantee stability of the FE and FDTD numerical calcula-tions (HFSS, 2008; XFDTD, 2005).

RESULTS

Effect of Operation Frequency

A comparative evaluation regarding the depen-dency of implantable antenna performance upon operation frequency is carried out (Kiourti & Nikita, 2012b). Implantable antennas operating at 402, 433, 868 and 915 MHz (antennas #1-#4 of Table 1) are assessed in terms of resonance, radiation, and safety performance, for implanta-tion inside the skin-tissue (scalp) of the 13-tis-sue anatomical head model of Figure 3(e). A multi-tissue anatomically-based head model is selected in an attempt to obtain realistic simula-tion results. The antenna is implanted 3.6 mm under the skin, with its ground plane placed in parallel with the horizontal plane of the head model. Antenna designs exhibit identical physical (volume of 203.6 mm3), but varying effective dimensions in order to accommodate the different operating frequencies. Longer meanders increase the effective length of the current on the radiating patches, and work to advantageously reduce the exhibited operation

Table 2. Dielectric parameters (Gabriel, 1996) and mass densities of the tissues used in the study

Tissue Type 402 MHz 433 MHz 868 MHz 915 MHzMass

Densityεr σ [S/m] εr σ [S/m] εrσ

[S/m] εr σ [S/m]

skin (scalp) 46.74 0.689 46.08 0.702 41.58 0.856 41.33 0.872 1100

bone (skull) 13.10 0.090 13.07 0.094 12.48 0.139 12.44 0.145 2200

dura 46.65 0.827 46.38 0.835 44.51 0.951 44.39 0.966 1100

CSF 70.97 2.252 70.64 2.260 68.71 2.399 68.61 2.419 1020

grey matter 57.39 0.738 56.83 0.751 52.88 0.929 52.65 0.949 1030

white matter 42.05 0.445 41.67 0.454 38.99 0.581 38.84 0.595 1030

muscle 57.11 0.797 56.87 0.805 55.11 0.932 54.99 0.948 1040

cartilage 45.45 0.587 45.15 0.598 42.77 0.768 42.6 0.789 1100

vitreous humor 69.00 1.529 68.99 1.534 68.91 1.627 68.89 1.641 1000

lens 48.14 0.669 47.96 0.675 46.63 0.784 46.55 0.798 1100

eye sclera 57.66 1.005 57.38 1.014 55.36 1.155 55.23 1.173 1100

spinal cord 35.39 0.447 35.05 0.456 32.63 0.565 32.49 0.578 1040

cerebellum 55.94 1.031 55.14 1.048 49.66 1.248 49.35 1.269 1030



frequency. Therefore, the patch surface area of the 433, 868, and 915 MHz resonating antennas is found to be increased by 2.2%, 24.9%, and 25.4% as compared to that of the 402 MHz antenna. Numerical simulations are carried out in Remcom XFDTD, using the Finite Difference Time Domain (FDTD) method (XFDTD, 2005):

1. Resonance Performance: The reflection coefficient frequency responses of the antennas implanted inside the 13-tissue anatomical head model are shown in Figure 4(a). The antennas resonate at the desired operation frequencies of 402, 433, 868, and 915 MHz, and exhibit broad bandwidths which cover the frequency bands of interest. Computed values of the exhibited 10-dB bandwidths (10 dB-BW) are shown in Table 3. Bandwidth enhancement with increasing frequency is attributed to the larger patch surface area of the antennas (Notis, 2004);

2. Radiation Performance: Radiation per-formance of the antennas is evaluated in terms of the radiation patterns and maxi-mum gain values achieved. The 3-D far-field gain radiation patterns exhibited by the antennas inside the 13-tissue anatomical head model are shown in Figure 4(b). Since the anatomical head model is an asymmetri-cal, inhomogeneous dielectric structure, radiation patterns, which depend on the structure and shape of the implantation site, are not symmetric either. Increased tissue absorption at high frequencies causes at-tenuation which deteriorates symmetry, and is consistent with the findings in (Chirwa, 2003; Xu, 2009) for ingestible antennas. Losses inside the body come from attenu-ation by the weakly conductive tissues and reflections at each of the boundaries of dissimilar tissue. Maximum far-field gain values (Gmax) exhibited by the antennas are

