Defining the anterior nucleus of the thalamus (ANT) as a deep brain stimulation target in refractory epilepsy: Delineation using 3 T MRI and intraoperative microelectrode recording

Defining he an e io nucleus o he halamus (ANT) as a deep b ain s imula ion a ge in e ac o y epilepsy: Delinea ion using 3 T MRI and in aope a i e mic oelec ode eco ding T. Mö önen a, ⁎,J.Ka isko b , J. Haapasalo a , T. Täh inen a , T. Kieka a c ,V.Kähä ä c ,J.Pel ola a , J. Öhman a , K. Leh imäki a a Depa men o Neu osciences and Rehabili a ion, Tampe e Uni e si y Hospi al, P.O. Box 2000, Tampe e 33521, Finland b Depa men o Neu osu ge y, Oulu Uni e si y Hospi al, P.O. Box 21, Oulu 90029, Finland c Medical Imaging Cen e, Tampe e Uni e si y Hospi al, P.O. Box 2000, Tampe e 33521, Finland abs ac a icle in o A icle his o y: Recei ed 30 No embe 2014 Recei ed in e ised o m 3 Feb ua y 2015 Accep ed 1 Ma ch 2015 A ailable online 5 Ma ch 2015 Keywo ds: An e io nucleus Deep b ain s imula ion Epilepsy Magne ic esonance imaging Thalamus Backg ound: Deep b ain s imula ion (DBS) is a minimally in asi e and e e sible me hod o ea an inc easing numbe o neu ological and psychia ic diso de s, including epilepsy. Ta ge ing poo ly defined deep s uc u es is based in la ge deg ee on s e eo ac ic a las in o ma ion, which may be a majo sou ce o inconsis en ea men e ec s. Aim o he s udy: In he p esen s udy, we aimed o s udy whe he a ecen ly app o ed a ge o epilepsy (an e io nucleus o halamus, ANT) is isualized in clinically es ablished 3 T MRI and whe he ANT is delinea ed using in aope a i e mic oelec ode eco ding (MER). We ha e especially ocused on indi idual a ia ion in he loca ion o ANT in s e eo ac ic space. We also aimed o demons a e he ole o indi idual a ia ion in in e p e- a ion o MER da a by p ojec ing samples on o AC–PC (an e io and pos e io commissu e) and ANT-no malized coo dina e sys ems. Me hods: De ailed analysis o ANT delinea ions in 3 T MRI sho au in e sion eco e y (STIR) images om eigh pa ien s unde going DBS o e ac o y epilepsy was pe o med. Co onal and sagi al c oss-sec ional models o ANT we e plo ed in he AC–PC coo dina e sys em o s udy indi idual a ia ion. A o al o 186 MER samples collec ed om 10 DBS ajec o ies and 5 pa ien s we e analyzed, and he loca ion o each sample was calcula ed and co ec ed acco dingly o he loca ion o he final DBS elec ode and p ojec ed o he AC–PC o coo dina e sys em no malized o ANT. Resul s: Mos o he key s uc u es in he ana omic a las a ound ANT (mammillo halamic ac and ex e nal medulla y lamina) we e iden ified in STIR images allowing isual delinea ion o ANT. We obse ed a high deg ee o ana omical a ia ion in he loca ion o ANT, and he c oss-sec ional a eas o e lapped by s udy pa ien sdec easedinalinea ashionwi haninc easingnumbe o pa ien s.MERin o ma ion om10in- di idual ajec o ies co ela ed wi h STIR signal cha ac e is ics by demons a ing a spike-nega i e zone, p esumably whi e ma e laye , a he la e al aspec o ANT in ANT-no malized coo dina e sys em as p e- dic ed by STIR images. Howe e , MER in o ma ion p ojec ed o he AC–PC coo dina e sys em was no able o delinea e ANT. Conclusions: ANT is delinea ed in 3 T MRI by isualiza ion o a hin whi e ma e lamina be ween ANT and o he nuclea g oups ha lack spiking ac i i y. Di ec a ge ing in he an e io halamic a ea is supe io o indi ec a ge ing due o ex ensi e indi idual a ia ion in he loca ion o ANT. Wi hou de ailed imaging in- o ma ion, howe e , a single ajec o y MER has li le localizing alue. © 2015 The Au ho s. Published by Else ie Inc. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). 