Silicone made contractile dielectric elastomer actuators inside 3-Tesla MRI environment

2008 IEEE/RSJ International Conference on Intelligent Robots and Systems Acropolis Convention Center Nice, France, Sept, 22-26, 2008

Silicone Made Contractile Dielectric Elastomer Actuators Inside 3-Tesla MRI Environment Federico Carpi, Azadeh Khanicheh, Constantinos Mavroidis and Danilo De Rossi

Abstract—New actuators are greatly demanded today in order to develop magnetic resonance imaging (MRI)compatible mechatronic systems capable of extended and improved capabilities. They are particularly needed for MRIguided interventional or rehabilitation procedures. Actuators based on dielectric elastomers, a specific class of electroactive polymers, appear as suitable candidates for new MRIcompatible technologies, due to their intrinsic material properties and working principle. This paper presents the first investigation on the MRI compatibility of a recently developed linear contractile actuator made of a silicone elastomer. The assessed absence of any degradation of both the actuator electromechanical performance in the MRI environment and the quality of images acquired from a phantom demonstrated the MRI compatibility of the actuator. These results suggest the suitability of this soft actuation technology as a possible new entry in the class of MRI compatible mechatronic systems. Index Terms—Actuator, contractile, dielectric elastomer, electroactive polymer, folded, MRI-compatible, MRI compatibility, silicone.

M

I. INTRODUCTION

AGNETIC resonance imaging (MRI) has today become an essential powerful tool for several clinical applications and basic research studies. It is mainly used as a non-invasive and safe technique for both diagnostic imaging (especially well suited for soft tissues) and basic research in neurosciences. Nevertheless, different types of applications of the MRI technique, such as image-guided interventions and progress monitoring of rehabilitation therapies, are rapidly spreading as well. These kinds of applications require the use of specific operative mechanical systems, such as manipulators, probes or rehabilitation tools. Their actuation can be either passive (manual) or active (robotic), depending on the required functionalities. In all cases, the interventional or rehabilitation actuated system has to be introduced inside (or

F. Carpi is with the Interdepartmental Research Centre “E. Piaggio”, Faculty of Engineering, University of Pisa, Pisa, Italy (phone: +39-0502217064; fax: +39-050-2217051; e-mail: [email protected]). A. Khanicheh is with the Department of Mechanical and Industrial Engineering, Northeastern University, Boston, MA 02115 USA, (e-mail: [email protected]). C. Mavroidis is with the Department of Mechanical and Industrial Engineering, Northeastern University, Boston, MA 02115 USA, (e-mail: [email protected]) D. De Rossi is with the Interdepartmental Research Centre “E. Piaggio”, Faculty of Engineering, University of Pisa, Pisa, Italy (e-mail: [email protected]).

978-1-4244-2058-2/08/$25.00 ©2008 IEEE.

very close to) the MRI scanner. This poses a technical challenge in the case of robotic systems. In fact, the adopted actuators should be capable of withstanding the harsh magnetic environment and, at the same time, their functioning should not provide image artifacts. These two basic conditions define the field of so-called MRIcompatible mechatronics. Conventional electromagnetic motors are in general intrinsically not suited for this field of application, due to their working principle. Accordingly, different types of actuation technologies are currently being investigated as possible alternatives. In particular, although hydraulic, pneumatic and electrostatic solutions are being considered, ultrasonic piezoelectric motors show a good MRI compatibility, so that they have been the most favourite choice so far [1]. In order to develop new MRI-compatible mechatronic systems capable of extended and improved capabilities (especially characterised by lower specific weight, higher mechanical compliance, complete absence of acoustic noise, ease of scalability and significantly lower cost), the investigation of different and less-conventional actuation technologies has begun in the last few years. As an example, electrorheological fluids have recently been proven to offer a valid means to electrically modulate resistive forces inside a MRI environment [1,2]. As a different technology, so-called electroactive polymer (EAP) actuators [3] may consist of a really valuable choice, due to their intrinsic material properties and working principles, as described in the following section. In this context, the aim of this paper is to present results of the first investigation on the MRI compatibility of a new linear contractile actuator, recently developed with one of the most performing classes of EAP materials: the dielectric elastomers (DE). II. DE ACTUATORS: HIGH-PERFORMANCE EAP TECHNOLOGY Electroactive polymers represent a broad family of ‘smart materials’, whose electromechanical actuation properties are significantly studied today in order to develop ‘artificial muscles’ [3,4]. These materials are capable of responding to different types of electrical stimuli, showing a transduction of the input electrical energy into an output mechanical work. The transduction mechanisms and properties strictly depend on the specific EAP category considered, according to different physical/chemical effects. In particular, the EAP