Figure 4. Effect of operation frequency: (a) reflection coefficient frequency responses, (b) far-field gain radiation patterns, and (c) local SAR distributions (net-input power of 4.927 mW) for the 402, 433, 868 and 915 MHz antennas under study (#1-#4 of Table 1) implanted inside the skin of the 13-tissue anatomical head model of Figure 3(e)



shown in Table 3. Because of the miniature antenna size and high tissue loss, low values of gain are recorded. Higher gain values at increased frequencies are attributed to the enhanced copper surface area of the patches;

3. Safety Performance: Assessment of the antennas’ safety performance involves evaluation of their compliance with the IEEE C95.1-1999 (IEEE, 1999) and IEEE C95.1-2005 (IEEE, 2005) safety guide-lines for the SAR. The IEEE C95.1-1999 guidelines restrict the SAR averaged over any 1 g of tissue in the shape of a cube (1 g-avg SAR) to less than 1.6 W/kg, while the IEEE C95.1-2005 guidelines restrict the SAR averaged over any 10 g of tissue in the shape of a cube (10 g-avg SAR) to less than 2 W/kg. The net-input power to the antennas is initially set to 1 W, and the maximum 1 g-avg and 10 g-avg SAR values are recorded. Based on these, we calculate the maximum allowable net-input power levels which satisfy the IEEE C95.1-1999 (P1999) and IEEE C95.1-2005 (P2005) restrictions for the SAR. Results are shown in Table 3. The IEEE C95.1-1999 guidelines are found to be much stricter, limiting the net-input power to more than six times lower than that imposed by the IEEE C95.1-2005 guidelines. Local SAR distributions generated in the surrounding tissues are shown in Figure 4(c), for the FDTD slices where maximum local SAR values have been recorded (net-input power

of 4.927 mW). Simulation results related to the safety performance of the antennas are justified by their identical physical but varying effective dimensions. At higher operation frequencies, electric field, or, equivalently, current density, is more uniformly distributed across an increased surface area of the radiating patches. Therefore, lower maximum SAR values are exhibited, along with more expanded field distribution in the surrounding tissues. However, it is important to highlight that if the antennas were scaled in physical rather than in effective size, then high operation frequencies, and, thus, tissue conductivities, would be expected to result in increased SAR values and more concentrated SAR distributions (Scanlon, 2000).

Effect of Tissue Anatomy and Dielectric Parameters

An implantable antenna operating at 402 MHz (antenna #5 of Table 1) is assessed in terms of resonance, radiation, and safety performance, for implantation inside the skin-tissue (scalp) of the five numerical head models of Figure 3(a) through Figure 3(e) (two spherical and three anatomical). Seven dielectric parameter scenarios are considered: reference dielectric parameters at 402 MHz (Table 2), variation in all tissue permittivities (εr) by ±20%, variation in all tissue conductivities (σ) by ±20%, and si-multaneous variation in all permittivities (εr) and

Table 3. Effect of operation frequency: 10 dB bandwidth (10 dB-BW), maximum far-field gain (Gmax), and maximum allowable net-input power levels imposed by the IEEE C95.1-1999 (P1999) and IEEE C95.1-2005 (P2005) guidelines for the 402, 433, 868 and 915 MHz antennas under study (#1-#4 of Table 1) implanted inside the skin of the 13-tissue anatomical head model of Figure 3(e)

Oper. Freq. 10 dB-BW [MHz] Gmax [dB] P1999 [mW] P2005 [mW]

402 MHz 27 -36.90 4.927 30.02

433 MHz 28 -35.99 5.166 30.13

868 MHz 38 -35.14 5.388 30.28

915 MHz 40 -32.94 5.426 30.41



conductivities (σ) by ±20% (Kiourti & Nikita, 2012c). High variations in dielectric parameters are considered for a worst-case variability analysis. The aim is to quantify the sensitiv-ity of the performance of a scalp-implantable antenna to variations in head anatomy and dielectric parameters. The antenna is positioned symmetrically inside the skin-tissue of each head model, with its ground plane placed in parallel with the horizontal plane of the tissue model. Numerical simulations are carried out in Remcom XFDTD, using the Finite Difference Time Domain (FDTD) method (XFDTD, 2005):

1. Resonance Performance: The antenna is found to be well-matched at 402 MHz for all tissue model and dielectric parameter scenarios under study. Maximum, mini-mum, and average values of the exhibited return loss (RL), defined as the incident to reflected power ratio, are shown in Table 4.

