tavvJ
TRENDSin CognitiveScierc€s Vot.7No.l2 D6csmbe.2OO3 Sensorysubstitution and the human- machineinterface Paul Baeh-y-Ritaland Stephen W. Kercelz
l?e?.anmglE-9fonhopedics and RehabiritationMedicine, and BiomedicarEngineering, univorsity of wisconsin, MadisonWl 53706.USA 2Endogenous SystemsResearch Group, New EnglandInstitute, University ot NewEngland, Biddeford ME 04005,USA
Recent advances in the instrumentation technology ol advances have led to the possibility of new prosthetic sensory substitution presemed have new opportunities devices being potentially accessible at much lower cost to to dsv€lop systems for compsnsation of sansory loss, millions ofpatients. ThiB is Btimulating the interest ofboth ,n sensory substitution {€.g. ot sight or vestibular func. research groups and industry leading to the establish- tion), intormation trom an artificial roceptor is coupled ment of new research and development efrorts in many tho to brain via a human-machine interfaco. The brain countries. Thia recent exploeion of interest in sensorv is able to use this intormation in place of that usually substitution suggest€that now is a good time to review transmitted from an intaqt ssnss organ. Both auditory progress in the area. and tactilo systems show promiss for practical sensory substitution intartace sitos. This research provides Brain plasticity experimental tools tor examining brain plasticity and Brain plasticity can be defined as the adaptiee capacities has implications fot porceptual and cognition studios ofthe central nervous system - its ability to modify its own more gonerally. structural organization and fiuctioning, IlT. It permits atr adaptive (or a maladaptive) responseto functional demand, Persons who become blind do not lose tÄe capacity to Mechanisms of brain plasticity include neurochemical, see. Usually, they lose the peripheral sensory system sJmaptic,receptor, and neuronal structural changes[19_ 21]. (the retina), but retain central visual mechanisms. Plastic changes in functional representation (usually Similarly, persons who becomedeafor are without balance occurring in response to a combination of a need and usually lose only the peripheral structures relating to training) do not appear to change the original functional sound transduction (the cochlea) or positional orientation representation [22]. the mechaoisms of representational (the vestibular apparatus). The input fiorn a sensory known, but 9gfl,are -not they_probablyinclude the substitution system can reach many brain structures unmaar Box 1, PhilGophicrl considgrationsfor sensory sub€titution Are bllnd porson8 u3ing lactil€-vision sensory Eubstitution (TVSS) pre-position h€r hand in r€spohse to lh€ therapist's pro-positioningof actually s€6lng? H6il l4l and Morgan l5l suggest thät becaus€ blind h€rhand. Following thetraiectoryotth€ ball provld€sconflrmatoryand subj€cts aro being giv€n Bimilar intormalion to that which caus$ the corr€ctiv€information. This is fully conqstsnt witbllaelnan's notion of sightad to soe and are capable ot giving similat rcspons63, on6 is l€A üealtorcnce and rcafrercnce I47l d Merleru-Ponvi)otion ol the with littlo altornativebut to admit th6tthey ars ss6ing (and not meroly int€nrional arc 14a,491.Ths essenli€lbät..i..lh€t-bl|fi/d subjecis us€ 'se6ing' or having 6oms oth€r s6nsstion). the cam€ra data in ths intentional orc iust as a sightod porson woutd fn his €ssav Molyneux's Auestion lan exploration of th6 quo$ion posed in 1688 by William Molynoux to the philosophor John Lock€: Hurleyand Noö int€rpr€tsonsory substitution as a form of controlled 'TVSS would a p€ßon blind from birth be sble to distinouish visually a cubo synoathesia, and state that effocts a new o)d€rnsl intormodal from 3 glot € upon $rdd€n aoquisition of vision?), Morgan l5l of€rs two mdpping from diEtal souroes of v;susl inp,ut to frefiphoral tactil€ inputs basic argum€n$ for this position, First, the structural naturo of lhe and on to the sohatosensory conax, A3 6 result, the qualitatlv6 perceptualsystom doos not offer any criteria lor disringuishing s€€inO expressionot somatosensorycod€x atter adaptation appgorsto changs from not s6oing. Tho ho߀shos crab is offered ss an examplo ot s