1. In oduc ion Deep b ain s imula ion (DBS) is a widely used minimally in asi e and e e sible me hod o ea a ious neu ological and psychia ic diso de s by s imula ing deeply loca ed s uc u es in he b ain, and i s use is es ima ed o expand in he u u e. Implan a ion o elec odes in o a a ge no isible in cu en imaging me hods can be pe o med using s e eo ac ic p inciples, whe e he a ge s uc u e loca ion is Neu oImage: Clinical 7 (2015) 823–829 Abb e ia ions: AC, an e io commissu e; ANT, an e io nucleus o halamus; DBS, deep b ains imula ion;MCP,mid-commissu alpoin ;MER,mic oelec ode eco ding;PC,pos e- io commissu e;STIR,sho auin e sion eco e y;3D-T1W,3-dimensional(mul i-plana econs uc ion)T1-weigh edimage;T2W,T2-weigh edimage;VA, en alan e io nucleus o halamus. * Co esponding au ho a : Depa men o Neu osciences and Rehabili a ion, Tampe e Uni e si y Hospi al, P.O. Box 2000, Tampe e 33521, Finland. Tel.: +358 3 311 69410. E-mail add ess: imo.mo [email protected] (T. Mö önen). h p://dx.doi.o g/10.1016/j.nicl.2015.03.001 2213-1582/© 2015 The Au ho s. Published by Else ie Inc. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). Con en s lis s a ailable a ScienceDi ec Neu oImage: Clinical jou nal homepage: www.else ie .com/loca e/ynicl fi s defined in he a las b ain ela i e o common isible ana om- ic landma ks such as he an e io commissu e (AC) and pos e io commissu e (PC), and he same coo dina es ela i e o he AC–PC line a e used o access he non- isible a ge in he pa ien b ain (indi ec a ge ing me hod). Due o he de elopmen o imaging echniques, indi ec a ge ing has mos ly been eplaced by di ec a ge ing based on isualiza ion o he desi ed a ge s uc u e in he pa ien 3s images. In heo y, he use o indi ec (AC–PC based) a ge ing is limi ed due o po en ial ana omical a ia ion in he loca ion o he a ge s uc u e be ween indi iduals and s e eo ac ic a las in o ma ion, whe eas he di ec a ge ing me hod is limi ed due o a lack o adequa e a ge specific MRI imaging me hods. Ad ances in sophis ica ed imaging modali ies ha e led o he eplacemen o indi ec a ge ing wi h di ec a ge ing in mos indica ions. Howe e , due o a lack o di ec imaging me hods, indi ec a ge ing is s ill widely used in DBS o he an e io nucleus o he halamus (ANT). DBS o he an e io nucleus o he halamus (ANT) has been sug- ges ed as a ea men op ion in e ac o y ocal epilepsy by a la ge- scale andomized con olled ial (Fishe e al., 2010)andse e al open-label pilo s udies (Hodaie a al., 2002;Ke igan e al., 2004; Lee e al., 2012). Howe e , ANT is poo ly iden ified in cu en ly used MRI echniques adop ed om mo emen diso de su ge y. A ecen epo sugges s ha ANT bounda ies a e iden ified in 3 T MRI T1 images (Buen jen e al., 2014), mos likely due o isualiza- ion o he whi e ma e laminae a ound he ANT. Mo eo e , isual- iza ion o he whi e ma e laminae in he halamus by 3 T MRI sho au in e sion eco e y (STIR) images has been sugges ed (Yamada e al., 2010). ANT, wi h i s ex ensi e connec ions o he subicula co ex, cingulum, mammilla y bodies and e osplenial co ex, is an in eg al pa o he hippocampal sys em o episodic memo y, and is also hough o be in ol ed in execu i e and emo ional unc ions. Animal s udies ha e poin ed o ANT ha ing a ole in he gene aliza ion o ocal co ical seizu es. (Child and Bena och, 2013) Animal models ha e also p o en ha high- equency s imula ion o ANT can inhibi induced ic al ac i i y in he hippocampus (S ypulkowski e al., 2011). The aim o he p esen s udy is o s udy he delinea ions o ANT using 3 T MRI STIR sequence and o p o ide in o ma ion abou he sizeandshapeo ANTaswellas omeasu e hedeg eeo ana omical a ia ion/o e lap in he loca ion o ANT in an AC–PC based coo di- na e sys em be ween indi iduals. Secondly, we aimed o s udy whe he ANT can be delinea ed using mic oelec ode eco ding (MER), a me hod widely used in mo emen diso de su ge y, which can be ega ded as an al e na i e elec ophysiological imag- ing modali y di e en ia ing whi e