Anatomy of the numerical head model is found to insignificantly influence impedance matching among the 3- and 5-layer spherical and 10- and 13-tissue anatomical head models. Impedance matching within the 6-tissue ana-tomical head model strongly varies. For six of the dielectric parameter scenarios under study, the RL is degraded by 31.8% on average, as compared to the 3-layer spherical head model. An increased RL value (+9.0%) is computed only when tissue permittivities (εr) are decreased

by 20%. This is attributed to the fact that the 6-tissue head model exhibits a bump of skin-tissue at the upper side of the head (implantation site), resulting in a less smooth head curvature compared to the other head models under study. This, in turn, alters the dielectric loading to the antenna by the surrounding tissues and exterior free-space, and greatly affects its performance.

Maximum variations from the reference RL values when dielectric parameters are var-ied by ±20% are shown in Table 5, for each of the five head models under study. Impedance matching is degraded when tissue permittivities (εr) are increased as well as when only tissue conductivities are increased. However, imped-ance matching is improved when only tissue conductivities are decreased. All numerical head models confirm these findings:

2. Radiation Performance: Highest, low-est, and average values of the maximum exhibited far-field gain values (Gmax) of the antenna are shown in Table 4. All tissue model and dielectric parameter scenarios have been considered. Because of the small antenna size and high tissue loss, low values of gain are computed.

Highly comparable Gmax values are com-puted among the 3- and 5-layer spherical and 10- and 13-tissue anatomical head models (variations <6.4%). However, much lower gain values are observed within the 6-tissue anatomi-cal head model, as attributed to its anatomical

Table 4. Effect of tissue model anatomy and dielectric parameters: maximum, minimum, and average variations in return loss (RL), maximum far-field gain (Gmax), and maximum allowable net-input power levels imposed by the IEEE C95.1-1999 (P1999) and IEEE C95.1-2005 (P2005) guidelines for the 402 MHz antenna under study (#5 of Table 1) implanted inside the skin of the head models of Figure 3(a) through Figure 3(e) when seven dielectric parameter scenarios are considered

RL Gmax P1999 P2005

Minimum 25.05 -34.2 5.503 26.011

Maximum 9.51 -49.9 6.946 27.892

Average 16.40 -39.2 6.594 26.950



differences around the antenna implantation site. Compared to the 3-layer spherical head model, the values of Gmax within the 6-tissue anatomical head are decreased by 34.6%, on average, for the dielectric parameter scenarios under study.

Maximum variations from the reference Gmax values when dielectric parameters are varied by ±20% are shown in Table 5, for each of the five head models under study. Compared to the reference dielectric parameter scenario, Gmax is always enhanced when only tissue permittivies (εr) are increased as well as when only tissue conductivities (σ) are decreased. An increase in tissue conductivities (σ) always degrades Gmax.

Far-field gain radiation patterns for the reference dielectric parameter scenario are shown in Figure 5(a). Asymmetry in radiation is attributed to the asymmetry of the heterogeneous numerical head models. Symmetry is strongly deteriorated in the case of the 6-tissue anatomi-cal head model, as attributed to differences in the dielectric environment around the implant location compared to the other head models. The 6-tissue model exhibits the least smooth head curvature, or, equivalently, the highest asymmetry in the area immediately surrounding the antenna. Radiation patterns are found to be similar for the same numerical head model with different dielectric parameters:

3. Safety Performance: Following the pro-cedure described previously, we compute the maximum allowable net-input power levels which guarantee conformance with the IEEE C95.1-1999 (P1999) (IEEE, 1999) and IEEE C95.1-2005 (P2005) (IEEE, 2005) guidelines. Maximum, minimum, and aver-age values of the P1999 and P2005 computed for varying tissue model and dielectric pa-rameters are shown in Table 4. On average, the IEEE C95.1-1999 guidelines are found to be stricter by more than four times as compared to the recent IEEE C95.1-2005 guidelines.

Anatomy of the head model is shown to insignificantly influence P1999 among the 3- and 5-layer and 10- and 13-tissue anatomical head models (variation by <2.5%), while higher varia-tions are observed within the 6-tissue anatomical head model (by 19.5%, on average, as compared to the 3-layer spherical head model). These are attributed to differences in the anatomy of the tissue model in the area which immediately surrounds the implant. However, values for P2005 are found to be almost independent of the numerical head model (variation by <3.7%).