intsrmodally. lo lake on aspocts of lh€ visual cha.ac16rol normal biologicalsystgm with f€wer r6ceptorsthan most mammals but which qualitative expr€ssionsoI visuäl con6x' {['t6], pp. 142-14:t). can novenholOssse€,Morgsn'5 socond argumsnt concerns bohsvioral Sevoraf chapters in P€rc€ption et lntetmodelite,.dit€d by Proust [50], squiv€lonc€;i{ blind sobiects r6c6ive (optical conversions ofl optical €xplore the classicalMolyn€ux qs€rstion.P6ch€ri€ l51l conclsd€s that intormation lhat would satlsfy crlteria lor seoing in ths slghled. and tho TVSS 6xperimenb permit a positive answ€r to the Molyneux r€spond in an indistinguish6bls m6nn6r, one probably haa to concode qu65lion, She considersthat TVSS offers the pos6ibility ots€parating rbatth€ blind a1666€ing. He consid€r€dthat'Th€r6 is notths slight€3t 69nsdlons from percoption:visualD€rceplion is not n€cessarilybased rsa60nto thinklhat lf two sensorym6Bsäg6s givethe animal sxsctly th€ upon visual s6nsations. She notesthat porc€ption i! an sctivo procoss, sam€ inlormalion, and l€ad to sxactly th€ sam€ b€havior, thoy will bo in Whlch movemenl plays a pan, Howev€r,w6 would add a cav€alifor I I percoived diff€r€ntly-ovon if they coms ovsr complelely different .ccu rsl6 id€ntilication and localizationin space,th€ blind personwould ll palhways' (151,p. 207). haveto bs 6ble to control the movehent, which t6k63minimal training I i Blind TVSS subj€ct6 do us€ the informbtion in the sam6 wey by TVSS sub,ects. We concluded from our TVSS studi€s th6t th€ i I psopl€ that sighted do [461.For €xample, lh€y sxhibit apDropristo mov€m€nt cän under cenain conditions bs im6gin6d r6thsr than I : anticipatory behsvioß. In s ball-rolling task. ths blind subj6ct will exscutsd änd stiff rgsuft in accuftt6 perc€ptlon [521. l point of contact with ar object provides a great deal of height with pitch and brightnese with loudneeein a left-to- practical information on object locatiou and identitrcation; right scan of any video frame. this yields a typical the HMI is at the hand where the receptors are activat€d regolution of 60 by 60 pixels at a rate of about one by the cane'e contact with an object. Stimulation of the ftame-per-second. hand is experienced at the end of the cane rather than in the hand. Tactilo-vision sribstitution: sesing vi! 3kin receptors In other worde, the sxperience is extemalized to the Tactile"vision Eensory substitution (TVSS) studies have point ofcontact between the object and the cane. [aically been carried out by numeroua reaearch groups for over a in sensory Bubstitution, tÄe HMI is simply a tran8ducer. It ceutüry. Many electro- and vibrotactile sensory substi- relays the information from artifcial sensors to the human tution HMIs have been developedwhich have been applied sensory interface, which carries the information to the to various surface ar.eas[26- 31]. In particular, the tongue brain. With experience, there is no perception of stimu- provides a practical HMI [32]. It is very sensitive and lation at the site of tüe HMI. highly mobile. Because it is in the protect€d environment of the moutfi, the sensory tecepto.s are close to the Auditoty-vision sensory substitution; seeing via tha sare eurface, The presence of an electrolytic solution, saliva, Substitutive sensory devices for blindness rehabilitation assures good electrical contact. The Tbngue Display Unit through auötion have been studied by D eYoldet et al. I24l . (Figure 1), or TDU (whictr we also identiS as a BrainPort Spatial localization and object recognition has been becauseofthe similarity to a computer USB port), has been demonstrated by Capelle et oL [25]. lbe syst€m couples developed to take advantage ofthese characteristics [331. a rough model ofthe human retioa with an inverse model The results obtained with a small electrotactile a|ray of tüe cochlea, using a püel-frequeDcy relationship. A developed for a study ofform prercepticn with a 6nger tip head-mount€d TV camera allorvs on.line translation of demonstrated that perception with electrical stimulation viaual pattema into sounds. With head or joystick moye- of the tongue is somewhat bett€r than with ffnger-tip nent, visual frames are grabbed at high-frequency and electrotactile stimulation, and