and g ay ma e based on hei fi ing p ope ies. Finally, we aimed o demons a e he ole o indi- idual a ia ion in in e p e a ion o MER da a by compa ing he MER da a p ojec ed o an AC–PC and ANT-no malized coo dina e sys em. 2. Ma e ials and me hods 2.1. Imaging Eigh pa ien s wi h e ac o y epilepsy unde wen magne ic esonance imaging o planning o su ge y. The sequences ob ained included iso opic 3D-T1W wi h 0.9 mm slices o mul i-plane e- cons uc ion, axial T2W, and STIRinall h eeimagingplanes.All images we e ob ained using a 3 T scanne (MAGNETOM T io 3 T, Siemens Heal hca e Sec o , E langen, Ge many) wi h a 12-channel head ma ix coil. The STIR sequence used was a 2D u bo in e sion eco e y sequence wi h a sho in e sion ime (TR/TE/TI = 8300 ms/22 ms/120 ms: acqui- si ion ime 7.05 min), wi h a slice hickness/gap o 2.0/0.2 mm, ma ix size 256 × 256 and field o iew o 235 mm. In a enous con as was used o he 3D-T1W images. 2.2. Delinea ion p ocedu e o he ANT in STIR images All images we e analyzed using Leksell Su giPlan so wa e (Elek a AB, S ockholm, Sweden). A fi s , midpoin s o he an e io com- missu e–pos e io commissu e (AC–PC) line we e defined in he 3D-T1-weigh ed images. Nex , a co- egis a ion o STIR and 3D-T1 con as images was pe o med. S e eo ac ic coo dina es ela i e o he mid-commissu al poin (MCP) o he poin s along he delin- ea ed bo de s o he ANT in he co onal and sagi al o ien a ions we e collec ed. The ea e , a c oss-sec ional model o ANT was calcu- la ed in he sagi al and co onal o ien a ions and plo ed on o an AC– PC based coo dina e sys em. These delinea ions we e also used o measu emen s o leng h, heigh , wid h and c oss-sec ional a ea o he ANT in each halamus (Table 1). To s udy ana omical o e lap be- ween indi iduals in s e eo ac ic space, each ana omical model was aken in o a common AC–PCbasedcoo dina esys em.Thedeg ee o ana omical o e lap was exp essed as o e lapping a ea (mm 2 , Fig. 3) as a unc ion o he numbe o pa ien s. Ta ge ing o implan- a ion was pe o med by using AC–PC based coo dina es as a s a ing poin and adjus ing he a ge di ec ly om he images o accoun o ana omical a ia ions. 2.3. Mic oelec ode eco ding Imaging esul s sugges ha ANT is delinea ed by myelin- ich whi e ma e lamina. To es his hypo hesis, a o al o 186 in aope - a i e mic oelec ode eco ding samples eco ded om 10 ajec o- ies in fi e pa ien s we e analyzed wi h espec o he p esence o spikes. Single- ack MER da a we e collec ed om fi e pa ien s bila e - ally. Fi e ajec o ies we e pa a en icula (bila e al in wo cases) and fi e we e ans en icula . We used high impedance Mic oMac o elec ode and ISIS MER Sys em by Inomed Medizin echnik GmbH, Emmendingen, Ge many. Su ge y and hus also he mic oelec ode eco dings we e pe o med unde gene al anes hesia excep in one case. Reco dings we e ob ained wi h 1-mm in e als s a ing om 10 mm be o e he calcula ed a ge , wi h 0.5-mm in e als s a ing om 5 mm be o e he a ge poin , and ex ending o 3–5 mm beyond he a ge . Rep esen a i e 2-s samples we e chosen o e e y eco ding dep h. The spike equency was calcula ed o each sample ( he numbe o N25 µV spikes/2s). In ans en icula ajec o ies, he mos supe io samples, clea ly ep esen ing he ce eb ospinal fluid, we e excluded. In pa a en icula ajec o ies, ex ending he measu emen beyond he a ge poin was mos ly dismissed o a oid en e ing he hi d en i- cle o he in e nal ce eb al ein. Each dep h along a planned ajec o y was conside ed o ep esen a poin in s e eo ac ic space. The loca ion o each eco ded sample was calcula ed ela i e o MCP and co ec ed acco ding o he loca ion o he final DBS elec ode in he pos ope a i e CT-p eope a i e MRI usion images o accoun o su gical e o . Sampleswe e analyzed on he basis o whe he o no hey had spikes ypical o halamic nuclei. A sample