Acta Biomaterialia 2 (2006) 531–536www.actamat-journals.com

Brief communication

Nanocomposite polymeric electrolytes to record electrophysiological brain signals in prolonged, unconventional or extreme conditions

Silvia Licoccia a,¤, M. Luisa Di Vona a, Paola Romagnoli a, Livio Narici b,Massimo Acquaviva c, Simone Carozzo c, Stefano Di Marco c, Moreno Saturno c,

Walter G. Sannita c,d, Enrico Traversa a

a Department of Chemical Science and Technology, Università di Roma Tor Vergata, Via della Ricerca ScientiWca 1, 00133 Rome, Italyb Department of Physics, Università di Roma Tor Vergata, Via della Ricerca ScientiWca 1, 00133 Rome, Italyc Department of Motor Science and Rehabilitation, Università di Genova, L.go Benzi 10, 16132 Genova, Italy

d Department of Psychiatry, State University of New York, Stony Brook, NY 11794, USA

Received 10 February 2006; received in revised form 2 May 2006; accepted 11 May 2006

Abstract

Chemically stable nanocomposite iono-conducting polymeric membranes (based on lithium salts and nanocrystalline oxide powdersdispersed in a polymethyl methacrylate matrix) performed successfully in the recording of human brain responses to visual stimulation.Impedance was higher than that of conventional electrodes. However, the electrophysiological signals recorded by acid Al2O3 and neutralAl2O3 5 wt.% and 10 wt.% nanocomposite gel electrolytes were comparable to those obtained with standard electrodes, even without pre-liminary skin cleaning and in the absence of gel electrolytes allowing better contact with and skin–electrode ionic conductance. The elec-trochemical and mechanical characteristics of these membranes make them Wt for human and animal research, for clinical application(speciWcally in emergencies, prolonged electrophysiological recordings), or in unconventional or extreme conditions when Xuid electro-lytes are unsuitable (e.g., biomedical space research).© 2006 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Nanocomposites; Biological polymers and assemblies; Electrodes; EEG; Visual evoked responses; Impedance

1. Introduction

Human brain signals (e.g., electroencephalogram [EEG],responses to sensory inputs) are recorded through surfaceelectrodes and a Xuid electrolyte (gel) interface with thescalp. The gel allows Xexible contact with the skin (mini-mizing the eVects of movement) and ionic conductance.Voltage, impedance, signal-to-noise ratio and, ultimately,the signal quality depend on the eYciency of the skin/gel/electrode contacts and interfaces, electrode homogeneity,

* Corresponding author. Tel.: +39 067 2594386; fax: +39 067 2594328.E-mail address: licoccia@uniroma2.it (S. Licoccia).

1742-7061/$ - see front matter © 2006 Acta Materialia Inc. Published by Elsedoi:10.1016/j.actbio.2006.05.007

gel stability, etc. [1]. Accurate cleaning of the skin is usuallynecessary to reduce impedance.

Conventional Xuid gels are unstable (Xuidity and ionicconductance can change over time) and may spread overthe scalp, therefore increasing the recording surface beyondcontrol. Standard procedures may therefore be diYcult toapply or unreliable in long-term recordings (emergencyroom monitoring, sleep, epilepsy, etc.), with uncooperativepatients, in emergencies, in extreme recording conditions,or in unconventional research applications.

Microgravity makes Xuid gels unsuitable in biomedicalspace research. Alternatives are mandatory for human elec-trophysiological research onboard the International SpaceStation, such as the studies in the anomalous long termeVects in astronauts project (ALTEA) [2,3]. This project

vier Ltd. All rights reserved.

532 S. Licoccia et al. / Acta Biomaterialia 2 (2006) 531–536

will allow systematic electrophysiological investigation ofthe human visual system in space; its goal is to investigatethe anomalous perception of light reported by astronautsever since the Apollo and Skylab missions [4,5]. This phe-nomenon is conceivably due to the direct action of heavynuclei on the retina or visual system [6], as suggested byearly experiments in accelerators [3] and measurements onthe Mir station [7–14].