the tongue requires only generate in real-time tlte corresponding complex Eounds -3?o of llhe voltage (6-16V), and much less current which allow recognition. (0.4-2.0 nA), than the ffnger-tip [32,34].the electrodee After training, both early blind (EB) and blinclfoldeil are separated by 2.34 mm (cent€r to center), The preseut sighted subjects were able ta identifu vieual forms [141. modelhas a72 x 72matrixofelectrode'. The elect otactile Ttained EB subjecls acquired rules of visual depth stimulus consists of 40-ps pulses delivered sequentially to p€!'ception and increased their performance it locallzation each ofthe active electrodes in tlrc patt€m. Bursts ofthree of distant object€. pulses each are delivered at a rate of50 Hz with a 200-Ilz Another approadr to auditory-vision substitution has pulse rat€ within a bü6t [32,34]. been taken by Meijer [12]. He developed a prototype Early studiesare rewiewedelsewhere [1,31,351. In most system for a general video to audio mapping, associating etudies, optical images picked up by a TV camera were TBENDSin Cosninve Scienes Vol.7No.12 D€comb.r2OO3 concentrat€ on these sen8ations. This does not lead to comitive overload (Box 2). poor resolution sensory substitution system can )\,/\ 1 orovide lhe information neceasary for the perception of \l (e g poor \-\-.---< compleximages. The inadequaciesof the skin two-point resolution) do not appear as serioue barriers to eventual high performaace, because the brain extracts informationlom the patterns ofstimulation lt is possible t recognize a face or to accomplish hanil-eye coordinated lu"ks ädt ottly a few hundred pointe of stimulation An Videocable exoeriment with etationary TVSS displaying the tactile mätrü oo the subject'eback [2] ehowedthatblind subjects were able to bat a ball as it rolled otra table at a point that had to be predicted by ttre blinil subject The subject had to identifr the rolling balt, estimate the tim€ it would teke-to reach ihe edge of tÄe table, estimate the position on the table, identif the location in his Vsual field of the bat' (which anil correctly time his movement of the bat was Uäing t"ot'ed) to bat the ball The experimental resultwitl"o.rrt""tty tuo;eU trabed blind subjectswas that the mean fRENDS h CenfltE S.ten .e ofthe performance was nearly pefect. Recent experiment€ with motivated subjects show that they can perform 'v'l6h El€clrd Floüo'tääli" I Schemslic of . t.ctil.'vßion sonlory substhudon |WSS) hand-'eye' coordination tasks $'ithin the fiist session' to rhs dorsum or thE loneus !iä fr6'ibrs 't6drod6 ",*"ir'";";"i'*'"0 to lh'iongu' displ'Y unlt ITDU) Y" Oi"""O. wnh conn€crion '6 bv o hssd moutrtod subStitution: balancing via ",t"t",ioOän caUre-nioesstno "".-tnout ot the mouth An lmsg€ cdptur'd Tac{ils-vestibularsensory vrä a vid6o cabl€ rh6 ääii*..i" 'r:,r"" ."r. .re rsnlmiftod ro th' mu puls6 trains 65ch corre skin recaptorc iOU .on*o" t'" uio"o in,o e patt'm ot 144 lN-voltsgs (BVD) experi€nce to tho olscttod6 srsv vi8 the with bilateral vestibular damage ip*ii"e t. pr""r' Th. pulso tralns 6ro carri€d Persons do6um of lh' pogtural \n'obblinCi' .iti". .,-.tt.. " una üt. sdhullte touch oshtors on ths functional difrcultiee that inclu'le a5 3n rm696 subl6d etp6ri6ncos"l*llod.3 rhe r$ulting str'6m of presents the longu6. Th6 "ns'iion g"it ana oscillopsia. lhie conötion """taUfeunique ofportunity to: (i) study a model of an open-loop (üt to re'establieh head- transduced intp a form of energy (vibratnry or direct- lrr^u., cottt"ot syet€m, and of vestibular sub8titution i"tt mediated by the skin receptors oostural control by means of "rl*utttlon) and an electrotactile "t""torr" of parts of the body (abdomen' back' t'high' i,rirrJu ft""a--o"Gd accelerometer üre recepton ltre use of Ä"r"*;o,""u"tut färehead).Vieual information reaches fttüitn"o.tgt the tongue sensory ""ä via produces a strong stabil- o"rleoti-l levels for analysis and interpretation u""iil"tu" änsory substitotion head and body coordination in subjects ,o-i""tttory pathways and structures A&er .training ization efiect on TVSS, subjects report experiencing images rn space' with BVD t111. with with great di6culty' tL etin. fUey tearn to make perceptual BVD patients