137

family comprehends ionic-type materials (including gels, conducting polymers, ionic polymer-metal composites and carbon nanotubes) and electronic-type materials (including dielectric elastomers, liquid crystalline elastomers, piezoelectric polymers and electrostrictive polymers) [3,4]. All the EAP materials typically offer intrinsic high mechanical compliance, lightness, ease of processing and, in many cases, low costs. Nevertheless, each EAP category shows very different electromechanical properties, suitable for different actuation purposes [3,4]. Among all the EAP materials, dielectric elastomers offer at present one of the best combinations of overall performances [5,6]. In fact, they are basically the unique EAP technology capable of combining very large active strains at still considerable stresses, high energy densities, good efficiencies, fast response, high reliability and lifetime, along with low costs. All these appealing features are paid with a typical need of high driving voltages, in order to reach the required electric fields (usually of the order of 10100 V/µm). The electric field drives the DE actuation according to a simple basic mechanism, which can be described as follows. Let’s consider an elementary configuration of a DE actuator, consisting of a thin layer of an insulating elastic material sandwiched between a couple of compliant electrodes. By charging the electrodes with a high voltage, an electrically induced compression of the elastomer occurs at constant volume, providing a thickness squeezing and an area expansion [5,6]. By exploiting such a basic actuation mechanism, several types of DE actuators, shaped according to a large number of possible configurations, have been demonstrated [7]. Dielectric elastomers (and, more generally, all the EAP materials) permit to build actuators that consist just of polymers (without any ferromagnetic inclusion) and that reasonably appear as suitable candidates for new MRIcompatible technologies. The next section provides a brief overview on previously reported experiences about MRIcompatible DE actuators. Very few investigations on the MRI-compatibility of DE actuators have been described so far. In particular, to the best of our knowledge, just two examples have been reported. They both concern the development of radiofrequency (RF) reconfigurable coils for flexible imaging capabilities [8,9]. According to some MRI techniques, individual coils can be placed on the patient surface, in order to enhance the quality and the resolution of images acquired from a certain region of interest. The image quality depends on the size and/or the position of the coil; therefore, during an imaging session it can happen that traditional coils have to be moved and/or replaced, in order to improve image quality. This usually has to be done by removing the patient from the MRI scanner. For the sake of avoiding these operations, the use of coils being automatically reconfigurable through MRI-compatible actuators would

provide a great advantage. For this purpose, two solutions based on different DE actuators have been proposed [8,9]. The first adopts a circular actuator used to modify the area included within the RF signal carrier loop. The second solution relies on bi-stable (also called binary) actuators that can change both the area and the position of the coil loop. Magnetic resonance images acquired from water-filled phantoms by using these active coils demonstrated their MRI-compatibility [8,9]. As an observation, we mention here that bi-stable actuators are also currently being studied in order to develop a MRI-guided needle positioning device, intended for prostate cancer treatments [10]. Moreover, it is also worth stressing that these studies are actually the sole that have been reported so far, not only within the field of dielectric elastomers, but also among the entire EAP domain. III. CONTRACTILE FOLDED DE ACTUATORS The actuators presented in [8,9] are just two examples of possible configurations for DE devices. They are made of the most common dielectric elastomer material currently studied for actuation: it consists of an acrylic polymer available as a film, commercialised by 3M under the brand VHB 4905/4910. More generally, the state of the art of the DE technology offers today, as mentioned above, actuators based on a large number of different configurations and, frequently, of different materials as well [7]. Among them, linearly contractile actuators can have a certain importance for several applications. This is especially the case of applications that require (or would largely benefit from) functional actuation properties similar to those of contractile natural muscles. As an example, new types of envisaged mechatronic systems could take advantage of soft, lightweight and contractile linear actuators, capable of providing new actuation properties, currently precluded to any conventional motor technology. In particular, they could open new paradigms in the field of wearable mechatronic systems for rehabilitation, where at present no conventional actuation technology (typically too heavy, stiff and noisy) is capable of offering comparable performances. A simple configuration suitable to easily obtain a DE contractile actuator was recently described [11]. It is represented by the so-called folded actuator. This device consists of a monolithic DE strip, which is first coated with compliant electrodes and then folded several times, so as to obtain the compact structure shown in Fig. 1. Following the application of a high voltage difference between the two compliant electrodes, the thickness of the insulating layer is electromechanically squeezed along its entire length (with the exception of the folds). This thickness squeezing is concomitant with a surface expansion of the strip, due to the characteristic volume constancy of elastomeric materials. As a result, the device shows an axial contraction combined with lateral expansions [11]. For any given driving electric field (i.e. the ratio between the applied

138

voltage and the thickness of the elastomeric layer), the amplitude of the achievable contraction depends on both the electrical and the mechanical properties of the considered elastomer.