Maximum variations from the reference P1999 and P2005 values when dielectric param-eters are varied by ±20% are shown in Table 5,

Table 5. Effect of tissue model anatomy and dielectric properties: maximum variations from the reference values of return loss (RL), maximum far-field gain (Gmax), and maximum allowable net-input power levels imposed by the IEEE C95.1-1999 (P1999) for the 402 MHz antenna under study (#5 of Table 1) implanted inside the skin of the head models of Figure 3(a) through Figure 3(e) when variations by ±20% in dielectric parameters are considered

RL Gmax P1999 P2005

3-layer spherical -37.6% +4.7% -0.32% +0.74%

5-layer spherical -39.9% +1.6% +0.38% -0.81%

6-tissue anatomical +35.2% +4.6% -0.57% -0.75%

10-tissue anatomical -38.9% +4.0% -0.70% +1.47%

13-tissue anatomical -37.0% +3.5% -0.53% -0.71%



for each of the five head models under study. Regardless of the numerical head model, it be-comes evident that reduced values are computed for P1999 and P2005 when tissue conductivities (σ) are decreased.

Distributions of the 1 g-avg and 10 g-avg SAR considering the reference dielectric pa-rameter scenario are shown in Figure 5(b) and Figure 5(c), respectively. In general, SAR dis-tributions inside the 3- and 5-layer spherical and the 10- and 13-tissue anatomical head models are highly comparable. More concentrated SAR distributions are observed within the 6-tissue anatomical head model, which are attributed to its differentiated anatomical features around the implant location. SAR distributions inside the same head model with different dielectric parameters are found to be similar.

Effect of Implantation Site

The effect of implanting the exact same an-tenna within different parts of the human body (implantation sites) is hereafter investigated (Kiourti & Nikita, 2012d, 2012e). An implant-able antenna operating at 402 MHz (antenna #6 of Table 1) is assessed in terms of resonance, radiation, and safety performance, for implanta-tion inside the skin-tissue of the human head, arm, and trunk. Given the minor effects of tissue anatomy and dielectric parameters on the per-formance of implantable antennas, as indicated in the previous sections, canonical models of the human head (Figure 3(a)), arm (Figure 3(f)), and trunk (Figure 3(g)) are considered, along with reference dielectric parameters at 402 MHz (Table 2). The antenna is positioned

Figure 5. Effect of tissue model anatomy: (a) far-field gain radiation patterns, (1) 1 g-avg SAR distributions, and (c) 10 g-avg SAR distributions for the 402 MHz antenna under study (#5 of Table 1) implanted inside the skin of the head models of Figure 3(a) through Figure 3(e) when reference dielectric parameters are considered



symmetrically by 2.5 mm under the skin-tissue of each tissue model, with its ground plane placed in parallel with the horizontal plane of the model. Numerical simulations are carried out in Ansoft HFSS, using the Finite Element (FE) method (HFSS, 2008):

1. Resonance Performance: The reflection coefficient frequency responses exhibited by the antenna within each of the tissue models under consideration are shown in Figure 6(a). Operation of the scalp-implantable antenna inside the skin-tissue of other parts of the body is found to cause insignificant changes in the exhibited reso-nance performance. This can be attributed to the fact that the antenna exhibits similar dielectric loading from the surrounding tissues and exterior air in all three implan-tation scenarios. The antenna resonates at 401.6 MHz, 401.4 MHz, and 397.8 MHz, within the head, arm, and trunk models, with broad bandwidths which cover the MICS band. Computed values of the ex-hibited 10-dB bandwidths (10 dB-BW) are given in Table 6;

2. Radiation Performance: The 3-D far-field gain radiation patterns of the antenna implanted inside the skin-tissue of the canonical head, arm, and trunk models are shown in Figure 6(b). Since the antenna is electrically very small, it radiates nearly omni-directional, monopole-like radiation patterns, which are, however, affected by the surrounding heterogeneous and asymmetrical medium. Maximum far-field gain values (Gmax) are shown in Table 6. Low gain values are computed, which are attributed to the small antenna size and surrounding tissue loss, and are consistent with the radiation performance of minia-ture implantable antennas reported in the literature (e.g. (Kiourti, 2011b)). Given the fact that dielectric loading of the antenna is similar in all three implantation scenarios, highly comparable results are recorded for its radiation performance;