lacking spiking ac i i y was hypo hesized o ep esen whi e ma e . Due o ex ensi e ana omical a ia ion in he loca ion o ANT in he AC–PC based coo dina es, i was expec ed ha he samples om di e - en indi iduals migh no be su ficien o es whe he ANT is delinea ed using MER. To o e come ana omical a ia ion we decided o use bounda ies o ANT in STIR images as a ame o ano he coo dina e sys em ins ead o AC–PC. In his ANT-no malized coo dina e sys em he me- dial and in e io bo de s o ANT in a co onal STIR image h ough he mammillo halamic ac was defined as he ze o le el, and he la e al andsupe io bo de so heANTas1(Fig. 4C). SPSS 17.0 was used o isualize a ia ion in he ANT loca ion as well as o demons a e co ela ion be ween MER and imaging in o ma ion. Consen was ob ained om hepa ien s and he s udy was app o ed by he E hical Re iew Boa d o Pi kanmaa Hospi al Dis ic . 824 T. Mö önen e al. / Neu oImage: Clinical 7 (2015) 823–829 Fig. 1. Compa ison be ween Schal enb and a las and 3 T MRI STIR images. A las images ep oduced wi h pe mission om Thieme Medical Publishe s, Inc. Abb e ia ions: Ap , an e io p incipal nucleus; Cd, nucleus cauda us; DM, do somedial nucleus o halamus; DSF, do sal supe ficial nucleus; eml, ex e nal medulla y lamina; Gp, Globus pallidum; iml, in e nal medulla y lamina; MTT, mammillo halamic ac ; OT, op ic ac ; Pu , pu amen; Pu, pul ina ; Ru, nucleus ube ; STN, sub halamic nucleus; S , s ia medulla is. Table 1 Mo phological ea u es o ANT in s udy pa ien s. Pa ien Leng h (mm) Heigh (mm) Wid h (mm) C oss-sec ional a ea Co onal (mm 2 ) C oss-sec ional a ea Sagi al (mm 2 ) Le Righ Le Righ Le Righ Le Righ Le Righ 1 10.8 13.0 4.1 3.6 5.5 4.8 11.3 14.5 37.5 35.8 2 10.0 11.5 3.5 4.0 5.3 5.6 13.3 12.5 29.3 32.5 3 7.5 7.8 2.3 3.9 5.3 5.8 6.3 12.3 12.8 21.3 4 10.0 11.7 3.6 4.5 5.7 5.6 19.3 16.0 28.8 37.0 5 13.3 12.8 3.5 3.5 5.4 5.6 19.8 13.3 40.3 37.3 6 12.0 9.0 4.9 3.6 5.0 5.1 16.5 17.8 42.0 24.0 7 11.3 10.5 5.4 4.2 5.6 5.6 14.0 15.0 38.3 27.0 8 9.0 10.3 3.5 3.8 6.1 5.1 12.5 10.8 22.8 30.8 Mean 10.0 10.1 3.8 3.9 5.5 5.4 14.1 14.0 31.5 30.7 825T. Mö önen e al. / Neu oImage: Clinical 7 (2015) 823–829 3. Resul s 3.1. 3 T STIR images and b ain a las A de ailed compa ison o STIR images (pa ien 4 is shown due o mos demons a i e ana omy) and he a las by Schal enb and and Wah en (1998) in axial, sagi al and co onal o ien a ions is p esen ed in Fig. 1. We ound ou ha mos o he key s uc u es a ound ANT we e in ac isualized in STIR images. The mos impo an and mos clea ly isualized s uc u e was he mammillo halamic ac ha was clea ly isualized in he co onal and axial images and occasionally also in he sagi al images. Co onal STIR images also isualize a hin a ea o low signal in ensi y hypo hesized o ep esen a whi e ma e lamina be ween ANT and he en al an e io nucleus (VA). The in e o-medial aspec o he ANT was less clea ly delinea ed simila ly o he s e eo ac ic a lases. In he axial images, ANT was mos clea ly delinea ed in i s an e i- o and la e al aspec , whe e ANT was delinea ed by p esumably whi e ma e s uc u es ha ing low signal in ensi y in STIR images co ela ing wi h ex e nal medulla y lamina in ana omic a las (Mai e al., 2007). In some pa ien s, he in e nal medulla y lamina sepa a - ing he do somedial halamus and la e al nuclea g oup was isible in he axial images (Fig. 1). Sagi al images occasionally demons a - ed (depending on he le el o he image slice in a pa icula pa ien ) mammillo halamic ac and in e nal medulla y lamina in e io o ANT. Dis inc do sal supe ficial (DSF) subnucleus o he ANT was is- ible in a subg oup o pa ien s in sagi al images. Co onal STIR images showed he do sal supe ficial nucleus as a mo e fla nucleus loca ed pos e io , sligh ly mo e supe io and la e al compa ed o he an e i- o p incipal (Ap ) and an e omedialis subnuclei (AM, no shown). O e all, we obse ed a high deg ee o co ela ion be ween STIR im- ages and a ailable ana omical a lases, in he sense ha s uc u es in a lases we e eadily iden ifiable om he images. 