The blocking force was measured by arranging the device upright and by connecting it (through a Nylon wire) to a miniature S beam load cell (LSB200, FUTEK Inc, U.S.A.). The experimental set-up is presented in Fig. 2b. The necessary high voltages were generated by using a high voltage power

Fig. 1. Working principle of a folded DE actuator: (a) rest condition; (b) electrically contracted condition.

Silicone made prototypes of folded actuator have been recently demonstrated [11]. Possible uses of such actuators for mechatronic rehabilitation systems (or even for different applications) may require their operation within MRI environments. So far, no data on the MRI compatibility of silicone-made dielectric elastomer actuators are available. Although such a compatibility could be reasonably expected (due to the material properties), this, of course, has to be demonstrated. Accordingly, in this work a prototype sample of a silicone based folded actuator was fabricated and tested within a MRI scanner for the first time, as described in the following. IV. MATERIALS AND METHODS A. Fabrication of the Actuator A prototype actuator was fabricated by using a soft commercial silicone (TC-5005 A/B-C, BJB Enterprises Inc., U.S.A.) as a dielectric elastomer and a custom-made silicone/carbon-black mixture (CAF 4, Rhodorsil, France / Vulcan XC 72 R, Carbocrom, Italy) as a compliant electrode material. The sample was obtained from a 1 mm-thick silicone strip and was assembled according to the same procedure described in [11]. Fig. 2a shows a picture of the prototype actuator, which had a resulting active length (at rest) of 65 mm and an active cross-section (at rest) of 16 mm × 21 mm. B. Measurement of the Actuator Active Stress The blocking force actively generated by the actuator in response to different electric fields was measured. In particular, the maximum applied field was 8 V/µm, which represented a safe limit for the prototype, in consideration of the low dielectric strength of the specific silicone adopted for its fabrication [11].

Fig. 2. Prototype folded DE actuator (a) and experimental set-up for the blocking force measurements (b).

supply (Model 610C, TREK, U.S.A.). The collected force data were then used to calculate the active stress of the device, as the ratio between its blocking force and its active cross-sectional area. C. Testing of the MRI Compatibility of the Actuator In order to study the MRI compatibility of the actuator, the following two fundamental requisites for any MRIcompatible device were investigated: i) its capability of withstanding the strong magnetic fields of the MRI scanner and sensitive imaging sequences, without any significant performance degradation; ii) its capability of functioning without inducing any significant artifacts on the acquired images. Aimed at performing these evaluations, two types of tests were carried out. The first consisted in repeating the measurement of the stress-field characteristic for the actuator arranged inside the MRI scanner in action. The comparison of this characteristic with that obtained outside the scanner provided a useful indication about any eventual alteration of the actuator electromechanical performances. As a second type of test, a signal-to-noise ratio (SNR) investigation was performed on images acquired from a phantom, consisting of a container filled with a saline solution. The SNR comparison of the MRI images for different actuator performance conditions (different output stresses for different electric fields) with the control image (without the actuator in the MRI scanner) was used to investigate any possible degradation of the MRI images during the operation of the actuator. A 3-Tesla Siemens Trio whole-body MRI scanner equipped with 12-channel Siemens TIM head coil (Athinoula A. Martinos Center for Biomedical Imaging,

139

Massachusetts General Hospital, Boston, MA) was utilized for all tests conducted on the actuator. Images were collected using gradient-echo echo-planar imaging (GE-EPI), commonly used in functional imaging. Acquisition parameters were: TR/TE=2000/30 ms, voxel size (3.1mm)×(3.1mm)×(5.0mm), 128×128 acquisition matrix/ 200mm×200mm FOV, 33 slices, 90o flip angle.

applied electric field. Each electric field was applied for 120 seconds and the related average force with its standard deviation (error bars) is reported in Fig. 4. Both the stressfield curves show a typical parabolic trend, characteristic of any DE actuator [5,6]. No significant difference is evident from these data series for inside and outside MRI scanner tests, which suggests that the effect of surrounding magnetic field and EPI sequences on the 1.6