3. Safety Performance: Issues related to patient safety limit the maximum allowable net-input power to the antenna. An SAR

Figure 6. Effect of implantation site: (a) reflection coefficient frequency responses, and (b) far-field gain radiation patterns for the 402 MHz antenna under study (#6 of Table 1) implanted inside the skin of canonical head (Figure 3(a)), arm (Figure 3(f)), and trunk (Figure 3(g)) models



analysis is therefore carried out for all im-plantation scenarios under consideration, and conformance with the IEEE safety guidelines for the SAR is assessed. Based on the procedure described previously, the maximum allowable net-input power levels which guarantee conformance with the IEEE C95.1-1999 (P1999) and IEEE C95.1-2005 (P2005) guidelines are computed. Results are summarized in Table 6. The old IEEE C95.1–1999 guidelines are found to limit the maximum allowable net–input power to more than ten times lower than that imposed by the recent IEEE C95.1–2005 guidelines. Similarities in the simulation results for all implantation scenarios under consideration are justified by the similar dielectric loading of the antenna.

CONCLUSION

In this paper, a parametric model of a miniature implantable patch antenna was initially sug-gested, which can be appropriately modified to address any antenna design (e.g. size, material etc) or tissue model (e.g. implantation site) requirements. Dependence of the resonance, radiation, and safety performance of implant-able antennas upon (a) operation frequency, (b) tissue anatomy and dielectric properties, and (c) implantation site was further studied.

Selection of the operation frequency was highlighted as crucial for implantable antennas. Antennas of identical physical but varying ef-fective dimensions were found to achieve en-

hanced gains (10.7% increase in the maximum far–field gain at 915 MHz as compared to 402 MHz), reduced SAR values (9.2% and 1.3% decrease in the 1 g– and 10 g–averaged SAR), increased maximum allowable net–input power levels (10.1% and 1.3% increase imposed by the IEEE C95.1–1999 and IEEE C95.1–2005 safety guidelines [10]), and more expanded SAR distributions. Results are attributed to the enhanced patch surface area of the antennas at higher frequencies.

Taking tissue anatomy and dielectric pa-rameters uncertainties into account was also shown to be significant for implantable antenna design and performance evaluation. Simulation results within five different numerical head models and seven dielectric parameter scenarios indicated the need for designing implantable antennas with enhanced bandwidth in order to compensate for potential detuning and imped-ance mismatch. A conservative evaluation of the link-budget between the implantable antenna and exterior equipment was also characterized as necessary in order to deal with potential degraded gain values and deteriorated symme-try in radiation. Compared with the reference dielectric parameter scenario within the 3–layer spherical head model, maximum variations of -55.1%, -39.2%, -19.9%, and +3.7%, were com-puted in return loss, maximum far–field gain, and maximum allowable net–input power levels imposed by the IEEE C95.1-1999 and IEEE C95.1-2005 safety guidelines, respectively. Compliance with the recent IEEE C95.1-2005 guidelines was found to be almost insensi-tive to tissue properties, as opposed to IEEE

Table 6. Effect of implantation site: 10 dB bandwidth (10 dB-BW), maximum far-field gain (Gmax), and maximum allowable net-input power levels imposed by the IEEE C95.1-1999 (P1999) and IEEE C95.1-2005 (P2005) guidelines for the 402 MHz antenna under study (#6 of Table 1) implanted inside the skin of canonical head (Figure 3(a)), arm (Figure 3(f)), and trunk (Figure 3(g)) models

10 dB-BW Gmax P1999 P2005

head 35.2 -50.98 1.664 20.853

arm 35.3 -54.16 1.676 20.951

trunk 35.4 -53.26 1.678 20.962



C95.1-1999. Furthermore, canonical models were proved to be suitable for design and per-formance evaluation of implantable antennas, as long as anatomy of the tissue model around the implantation site is not significantly altered.

Finally, implantation of a specific antenna inside different parts of the human body was shown to insignificantly affect its performance. More specifically, if an antenna which has been designed for implantation inside a specific type of tissue (e.g. skin-tissue) is operating inside the same type of tissue of different parts of the body (e.g. skin-tissue of the human head, arm, or trunk), then its resonance, radiation, and safety performance will remain almost unaltered. Results are attributed to the similar dielectric loading exhibited by the surrounding tissues and exterior air on the antenna in each of the implantation scenarios.