Ionic transport in polymeric gel electrolytes (salt and aplasticizer, often a low molecular weight organic solvent,dissolved in a polymer matrix) is generally described as theionic species being moved throughout the electrolyte by themotion of polymer chains. Solvents improve ionic mobilityeither by favoring the segmental motion of polymer chainsor by acting as media for ionic transport [15,16].

We previously reported the development of polymericmembranes based on the dispersion of LiClO4 solutions in1,2-diethoxyethane (DEE) in a polymethyl methacrylate(PMMA) polymeric matrix: satisfactory results wereobtained as to electrochemical characteristics (conductivity,chemical stability, lack of toxicity, proper mechanical char-acteristics), ease of utilization and applicability for EEGrecording purposes, in spite of high (>100 k�) skin/elec-trode interface impedances [17].

It has been demonstrated that the development of nano-composite materials prepared by Wlling polymeric matrixeswith nanocrystalline oxides is eVective in improving theperformances of polymeric electrolyte membranes. Theapproach of using nanocomposite membranes seems to bequite versatile given that it was successful both for H+ andLi+ conductors and with diVerent types of polymers [18].Thus, we have now developed new iono-conducting com-posite solid electrolytes with the addition of nanosizedoxide particles, and successfully tested them in electrophys-iological measurements.

2. Materials and methods

All reagents (Aldrich) were reagent grade and were usedas received.

DiVerent nanocrystalline oxides were examined: acidic,neutral and basic Al2O3 (Aldrich, 6 nm) and TiO2 (12 nm,synthesized as reported elsewhere [19]) at concentrations of5 wt.% and 10 wt.%. The preparation of the membranesinvolved homogeneous dispersion of the selected ceramicpowder in the electrolyte solution, formed by LiClO4, DEE,and PMMA (MW 996 000) along with EtOH/H2O.

In a typical preparation, LiClO4 (0.85 g, 8 mmol) was dis-solved in DEE (8.4 mL). PMMA (2.28 g), EtOH (1 mL) andH2O (0.11 mL) were added, together with the appropriateamount of ceramic powder. The mixture was heated at70 °C until complete gelliWcation occurred (ca. 2 h). The gelwas then cast on a petri dish and cooled to RT for 24–48 hto achieve a self-standing Wlm. Average membrane thick-ness was 1 mm.

Ionic conductivity was measured at room temperatureby electrochemical impedance spectroscopy in the fre-

quency range 1 Hz to 100 kHz, using blocking type stainlesssteel electrodes. The measurements were carried out using aSolartron frequency response analyzer (model 1260) scan-ning over a 1 Hz to 100 kHz range. Data were processedwith appropriate Wtting software.

Conventional cortical responses to visual (contrast)stimulation were recorded simultaneously by standard der-mal electrodes and by nanocomposite polymeric electro-lytes. Healthy subjects with previous experience of theprocedures and methods used in this study served as volun-teers. Visual stimuli were vertical gratings with a sinusoidalluminance proWle subtending 9.0° (viewing distance: 75 cm;spatial frequency: 2.5 cycle/°; contrast: 80%; mean lumi-nance: 30 cd m¡2) that were reversed at 2.1 Hz for 29.9 s.Following proper scalp cleaning, Grass™ dermal electrodeswere positioned 5 cm laterally to the inion and 5% of theinion-to-nasion distance above the inion. The referenceelectrode was mid-frontal, the ground electrode at the ver-tex; Xuid gel was used for all electrodes. All of the nano-composite polymeric electrolytes developed in this studywere used as skin electrodes; membranes were positionedon the scalp a few millimetres apart from the conventionalelectrodes, without skin preparation and under the assump-tion of a reasonable topographic approximation. Impedancewas measured periodically. The system producing the stim-ulus (Neurotech, VENUS) also recorded signals on the 0.5–150.0 Hz bandpass. Each response was measured as theaverage of 100 responses free of artifacts; the system alsoprovided a standard measure of the signal-to-noise ratio,which gives an indirect, albeit reliable indication of therecording quality [20,21]. Stimulus conditions and recordingprocedures were consistent with the international guidelinesfor human visual testing [22].