can stand ouly i""t""a int€grate a range ol judgments"f "" visual means of interpretation' sucb as making enormous conscious efforLsto using acceler- paiallax, looming and zooming, and depth tactile cu€s. Using head-mounted il"Ip*tit", "i"".i'""a via a tongue HMI, subjects arc able to estimatesFl. ometere coupled source of data in a matter of secoDds' Once the subject has learned with oue motor sysEm aaapt m tfre "et" of the subjects ehow that t'he stability ,"... h-d-h"ld camera, using the correspondingkines- H"ai motion studies victim obtains from the acceleromet€r data ülti" the camera can be swikhed to another that the that of the pereon with normal vestibular "y"i"*1,L u head-mount€d), with no loss of perceptual apDroaches head motion studies show that it is Dot a """i"- Äa when the HMI' the electro- or vibrotactile function. Other (e g efect. "ä""ti".uiun. i" *ou"A fto* one area of ekin tß aoother from olacebo loss ü"iu.k o tr," ubdomenortothe forehead),there i8 no human-machinetochnologios spatial localization' even when the array is lmDlanted interfaces have provided th.e- "i-""*"" t"ä. back tD faont, because the ttained btind iJ*urr-*u"ttio" -majo.r to sensory substitutioD Although "irr"ir.a * not perceiving the image on the skin' but is challenges Practical has emphasized tactile sensory substitutioD' ""ii""t it correctlyin space[21 ifti" r"ä"* locating--S;üÄ are being studied' Mussa-Ivaldi an'l using'TVSS learn to treat the information orl"t t".lt"iqo"* several HMIs including EEG and skin in its proper context At one rDoment rvtti"" ttu"" rleviewed arrivi;s at the They emphasize t'hat feed- ai the skin has been gathered by intra-cortical techniques [36]. irr" uio'.r"ti"" for iearning snd control' requiring the ".tivingbut a[ another it relates to tlre usual u-""[ L ä" fV , "*a"a of a 'clo8ed loop'. Comparably, Suaning ".-"tinformation (pressure' tickle, wetness' etc )' ""t.Ufi"tt-""t."J iou"U have developed a 100-channel implanted ",riurr.orr" i" when h€/she scratches hie/her tfr" *Uj""t 'seen' technique [3?]' "ot matix"onfueed; nothing is Even during neurostirnulation--- U".f u"a". the al. bavl studied high-ilensif multi- sensory system, the subject lli"ofai" et iu"t o""fot-u""" with the from large numbers of when agked to u...y" u" toola to record iu"""i"" purely tactile sensationa "l""trod" "u"- hirpJ ic.,tEnd6rom M TBENDsin cognitivl sciarcds vot,7 No.12 o€c€mbor2003 8or( 2. Sen3ory ovo oad Normal ssnsory sysa€ms do nol usually ovsdoad, The central nervous thg thir,gs ths oy6 crn behold, but wo wguld nevor hav€ be€n able to säy system is ablo to s€l6ct only th6 information ne€dod to deal with th6 thal it wos po$ibl6 to det6rmin6 the id€ntity snd layout in thr€€ siluation atfordedbvthepaniculsrcont€xtInwhich itfindsitself.In 1970 dim€nslons of ö group of fomiliar obiocts if this system hed not b6en wefound thst manv€ttempts to creatosonsory aids led tothoproblem designedto deliver 400maxim ally discriminabl€E€ngations to the skin.' of sensoryoverload l3l. This was the result oI sottino out to provide ä s€t Ind€€d, tho perceptual sylt€ms ol living orodnisms €re the most of maximally discriminable Bonsations, which slmost imm6diately rernarkabls information-reduction machinß known. Th6y ars not reveal€da sharp limitstion in the ral€ at which poople can copo with seriouElylmpair6d even in sltuation6 wher6 a large proponion of th€ incoming intormarion when n is presented serially. Visual pärception input is liltered out or ignored, but th6y B16invariably handicapped thrives whsn lt is floodod whh intormation, torexample, giv€n a whole wh6n th€ inpul isveryimpovorishsd orartificially encod6d.Thus,it was pag€ ol text or a whol6 scone; how€ver. lt talters when the input is cl€ar ln 1970 lhst gvidonco of ov€rloöd reflect€d, at l6ast to som€ diminish6d ortemporally disiolnt6d: whon it isforc€d to 16adon€ word degreo,us6 of inappropri€toor poordisplays mor€than s limitatlon In at a time, or views th6 world through a calindicaltub€. At thal tilte,we tho inlormalion-hendling csp€citis6 of the ssnsory systom oI th6 statedrhat,'ltwould b€ rashto oredict rhat th6 skin willbe abl€ lo s66all perc6iverI3l. individual newone in the brain, as required to understand through the tongue cal provide clues regarding neural complex neuronal processing [381.Furt]rer, such technol- mechanisms of sensory substitution. Power spectra of tle ogy may enable clirect braia-to-computer interfacing for head posturogram in uormal subjects consist ofpattern8 of control of prosthetics or other devices. It has been dücrete peake in a range fiom 1-15 Hz. In BVD patients challenging to maintain functionality of large numbers the normal spectral pattern was severely depressedin the ofelectrodee for long periods oftirne itr behaving animals. 