ACKNOWLEDGMENT

This work has been supported by the Operational Programme “Education and Lifelong Learning”, co-financed by the European Union (Euro-pean Social Fund - ESF) and national funds, under the project ARISTEIA DEM-II-MED (“Implantable and Ingestible Medical Devices (IIMDs): Optimal-Performance-Oriented De-sign and Evaluation Methodology”). The work of A.K. was supported by the IEEE Antennas and Propagation Society Doctoral Research Award and the IEEE Microwave Theory and Techniques Society Graduate Fellowship for Medical Applications.

REFERENCES

Bao, J. Z., Lu, S. T., & Hurt, W. D. (1997). Complex dielectric measurements and analysis of brain tis-sues in the radio and microwave frequencies. IEEE Transactions on Microwave Theory and Techniques, 45(10), 1730–1741. doi:10.1109/22.641720.

Buchegger, T., Obberger, G., Reisenzahn, A., Hochmair, E., Stelzer, A., & Springer, A. (2005). Ultra-wideband transceivers for cochlear implants. EURASIP Journal on Applied Signal Processing, 18, 3069–3075. doi:10.1155/ASP.2005.3069.

Chirwa, L. C., Hammond, P. A., Roy, S., & Cum-ming, D. R. S. (2003). Electromagnetic radiation from ingested sources in the human intestine between 150 MHz and 1.2 GHz. IEEE Transactions on Bio-Medical Engineering, 50, 484–492. doi:10.1109/TBME.2003.809474 PMID:12723060.

Conil, E., Hadjem, A., Lacroux, F., Wong, M. F., & Wiart, J. (2008). Variability analysis of SAR from 20 MHz to 2.4 GHz for different adult and child models using finite-difference time-domain. Physics in Medicine and Biology, 53, 1511–1525. doi:10.1088/0031-9155/53/6/001 PMID:18367785.

European Radiocommunications Committee. (ERC). (1997). European Radiocommunications Committee recommendation 70-03 relating to the use of short range devices. In Proceedings of the European Con-ference of Postal and Telecommications Administra-tion (EPT) (CEPT/ERC 70-03), Tromso, Norway.

Federal Communications Commission (FCC). (1999). Medical implant communications service (MICS) federal register. [Washington, DC, USA]. Rules Regulations, 64(240), 69926–69934.

Finite Difference Time Domain, X. (XFDTD). (2005). Electromagnetic solver based on the finite difference time domain method. State College, PA: Remcom.

Gabriel, S., Lau, R. W., & Gabriel, C. (1996). The dielectric properties of biological tissues: III. Para-metric models for the dielectric spectrum of tissues. Physics in Medicine and Biology, 41, 2271–2293. doi:10.1088/0031-9155/41/11/003 PMID:8938026.

Gosalia, K., Lazzi, G., & Humayun, M. (2004). Investigation of microwave data telemetry link for a retinal prosthesis. IEEE Transactions on Micro-wave Theory and Techniques, 52(8), 1925–1932. doi:10.1109/TMTT.2004.832007.

Guillory, K., & Normann, R. A. (1999). A 100-chan-nel system for real time detection and storage of extracellular spike waveforms. Journal of Neuro-science Methods, 91, 21–29. doi:10.1016/S0165-0270(99)00076-X PMID:10522821.

High Frequency Structure Simulator (HFSS). (1008). Ver. 11, Ansoft Corporation, Pittsburgh, PA.



Huang, W., & Kishk, A. A. (2011). Embedded spiral microstrip implantable antenna. Hindawi Interna-tional Journal of Antennas and Propagation, 1-6.

IEEE. (1999). IEEE Standard for safety levels with respect to human exposure to radiofrequency electromagnetic fields, 3 kHz to 300 GHz. IEEE Standard C95.1-1999.

IEEE. (2005). IEEE Standard for safety levels with respect to human exposure to radiofrequency electromagnetic fields, 3 kHz to 300 GHz. IEEE Standard C95.1-2005. International Telecommunica-tions Union-Radiocommunications (ITU-R). Radio regulations, section 5.138 and 5.150. ITU, Geneva, Switzerland. Retrieved from http://itu.int/home

Karacolak, T., Hood, A. Z., & Topsakal, E. (2008). Design of a dual-band implantable antenna and development of skin mimicking gels for continuous glucose monitoring. IEEE Transactions on Micro-wave Theory and Techniques, 56(4), 1001–1008. doi:10.1109/TMTT.2008.919373.