Active Stress (kPa)

1.4

Outside MRI Scanner Inside MRI Scanner: During EPI Imaging

1.2 1.0 0.8 0.6 0.4 0.2 0.0

Electric Field (V/μm)

The imaging object was a cylindrical phantom filled with a solution of 1.24 g NiSO4 × 6 H2O / 2.62 NaCl per 1000 g H2O. The control image was acquired without any device and the phantom was not moved during the entire tests. Fig. 3 shows the contractile actuator being used inside the scanner during phantom testing with EPI sequence. The power supply for the actuator and the computer used to read the force sensor data were located outside of the MRI room. The force data were recorded at a 20 kHz sampling rate and they were low-pass (Butterworth) filtered with 60 Hz cut-off frequency. The force sensor was shown to be MRI-compatible in previous studies [2]. To minimize electromagnetic interference, the wires were properly shielded and cables of appropriate size and impedance were used. The cables were shielded by a braided copper meshing and passed through the penetration panel into the shielded MRI room. The actuator was attached to the scanner table in the approximate 30 cm distance from the iso-center of the magnet.

3 4 5 6 7 8 9 Electric Field (V/μm) Fig. 4. Static driving of the actuator inside and outside the 3-Tesla MRI scanner: active stress versus electric field diagrams.

1

0

2

2

4

1.0

6

8

10

Time (s)

Inside MRI Scanner, No Imaging

12

14

Inside MRI Scanner EPI Imaging

0.8 0.6 0.4 0.2 0.0

0

2

4

6

8

10

12

14

Time (s) Fig. 5. Sinusoidal driving of the actuator inside and outside the 3-Tesla MRI scanner: (a) applied electric field; (b) active stress response.

V. RESULTS

a

Results of the MRI compatibility tests are separately presented below. A. Actuator Active Stress Outside and Inside the MRI Scanner Tests with both dc and ac driving signals for the actuator were performed, as described below. 1) Static driving Values of the active stress measured outside and inside the MRI scanner are plotted in Fig. 4 as a function of the

8 7 6 5 4 3 2 1 0

0

Outside MRI Scanner

Active Stress (kPa)

Fig. 3. Arrangement of the folded DE actuator inside the MRI scanner during phantom tests with EPI sequence.

b

b’

c

d

e

c’

d’

e’

Fig. 6. Effect of static activations of the actuator on phantom images: (a) control; (b) actuator @ 0 V/µm; (c) actuator @ 4 V/µm; (d) actuator @ 6 V/µm; (e) actuator @ 8 V/µm; (a'-d') subtraction of the control (a) from (be).

140

TABLE I SIGNAL STABILITY AND SNR FOR PHANTOM IMAGES ACQUIRED WITH EPI. Condition

SNR c

Signal (a.u.) a

200.77

2402.29±59.17

b

200.19

Actuator @ 1 V/μm

2402.40±58.71

b

200.20

Actuator @ 2 V/μm

2399.42±58.78 b

199.95

Actuator @ 3 V/μm

2404.06±59.11

b

200.34

Actuator @ 4 V/μm

2404.15±58.51

b

200.35

Actuator @ 5 V/μm

2405.36±59.48 b

200.45

Actuator @ 6 V/μm

2404.05±59.42 b

200.34

Actuator @ 7 V/μm

2403.73±59.35

b

200.31

2404.44±59.06

b

200.37

Control (no actuator) Actuator @ 0 V/μm

Actuator @ 8 V/μm

empirical width of the noise profile. SNR values upon the introduction and operation of the folded DE actuator into the MRI scanner are reported in Table I. In all cases, simple paired t-tests comparing condition to control (two-tailed, Pvalue=0.05) failed to reach significance, with P-values ranging from 0.06 to 0.98. Results show that, in all cases, the loss of SNR observed was not significant regardless of the introduction and operation of the folded DE actuator.

2409.29±59.41

Mean of signal ROI ± standard deviation of the mean; b P-values ranging from 0.06 to 0.98 (two-tailed t-tests comparing condition to control; threshold set at P=0.05. c SNR calculated dividing mean of signal ROI to width of distribution of background intensities. Histograms of background intensities (‘image noise’) were fitted with Gaussian distribution: σ = 12. a

performance of the actuator is negligible. The maximum active stress was found to be about 1.5 kPa for an electric field of 8 V/µm.