3.2. Mo phology o ANT Mo phological measu emen s om each c oss-sec ional model o ANT based on delinea ion in 3 T MRI STIR images a e p esen ed in Table 1. The mean leng h o ANT (along he an e io –pos e io axis o ANT) in sagi al o ien a ion was ≈10 mm. The mean wid h o ANT in co onal images (pe pendicula o he supe io –in e io axis o ANT) was ≈5.5 mm and he mean maximal heigh o ANT in sagi al images (along he supe io –in e io axis o ANT) was ≈4mm(Table 1). The c oss-sec ional a ea o ANT in co onal images was ≈14 mm 2 and he c oss-sec ional a eain sagi alimages was ≈30 mm 2 .Fig. 2 also demon- s a es a high deg ee o a ia ion in he shape o ANT. 3.3. The loca ion o ANT in he AC–PC-based coo dina e sys em To s udy he a ia ion in he loca ion o ANT be ween indi iduals, each ana omical model was plo ed on o he same AC–PC based coo di- na e sys em ela i e o MCP (Fig. 2). Ou analyses e ealed a high deg ee o ana omical a ia ion in he loca ion o ANT be ween indi iduals as expec ed. The numbe o o e lapping c oss-sec ional models co esponding o indi idual pa ien s in he AC–PC based coo dina e sys em is p esen ed in Fig. 2. The o e lapping c oss-sec ional a eas declined in a linea ashion as unc ion o numbe o he pa ien s (Fig. 3). No ewo hily, no a ea obse ed in his pa ien g oup was o e lapped by all eigh pa- ien s in co onal o sagi al o ien a ions. Compa ed o heSchal enb and a las, ANT was consis en ly loca ed mo e supe io ly, an e io ly and la e ally (Fig. 2). 3.4. Delinea ion o ANT using mic oelec ode eco ding The spike equency in MER samples in he AC–PC based coo dina e sys em (Fig. 4A) and ANT-no malized (Fig. 4B)isdemons a edinFig. 4. We we e able o demons a e wo ana omically dis inc g oups o spikes using he ANT-no malized coo dina e sys em (Fig. 4C). When p ojec ed o he ana omical a las no malized o ANT, he fi s g oup is loca ed in an a ea co ela ing wi h ANT ( ed) and he second is co ela ed wi h VA (blue) (Fig. 4B). The p ojec ion o he same samples in o he AC– PC based coo dina e sys em ela i e o MCP shows ha samples o igi- na ing om di e en nuclei o e lapped o a significan deg ee (Fig. 4A). To u he es he hypo hesis ha ANT is su ounded by myelin- ich whi e ma e lamina, we ca ied ou a spa ial densi y analysis o MER samples which also included samples wi hou spiking ac i i y. The spa- ial densi y o he samples wi h spikes (ANT spikes o VA spikes based on equency analysis) and he samples wi hou spikes is p esen ed in he AC–PC based (Fig. 5A) and ANT-no malized coo dina e sys ems (Fig. 5B). The mos supe io samples om ans en icula ajec o ies (n=5)wi hfluid cha ac e is ics in MER we e ega ded as samples om he ce eb ospinal fluid and we e p esen ed as a sepa a e g oup. Two spa ially dis inc g oups o spikes we e iden ified simila ly o he Fig. 2. Ana omical a ia ion and c oss-sec ional o e lap be ween pa ien s.The ela ionship be ween each indi idual ana omical model and he whole pa ien g oup and ana omic a las is p esen ed. Each indi idual ANT model (one o eigh ) is demons a ed in he AC–PC coo dina e sys em. The colo s ep esen he deg ee o o e lap ( om 1–7) wi h he whole pa ien g oup, hus illus a ing he de ia ion o a pa icula pa ien om he whole pa ien g oup. The e ical dashed line in sagi al ANT models ep esen s he le el o he co onal ANT model in y-axis. Since he co onal models we e delinea ed om STIR image slices isualizing he mammillo halamic ac , he line also illus a es he ana omi- cal ela ionship be ween he junc ion o mammillo halamic ac and ANT in indi idual pa ien s. In he lowe pa o he image, he deg ee o o