Kiourti, A., Christopoulou, M., & Nikita, K. S. (2011a). Performance of a novel miniature antenna implanted in the human head for wireless biotelem-etry. IEEE International Symposium on Antennas and Propagation.

Kiourti, A., & Nikita, K. S. (2011b). Meandered versus spiral novel miniature PIFAs implanted in the human head: tuning and performance. In Proceed-ings of the 2nd ICST International Conference on Wireless Mobile Communication and Healthcare.

Kiourti, A., & Nikita, K. S. (2012a). A review of implantable patch antennas for biomedical telem-etry: challenges and solutions. IEEE Antennas and Propagation Magazine, 54(3), 210–228. doi:10.1109/MAP.2012.6293992.

Kiourti, A., & Nikita, K. S. (2012b). Miniature scalp-implantable antennas for telemetry in the MICS and ISM bands: design, safety considerations and link budget analysis. IEEE Transactions on Antennas and Propagation, 60(6), 3568–3575. doi:10.1109/TAP.2012.2201078.

Kiourti, A., & Nikita, K. S. (2012c). Numerical as-sessment of the performance of a scalp-implantable antenna: Effects of head anatomy and dielectric pa-rameters. Wiley Bioelectromagnetics. doi:10.1002/bem.21753.

Kiourti, A., & Nikita, K. S. (2012d). Performance of miniature implantable antennas for medical telemetry at 402, 433, 868 and 915 MHz. In Proceedings of the 3rd International ICST Conference on Wireless Mobile Communication and Healthcare (MobiHealth 2012).

Kiourti, A., & Nikita, K. S. (2012e). Detuning issues and performance of a novel implantable antenna for telemetry applications. In Proceedings of the 6th European Conference on Antennas and Propagation (EuCAP 2012).

Liu, W. C., Chen, S. H., & Wu, C. M. (2008). Implantable broadband circular stacked PIFA an-tenna for biotelemetry communication. Journal of Electromagnetic Waves and Applications, 22(13), 1791–1800. doi:10.1163/156939308786375181.

Notis, D. T., Liakou, P. C., & Chrissoulidis, D. P. (2004). Dual polarized microstrip patch antenna, reduced in size by use of peripheral slits. In Pro-ceedings of the European Conference on Wireless Technologies.

Sani, A., Alomainy, A., & Hao, Y. (2009). Numeri-cal characterization and link budget evaluation of wireless implants considering different digital hu-man phantoms. IEEE Transactions on Microwave Theory and Techniques, 57, 2605–2613. doi:10.1109/TMTT.2009.2029770.

Scanlon, W. G., Burns, J. B., & Evans, N. E. (2000). Radiowave propagation from a tissue-implanted source at 418 MHz and 916.5 MHz. IEEE Transac-tions on Bio-Medical Engineering, 47, 527–534. doi:10.1109/10.828152 PMID:10763298.

Scanlon, W. G., Evans, N. E., & McCreesh, Z. M. (1997). RF performance of a 418 MHz radio tele-meter packaged for human vaginal placement. IEEE Transactions on Bio-Medical Engineering, 44(5), 427–430. doi:10.1109/10.568919 PMID:9125828.

Schmid, G., Neubauer, G., Illievich, U. M., & Alesch, F. (2003). Dielectric properties of porcine brain tis-sue in the transition from life to death at frequencies from 800 to 1900 MHz. Bioelectromagnetics, 24, 413–422. doi:10.1002/bem.10122 PMID:12929160.

Shiba, K., Nukaya, M., Tsuji, T., & Koshiji, K. (2008). Analysis of current density and specific absorption rate in biological tissue surrounding transcutaneous transformer for an artificial heart. IEEE Transac-tions on Bio-Medical Engineering, 55, 205–213. doi:10.1109/TBME.2007.900550 PMID:18232363.



Shults, M. C., Rhodes, R. K., Updike, S. J., Gil-ligan, B. J., & Reining, W. N. (1994). A telemetry-instrumentation system for monitoring multiple subcutaneously implanted glucose sensors. IEEE Transactions on Bio-Medical Engineering, 41(10), 937–942. doi:10.1109/10.324525 PMID:7959800.