2) Sinusoidal driving The SNR was also calculated for phantom images acquired when the actuator was driven with the sinusoidal signal shown in Fig. 5a. In this case, the force sensor was replaced with a common thin elastic wire that connected the actuator to its supporting structure. Accordingly, by applying the sinusoidal signal, the actuator was able to contract and relax continuously. This permitted to investigate the MRI compatibility within an experimental condition (pseudo-isotonic test) different from that considered before (isometric test). As a result, once again no significant decrease was observed in the SNR, as shown in Fig. 7. This plot presents the SNR for one slice and over 18 acquisitions of the same slice, compared to the SNR of the control image (darker colour). 200

2) Sinusoidal driving Values of the blocking force outside and inside the MRI scanner were also measured in response to a sample sinusoidal signal, having a frequency of 1 Hz and an amplitude corresponding to an electric field of 7 V/µm. Fig. 5 shows the driving signal and the related stress response. As Fig. 5 shows, the performance of the actuator was not affected by the MRI environment and imaging sequences.

SNR

150

100

50

B. MRI images with and without the working actuator 1) Static driving Phantom EPI images were acquired first for the control (no device in the MRI scanner), as presented in Fig. 6a, then in the presence of the actuator and of its operation with different driving electric fields, as shown in Figs. 6b-6d. The phantom was not moved during the entire tests. To ensure that the introduction and operation of the folded DE actuator caused no degradation in the MRI images and did not cause any image artifacts, the variation of the SNR among the acquired images was investigated. In particular, the SNR was estimated in the image domain: the signal was calculated from the mean of a ROI (Region Of Interest) encircling the center of the image, while the width of the noise was computed by choosing a ROI near the image edge outside the object [12]. Image SNR was then computed from the following relation: SNR = S s bg , where S is the mean signal intensity in the central ROI and sbg, is the width of the histogram of the intensities of the background ROI, or the

0

Control

0

1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18

Number of Acquisitions

Fig. 7. SNR variation for phantom images acquired with EPI sequence during a sinusoidal activation of the actuator.

VI. DISCUSSION The presented results show two fundamental evidences. First, the electromechanical performances of the considered contractile actuator do not change when it is either dc or ac operated outside or inside the MRI scanner in action. Accordingly, the actuator capability of withstanding the strong magnetic fields of the MRI environment and the sensitive imaging sequences, without any measurable performance degradation, has been assessed. Second, no image artifacts and no significant decrease of the signal-tonoise ratio were obtained in the MRI images acquired from the phantom during the operation of the actuator inside the MRI scanner. The coexistence of these two basic outcomes demonstrates the MRI compatibility of the tested contractile DE actuator.

141

This result encourages further investigations aimed at developing MRI compatible mechatronic systems actuated by such a type of polymer devices. Lightweight and wearable structures appear to be suitable candidate systems that could benefit from an integration with such a new actuation technology. Examples of possible applications include haptic devices and robotic systems with force feedback for interactions with human motions. They could be used, for instance, for either basics investigations in neuroscience, virtual reality applied to medicine or rehabilitation exercises. The broad field of mechatronic systems for rehabilitation, in particular, is expected to represent one of the preferential study platforms for the next investigations in this field. In fact, as an example, orthotic dynamic systems for hand rehabilitation (hand splints) adopting the actuator type studied in this work are currently being studied [13]. MRI compatible versions of such wearable rehabilitation systems could represent a useful tool in order to follow rehabilitation trainings with MRI investigations. However, the development of this or any other possible application will also require future improvements of the electromechanical performances of the prototype actuators considered in this work. This could be achieved through the feasible adoption of most performing dielectric elastomers and improved fabrication procedures, as it is currently being pursued [13,14]. Nevertheless, regardless of any specific performance of the considered prototypes, the tests here presented permitted to assess, for the first time, the MRI compatibility of a silicone based soft actuation technology, which shows a great promise as a new entry in the class of MRI compatible mechatronic systems. In order to properly extend and deepen the investigations on the MRI compatibility of the broad family of DE actuators, future studies should evaluate the behaviour of devices capable of working even at greater electric fields, e.g. one order of magnitude higher than those allowed by the sample tested in this work. Such an investigation with variable electric fields at higher amplitudes could provide useful additional information on the electromagnetic compatibility between the actuation and the imaging system.

efficient electromechanical transduction, which are exclusive of the dielectric elastomer technology. Further investigations on devices working at higher electric fields are needed in order to extend and deepen the study of the MRI compatibility of the broad family of dielectric elastomer actuators. ACKNOWLEDGEMENTS The authors would like to thank Dr. Giorgio Bonmassar for providing help with MRI testing. REFERENCES [1]