Fig. 1. Complex impedance plane spectra for nanocomposite membraneswith various ceramic Wllers.

0 100 200 300 400 500 600 700 800 900 10000

100

200

300

400

500

600

700

800

900

1000

5% TiO2

5% acid Al2O3

5% basic Al2O3

5% neutral Al2O3

Zim

m[

Zreal [Ωcm]

Ωcm

]

S. Licoccia et al. / Acta Biomaterialia 2 (2006) 531–536 533

3. Results and discussion

As reported in a previous work [17], we have developediono-conducting polymeric membranes, based on lithiumsalts dispersed in a polymethyl methacrylate (PMMA)matrix, for electrophysiological measurements in space.The addition of water and ethanol as co-solvents to imple-ment the compatibility with biological Xuids and tissueswas eVective also in decreasing interfacial impedance. How-ever, we observed a degradation of gel membranes, with theinitial conductivity value (3.3£10¡3 S cm¡1) decaying twoorders of magnitude in two weeks, and membrane harden-ing due to solvent evaporation.

Nanocrystalline oxides were thus used as Wllers, based onthe eVects of inorganic Wllers in improving ionic conductiv-ity, mechanical strength and chemical/thermal stability of

polymeric electrolyte systems [23]. Moreover, the uniformdispersion of nanosize particles in polymers can lead to anultra-large interfacial area between the constituents per vol-ume of material, resulting in the development of new classesof materials with unique structure [24]. The addition ofnanocrystalline oxides improved the chemical stability ofthe gels, reducing the evaporation of the solvent chemi-sorbed to the oxide surface.

DiVerent nanocrystalline oxides were examined: acidic,neutral and basic Al2O3, and TiO2, at concentrations of5 wt.% and 10 wt.%.

Fig. 1 shows the complex impedance spectra for the vari-ous nanocomposite membranes. The conductivity valueswere quite large (10¡3–10¡4 S cm¡1) for all membranes.Doubling the amount of ceramic from 5 wt.% to 10 wt.%reduced conductivity to almost one half. The fast ion trans-

Fig. 2. Time evolution of the conductivity of nanocomposite samples containing 5 wt.% TiO2 (�) and 5 wt.% Al2O3 (�) and that of the reference ceramic-free sample (�). Data obtained at room temperature by electrochemical impedance spectroscopy (a). Selected complex impedance plane spectra for thetwo nanocomposites are shown in (b). Data have been normalized by dividing for the appropriate cell constant (l/A).

0 10 20 30 40 50 60

10-4

10-3

Con

duct

ivity

[Scm

-1]

Time [days]

0 200 400 600 800 1000

0

200

400

600

800

1000

200 400 600 800 1000

1 day

6 days

31 days

43 days

62 days

TiO2

1 day

6 days

15 days

43 days

63 days

Al2O3

Zim

m [

Ω c

m]

Zreal [Ω cm]

(a)

(b)

534 S. Licoccia et al. / Acta Biomaterialia 2 (2006) 531–536

port is promoted by the ceramic/polymer interactions, butat higher concentration lithium ion mobility is inhibited bythe larger content of oxide particles.

Fig. 2 shows the eVect of the Wller addition on the longterm stability of the membranes, where the conductivityvalues as a function of time measured for samples contain-ing 5 wt.% TiO2 and neutral Al2O3, respectively, are

reported together with those of a reference ceramic-freemembrane. While a rapid decay of conductivity values wasobserved for the ceramic-free sample after a few days, theconductivity for the nanocomposite gel was stable evenafter more than two months (Fig. 2(a)). No deviations fromlinearity were observed in the impedance spectra of allsamples after diVerent storage periods, therefore, suggesting

Fig. 3. Comparison of conventional average cortical response to visual stimulation as recorded from volunteers at scalp locations through surface skinelectrodes with standard Xuid gel (and proper skin preparation) or nanocomposite polymeric electrolytes. (A) Electrodes’ impedance measured over a 4 hperiod. (B) Averaged responses (amplitude spectra in inset). Each response is the average of 100 responses free of artifacts. The responses obtained inrepeated tests with standard thermal electrodes and gel are shown superimposed on top left to show variability. The signal-to-noise ratio is indicated byvertical bars on the bottom right of each response.