2-15 Hz fiequency range, and the power in the spectra Between 96 aDd ?04 electrodes were implant€d in each of shifted to the 0,15-2.0 Hz tange. Shift t{ lower fiequency three animals. Single-unit recordings were obtained for as correlsted with tüe severity ofthe vestibular damage [451. long as 18 months from at least 58 neurons and at earlier In some BVD patients using the device, part of the high time points up ta 247 single neurons were recorded in a fiequency power spectrum ia restored, whereas in others, single sessioD. These reeults lepreaent a significant it is not- In the8e casesthe evidence suggeststhat the TDU advance in the ability to extract large amourts of has the efrect of canceling some of the noise in the process information fiom the nervous syatem. of rnaintaining balance with non-vestibular cues. Both Cochlear irnplants for ileaf persons have been demon- processessu8gest that the nervous system is int€grating strated to be practical [391.The techniques rcported above the TDU data. may lead to implanted sensory substitution systems for other sensorylosses. Gonclusions Sensory substitution studies have clemonstrated the Nsqral cortclates of sen3ory substitution capacity of the brain to adapt t,o inforrnation relayed In addition to the subjective reports anil obeewed from an artificial receptor via an auditory or tactile HMI. performance of the subjects experiencing eensory substi- With training and with motor control of the input by the tution, tlre phenomenon hae been investigated with subject, percepts are accurately identifed aud spatially functional neuroirnaging. PET studies of sensory substi- located. Thus, blind persons obtain visual information tution mainly rely on cerebral perfusion mapping to reflect resultiog in visual percepts (e.g. of a ball rolling across a brain activity because changes in synaptic activity in the table) and can produce appropriate motor reaponsea brain result in changes of local perfusion. Several PET (e.g. catching the bsll) with a vision subetitution syeterD studieg-demonstratedthat in cognitive task8, Buch as con8isting of a TV camera controlled by the subjects PraiäiÄAog and tactile discrimination tasks t40,411, movemeDts, the electroDic device that transforms the ( occipitg),Äctivation was found during pattem recoghition signals from the camera into signals appropriat€ for \s-dl6- early btind 142,431.Another recent study of blind tÄe sensory re{eptors at the HMI, and a tactil€ array on subjects whols{sü hours oftraining in lett€r recognition the tongue. Similarly, persone who have lost vestibular using a g[4ueßu]stitution syst€m, showed activatiou in i.rnction can regain balance with a tactile vestibular system; severall'visualjlreas, including the cuneus, fusiform in ttris case the input to the tongue HMI ie from an ggus, Nfferior medial and superior occipital gyri ta4l. accelerometerwhich transforms tilt into electrical signals. Although sensory substitution Btualiesstrongly support Although braia imaging and physiological activity corre- the capacity of the brain to reorganize [1,21, the actual lates of sensory substitution have been described, precise mechanisms have not been firmly established. BraiD qeural mechanisms have not been identiffed. imaging studies including PET acans and transcranial The field of sensory subetitution provides the opportu- magnetic stimulation in Braille, and both auditory and niw to develop practical devices for persons with gensory TVSS (see above), and evoked pot€ntial stualies [2] have loes. These will be cosmetically acceptable, miniaturized provided evidence for reorganizatioo. Receptoa up- and and inexpensive, and