[2] [3] [4] [5] [6] [7]

[8]

[9]

[10]

[11]

VII. CONCLUSIONS This work demonstrated the MRI compatibility of folded contractile actuators made of a silicone elastomer. For this purpose, the absence of any measurable degradation of both the electromechanical performance of the actuator and the quality of images acquired from a phantom was assessed. These results suggest the suitability of this type of actuators for the development of new generations of MRI compatible mechatronic systems. Specific applications of such systems could take advantage from the intrinsic properties of the actuation material, such as lightness, mechanical compliance and

[12] [13]

[14]

142

N. V. Tsekos, A. Khanicheh, E. Christoforou and C. Mavroidis, “Magnetic resonance–compatible robotic and mechatronics systems for image-guided interventions and rehabilitation: a review study”, Annu. Rev. Biomed. Eng., vol. 9, pp. 14.1–14.37, 2007. A. Khanicheh, A. Muto, C. Triantafyllou, B. Weinberg, L. Astrakas, A. Tzika and C. Mavroidis, “fMRI-compatible rehabilitation hand device”, J. Neuroeng. Rehabil., pp. 3-24, 2006. Y. Bar-Cohen, Editor, Electroactive polymer (EAP) Actuators as artificial muscles. Reality, potential, and Challenges – 2nd edition, Bellingham, Washington: SPIE Press, 2004. J. D. W. Madden et al., “Artificial muscle technology: physical principles and naval prospects”, IEEE J. Oceanic Eng., vol. 29, no. 3, pp. 706-728, 2004. R. E. Pelrine, R. D. Kornbluh and J. P. Joseph, “Electrostriction of polymer dielectrics with compliant electrodes as a means of actuation”, Sensors and Actuators A-Phys. vol. 64, pp. 77-85, 1998. R. Pelrine, R. Kornbluh, Q. Pei and J. Joseph, “High-speed electrically actuated elastomers with strain greater than 100%”, Science, vol. 287, pp. 836-839, 2000. F. Carpi, D. De Rossi, R. Kornbluh, R. Pelrine and P. SommerLarsen, Editors, Dielectric elastomers as electromechanical transducers. Fundamentals, materials, devices, models & applications of an emerging electroactive polymer technology, Amsterdam: Elsevier Press, 2008. J. Vogan, A. Wingert, J.-S. Plante, S. Dubowsky, M. Hafez et al., “Manipulation in MRI devices using electrostrictive polymer actuators: with an application to reconfigurable imaging coils”, Proc. of the IEEE Int. Conf. Robotics Automation, New Orleans, LA, pp. 2498–504, 2004. D. F. Kacher, J. Vogan, A. Wingert, M. Hafez, J.-S. Plante, S. Dubowsky, F. A. Jolesz, "Development of a reconfigurable MRI coil using electrostrictive polymer artificial muscle actuators", Proc. of the International Society of Magnetic Resonance in Medicine, Kyoto, Japan, p. 3394, 2004. J.-S. Plante, L. M. Devita, S. Dubowsky, “A road to practical dielectric elastomer actuators based robotics and mechatronics: discrete actuation”, In Y. Bar-Cohen, Editor, Electroactive Polymer Actuators and Devices (EAPAD) 2007, Proceedings of the SPIE, vol. 6524, pp. 652406-1 – 652406-15, 2007. F. Carpi, C. Salaris and D. De Rossi, “Folded dielectric elastomer actuators”, Smart Materials and Structures, Vol. 16, pp. S300-S305, 2007. M. J. Firbank, A. Coulthard, R. M. Harrison, and E. D. Williams, "A comparison of two methods for measuring the signal to noise ratio on MR images," Phys. Med. Biol., vol. 44, pp. N261-4, 1999. F. Carpi and D. De Rossi, “Contractile monolithic linear actuators”, In F. Carpi, D. De Rossi, R. Kornbluh, R. Pelrine and P. SommerLarsen, Editors, Dielectric elastomers as electromechanical transducers. Fundamentals, materials, devices, models & applications of an emerging electroactive polymer technology, Amsterdam: Elsevier Press, 2008. F. Carpi, G. Gallone, F. Galantini and D. De Rossi, “SiliconePoly(hexylthiophene) blends as elastomers with enhanced electromechanical transduction properties”, Adv. Funct. Mater., vol. 18, no. 2, pp. 235–241, 2008.