S. Licoccia et al. / Acta Biomaterialia 2 (2006) 531–536 535

that the integrity of the samples was maintained over time(Fig. 2(b)).

The mechanical properties were compatible with themembrane application to the skin, with proper plasticadaptation. The improvement can be attributed to thewater retention capability of the ceramic Wller depending,among other factors, upon the concentration and particlesize of the oxide.

The nanocomposites developed proved reliable in therecording of standard human responses to visual stimula-tion as used in clinical routines and vision research. Theresults are reported in Fig. 3 in comparison with conven-tional skin electrodes. The skin/electrode interface impedanceof conventional electrodes (with gel) was in the 10–20 k�range after skin preparation (coeYcient of variability overrepeated measures: »23%); the impedance of electrolytesmembranes (without skin preparation) was higher but rela-tively stable over a 4 h testing period, with values of about60 k� and 70 k� for membranes containing 5 wt.% TiO2and alumina, respectively (coeYcient of variability: »12–39%). The signal-to-noise ratio was higher with nanocom-posites than with conventional electrodes, but still in arange acceptable for clinical testing and human research[22]. Despite high impedance, distortion of the waveformand frequency organization of cortical visual responsesrecorded through acid Al2O3 and neutral Al2O3 5 wt.% and10 wt.% nanocomposite gel electrolytes was qualitativelynegligible; the response amplitude over time from the stim-ulus was within 1 SD of the response obtained with standardgel. The response was substantially distorted when record-ing through TiO2 and basic Al2O3, with diVerences fromstandard response exceeding 2 SD and 3 SD, respectively(Fig. 3).

The results indicate that the best performances wereachieved when small particle size and high surface acidityimproved the water retention capability of the ceramicWller. The addition of the nanosized ceramic Wller was eVec-tive in improving the performance of the polymeric iono-conducting membrane.

4. Conclusions

In this study, chemically stable nanocomposite iono-conducting polymeric membranes, based on lithium saltsand nanocrystalline oxide powders dispersed in a PMMAmatrix, were developed. These membranes successfully per-formed in EEG recording and determination of visuallyevoked potentials. The membranes with nanosized oxideWllers showed improved long-term stability, ionic conduc-tivity, and mechanical properties compared to the corre-sponding membranes without oxide addition. Due to theirlow cost, ease of preparation, storage and use, stability, andquality of recorded signals, Al2O3 acid and Al2O3 neutral5 wt.% or 10 wt.% nanocomposites appear suitable forapplication in extreme recording conditions (e.g., in space)as well as in clinical routines or in human and animalresearch.

Acknowledgements

This work has been supported by ASI (Italian SpaceAgency) Project ALTEA, by MIUR (Italian Ministry forEducation, University and Research) and by the Depart-ment of Motor Science and Rehabilitation, University ofGenova, Italy.

References

[1] Geddes LA, Baker LE. Principles of applied biomedical instrumenta-tion. New York, NY: John Wiley and Sons; 1989. p. 386.

[2] Narici L, Bidoli V, Casolino M, De Pascale MP, Furano G, Modena I,et al. The ALTEA facility on the international space station. PhysMed 2001;12:255–7.

[3] Narici L, Bidoli V, Casolino M, De Pascale MP, Furano G, ModenaI, et al. ALTEA: Anomalous long term eVects in astronauts. Aprobe on the inXuence of cosmic radiation and microgravity on thecentral nervous system during long Xights. Adv Space Res 2003;31:141–6.