it is expected that they will be downregulation, both s;naptic and extras]'Daptic mech- widely used. Furthermore, as the experimenter has anisms have been proposed [2], but ffrm evidence is not yet control of the conditions of the sensory Bubstitution available, process, they are being developed as tools to explore Dynamic and spectrographic analysis ofhead posturo- perceptual, physiological and brain mechanism correlates gram accompanying functional vestibular substitution of a complex cognitive process, htpJ i6,l16nd6.6m 2003 IRENDS in Cognitive Sciancag vol.7 No.l2 D€csmb€r \EEE Intemationnl Worhshop on 1ofr Computins in In46trial present has led to the demonstration 2OOg Researchup to the Lpficatrbnr (Kercel. S W. ed )' pp 27-32, IEEE sensory substi- (19901 rmqanization of primsrv of t"U.Ui" p.ototypiot devices tüat use zz- läfins. wlf et al. F\rctioßat are t'hree cort€x in adult ovJl uonkevd after behaviodlv i"U"" t" ti"t"t" ti"t sensory function There First' and most ""-"tÄ"""., t"Äf" J N'urophrsiol' 63' a2-1o4 p"Ut" of future development "ti-,tl"tior, (2000r c€'"bral neoddni2ation ot ,3;;- "o"t"offa H.a.' Graimsn' J eds, ""ti"l""a to robust and relatively Pr$8 urs€nt. i! the development n*ii"" B*i" o"mag€, Oxford Universitv tn maLe "tu m€ntal imaserv of nJ"-oä"i"" iiplementatioas of the technologv rzel ix vlä"", lC. a sr' {20oi) Auditlrv r'rissered blind hunäm a t"ide range of patient'ssuferiog sensory \-- i"*l'* visual arsociation areas in oarlv ü."c"""ilte to "ft"o. beyond the restoration oflost senses' Neiroinoee 14, r29-r39 io"". Soooa, .o"i"g A real tiroc exp€ritnetriql pmtotrre ror same t€chnologv to expand r ru ä"*ir" i. J J. rrssal ü il p*tiulJ to use t'he of !.sioD r€habilitation ulins auditory 3lrbstiturion' "i"rra the use ofnight K- #;;"-;", io".ur, iti"t, for example, enabling IEEE nant BianQtt. Eng' 45' 1279-1293 - of skiD ""ntiu *ithout int€rfering wth normal vision' ze'x'rt""i".'i.- ut a. rzoirl Snart'Ibuch ausDonüation uoo"toru" of tk lhh. "irion a whole range of non- ;;;; with electrocutaneout disptav' Pd*zedina' fi;;u". A" hchnology enables EnuironnetLt dnd to ;";;; Hdp'a tnterfdar fot uttuat exferiments with human subjecte ". l""t"i* f"*-..t dleipemnr Svstens.IEEE pp 40-46 of brain plasticitv ancl adaptstion or t.he aMomen: ;il;-iö; unileretanding -zz i'ää"*t, ie" Nsws & Features on BioMedNet update afticlesand specialreports Everytwo weeks' start your day with IroMedNels own daily sciencenews, features, res€arch Opinion' Cell and CurrcntBiologY' Plus' subscribe to Zn)oy'eioUeäUet Uagazrnewhich containsfroe aniclesfrom lrcnd9 Current ConferenceReportor to get daily repons di'ect from major lifo sciencemeetings' httpi,//news.bmn'Gom Hereis whatyou will{ind in News& Features: Tod.Ys t'lews Dailv news and featuresfor life scientistg. Sig; up to receiveweekly email alerts at httpi//newsbmn com/alerts sp€cial Report lite sciences' s!""i"i i"-ä"pttt ,"pon on events of current imponance in the world o{ the Reearch UFdtte by laboratoryres€archers chogen bythe editorsof Brietcommentaryon the lat€sthot paperslrom acrosstheLife sciences,written their fields' tn. frenas ard öuu.nt Opinionsiori]nalt, and a panel of koy expefts in http;//updatebmn con/alerts öiä" rp to iu""iuu n".€a;h Upd;te email alerts on your chosensubject at BioMedNet Magazi''.e EioMedNotMagazineoffeßfresart|c|esfromTfends,cuÜentopinion,celtandBioMedNetlvewgwitha'ocusonissueso'genera|. Report' 8?bMed&etMagazinefeaturos Opinion-s' Forum scientificinterest. From tt. tatestsoox reuie*s; the rnostcu(ent Special piecesand much more in an easilyaccessible form6t. lt pä", ö."tur"n"" n"poner, Historicalieripectives, scienceand society is publishedevery 2 weeks afso providesexcatang revrews, news and feätures,and primary research.BioMedNetMagazine Sign up to receivewoeklY email alerts at http,/news bmn com/alerts ConfeionceRePorter report journarsts coverdoz€nsotsessions at maior conferenoes,providing a quick but comprehensive BioMedNo(s orpeftscience Reporter'seosY_to-."ad you missea.rar moäiffoiÄutiu" Gn conferenceoverview, Con{erence of what rnight have "n "tOinary summaries6re updated daily throughout tho meeting' Sign up to receiveemail alefts al http://newsbmn com/alens