Soontornpipit, P., Furse, C. M., & Chung, Y. C. (2005). Miniaturized biocompatible microstrip antenna using genetic algorithm. IEEE Transactions on Antennas and Propagation, 53(6), 1939–1945. doi:10.1109/TAP.2005.848461.

Sullivan, D. M., Borupa, D. T., & Gandhi, O. P. (1987). Use of the finite difference time domain method in calculating EM absorption in human tissues. IEEE Transactions on Bio-Medical Engineering, 34(2), 148–157. doi:10.1109/TBME.1987.326039 PMID:3557494.

Wegmueller, M. S., Oberle, M., Kuster, N., & Fichtner, W. (2006). From dielectrical properties of human tissue to intra-body communications. In Proceedings of the World Congress on Medical Physics in Biomedical Engineering.

Wessels, D. (2002). Implantable pacemakers and defi-brillators: Device overview and EMI considerations. In Proceedings of the IEEE International Symposium on Electromagnetic Compatibility.

Wiart, J., Hadjem, A., Wong, M. F., & Bloch, I. (2008). Analysis of RF exposure in the head tissues of children and adults. Physics in Medicine and Biology, 53, 3681–3695. doi:10.1088/0031-9155/53/13/019 PMID:18562780.

Xu, L., Meng, M. Q. H., Ren, H., & Chen, Y. (2009). Radiation characteristics of ingestible wireless de-vices in human intestine following radio frequency exposure at 430, 800, 1200, and 2400 MHz. IEEE Transactions on Antennas and Propagation, 57, 2418–2428. doi:10.1109/TAP.2009.2024459.



Konstantina S. Nikita (M’96–SM’00) received the Diploma in electrical engineering (1986) and the Ph.D. degree (1990) from the National Technical University of Athens (NTUA), Greece. She then received the M.D. degree (1993) from the Medical School, University of Athens, Greece. From 1990 to 1996, she worked as a Researcher at the Institute of Communication and Com-puter Systems, NTUA. In 1996, she joined the School of Electrical and Computer Engineering, NTUA, as an Assistant Professor, and since 2005 she serves as a Professor, at the same School. Her current research interests include biological effects and medical applications of radiofre-quency electromagnetic fields, medical imaging, biomedical signal and image processing and analysis, biomedical informatics, simulation of physiological systems. Dr. Nikita has authored or co-authored 160 papers in refereed international journals and chapters in books, and over 280 papers in international conference proceedings. She is author or co-author of two books (Simulation of Physiological Systems, Medical Imaging Systems) in Greek and co-editor of one book (Computational Electromagnetics: State of the Art and Future Trends) in English published by Springer. She holds two patents. She has been the technical manager of several European and National Research and Development projects in the field of biomedical engineering. She has been honorary chair/ chair/member of the program/organizing committee of more than 70 international conferences on the same fields (IEEE-IST, IEEE-BIBE, IEEE-ITAB, ICST-Mobihealth etc). She has served as keynote/invited speaker at several international conferences, symposia and workshops organized by NATO, WHO, ICNIRP, IEEE, URSI, COMCON, PIERS etc. She is member of the Editorial Board of the IEEE Transactions on Biomedical Engineering and guest editor of several international journals on biomedical engineering subjects (IEEE Transactions on Information Technology in Biomedicine, Computerized Medical Imaging and Graphics, IOP Measurement Science and Technology, etc). She has been the adviser of twenty completed PhD theses, several of which have received various awards. She has served as external evaluator in numerous university promotion committees and in international and national committees for grant proposal applications. Dr. Nikita has received various honors/awards, among which, the prestigious Bodossakis Foundation Academic Prize for exceptional achievements in “Theory and Applications of Information Technology in Medicine” (2003). She has been a member of the Board of Directors of the Hellenic National Academic Recognition and Information Center, a member of the Board of Directors of the Greek Atomic Energy Commission and a member of the National Council of Research and Technology. Dr. Nikita is a Senior member of the Institute of Electrical and Electronics Engineers (IEEE), a member of the Technical Chamber of Greece and a member of the Athens Medical Association. She is also the founding chair and ambassador of the IEEE-Engineering in Medicine and Biology Society, Greece chapter, vice chair of the IEEE Greece Section and deputy head of the School of Electrical and Computer Engineering of the NTUA.

design of implantable antennas for medical telemetry: dependence

Documents