[4] Pinsky LS, Osborne WZ, Bailey JV, Benson RE, Thompson LF. LightXashes observed by astronauts on Apollo 11 through Apollo 17. Sci-ence 1974;183:957–9.

[5] Pinsky LS, Osborne WZ, HoVman RA, Bailey JV. Light Xashesobserved by astronauts on Skylab 4. Science 1975;188:928–30.

[6] Casolino M, Bidoli V, Morselli A, Narici L, De Pascale MP, PicozzaP, et al. Dual origin of light Xashes seen in space. Nature 2003;422:680.

[7] Charman WN, Dennis JA, Fazio GG, Jelley JV. Visual sensation pro-duced by single fast particles. Nature 1971;230:522–4.

[8] McAulay IR. Cosmic ray Xashes in the eye. Nature 1971;232:421.[9] Charman WN, Rowlands CM. Visual sensation produced by cosmic

ray muons. Nature 1971;232:574.[10] McNulty PJ. Light Xashes produced in the human eye by extremely

relativistic muons. Nature 1971;234:10.[11] McNulty PJ, Pease VP, Pinsky LS, Bond VP, Schimmerling W, Vos-

burgh KJ. Visual sensations induced by relativistic nitrogen nuclei.Science 1972;178:160–2.

[12] Budinger TF, Lyman JT, Tobias CA. Visual perception of acceleratednitrogen nuclei interacting with the human retina. Nature 1972;239:209–11.

[13] McNulty PJ, Pease VP, Bond VP. Visual phenomena induced by rela-tivistic carbon ions with and without Cerenkov radiation. Science1978;201:341–3.

[14] Avdeev S, Bidoli V, Casolino M, De Grandis E, Furano G, MorselliA, et al. Eye light Xashes on the MIR space station. Acta Astronaut2002;50:511–25.

[15] Bohnke O, Frand G, Resrazi M, Rousselet C, Truche C. Fast iontransport in new lithium electrolytes gelled with PMMA. 1. InXuenceof polymer concentration. Solid State Ionics 1993;66:97–104.

[16] Scrosati B, editor. Application of Electroactive Polymers. London:Chapman and Hall; 1993.

[17] Romagnoli P, Di Vona ML, Narici L, Sannita WG, Traversa E, Lic-occia S. Synthesis and characterization of novel ionoconductor gelsfor biomedical applications in space. J Electrochem Soc 2001;148:J63–6.

[18] Licoccia S, Traversa E. Nanocrystalline oxides improve the perfor-mances of polymeric electrolytes. In: Fecht HJ, Werner M, editors.The nano–micro interface: Bridging the micro and nano worldstogether. Weinheim: Wiley-VCH Verlag; 2004. p. 289–301.

[19] Traversa E, Di Vona ML, Licoccia S, Sacerdoti M, Carotta MC,Crema L, et al. Sol–gel processed TiO2-based nano-sized powders foruse in thick-Wlm gas sensors for atmospheric pollutant monitoring. JSol–Gel Sci Technol 2001;22:167–79.

[20] AminoV MJ. Electrodiagnosis in clinical neurology. fourth ed. NewYork, NY: Churchill-Livingstone; 1999.

536 S. Licoccia et al. / Acta Biomaterialia 2 (2006) 531–536

[21] Husar P, Berkes S, Gotze A, Henning G, Plagwitz KU. ImprovingSNR (signal to noise ratio) in multichannel EEG recording. BiomedTech (Berlin) 2002;47(Suppl 1):566–9 [in German].

[22] Deuschl G, Eisen A. Recommendations for the practice of clinical neu-rophysiology: Guidelines of the International Federation of ClinicalNeurophysiology. Amsterdam: Elsevier Science; 1999 [EEG Suppl 52].

[23] Croce F, Appetecchi GB, Persi L, Scrosati B. Nanocomposite poly-mer electrolytes for lithium batteries. Nature 1998;394:456–8.

[24] Vaia RA, Giannelis EP. Polymer nanocomposites: Status and oppor-tunities. MRS Bull 2001;26(5):394–401.