e lap is p esen ed in he whole pa- ien g oup oge he wi h delinea ion o ANT in he Schal enb and a las (dashed ou lines). Theda a demons a e ha ANT was loca ed mo e la e ally, supe io ly and an e io ly in 3 T MRI in his pa ien g oup compa ed o he ana omical a las. 826 T. Mö önen e al. / Neu oImage: Clinical 7 (2015) 823–829 spike equency analysis in ANT-no malized coo dina e sys em. Impo - an ly, he spike nega i e samples we e mos equen in an a ea be ween hese wo g oups o spikes in ANT-no malized coo dina e sys- em. Ins ead, when he loca ion o he same samples was exp essed el- a i e o MCP using he AC–PC coo dina e sys em he spa ial densi y o he di e en sample g oups peaked in a ela i ely small a ea wi hou clea ly sepa a ed densi y peaks om di e en o igins o samples as sugges ed by ANT-no malized coo dina e sys em. A co ela ion be ween localiza ion o he spikes and STIR image signal cha ac e is ics is demons a ed in Fig. 4C, whe e wo g oups o spikes a e e iden in an a ea co ela ing wi h ANT and VA in STIR image, and lack o spikes in an a ea wi h low signal in ensi y consis en wi h whi e ma e lamina. 4. Discussion In he p esen s udy, we aimed o s udy he a ia ion in he loca ion o ANT and combine imaging in o ma ion o he elec ical p ope ies o he an e io halamic issue. We we e able o demons a e ha ANT is delinea ed using 3 T MRI STIR images isualizing whi e ma e laminae a ound ANT. The clea need o di ec isualiza ion o ANT o clinical p ac ice in ea ing e ac o y epilepsy wi h DBS was demons a ed by he high deg ee o in e indi idual a ia ion and low deg ee o ana om- ical o e lap in he loca ion o ANT in he commonly used AC–PC coo di- na e sys em. Using isible bounda ies o ANT in 3 T MRI STIR images as a ame o he coo dina e sys em, i was possible o combine a las/STIR images wi h MER samples showing spikes o igina ing om ANT and VA. These samples we e sepa a ed by samples lacking spiking ac i i y, hus confi ming ou hypo hesis based on signal cha ac e is ics in STIR imaging abou whi e ma e laminae a ound ANT. The size and shape o ANT based on 3 T MRI imaging isually co e- la es wi h a ailable a las in o ma ion (Mai e al., 2007;Schal enb and and Wah en, 1998). A las in o ma ion sugges s ha he an e io pa o he ANT (co esponding o he an e io p incipal (AP ) and an e io medial (AM) subnuclei) is loca ed in e io ly and medially while mo e pos e io ly ANT uns sligh ly la e ally and supe io ly and has a mo e fla appea ance (Mai e al., 2007). The leng h o he nucleus in ou s udy was app oxima ely 10 mm wi h a heigh o 5 mm and a wid h o 4 mm, which is compa able o a las da a (Mai e al., 2007; Schal enb and and Wah en, 1998). In mos pa ien s, he igh ANT was loca ed sligh ly mo e an e io ly compa ed o he le side. I should be no ed ha ANT was iden ifiable usually only in a ew images om each pa ien (due o e.g. 2-mm slice hickness). I may be specula ed Fig. 3. The c oss-sec ional a ea o e lapped by s udy subjec s in sagi al and co onal o ien a ions. Fig. 4. Spike equency da a p ojec ed o he halamus using AC–PC and ANT-no malized coo dina e sys ems.The equency o spikes in MER samples and he loca ion o he sample p ojec ed o he AC–PC based coo dina e sys em (A) and ANT-no malized coo dina e sys em (B, C). Using ANT-no malized coo dina e sys em wo ana omically sepa a e g oups o spikes we e iden ified (B). The p ojec ion o spikes on o he a las o Mai (pla e 12 mm pos e io o AC) e ealed high a co ela ion be ween he medial/supe io g oup o spikes ( ed) wi h he bounda ies o ANT and a la e al/in e io g oup o spikes wi h he bounda ies o en al an e io nucleus (VA) (blue). P ojec ion o he samples o he AC–PC based coo dina e sys em (A) demons a es ha he g oups o spikes a e spa ially o e lapped o a significan deg ee wi hou clea , ecognizable p ojec ion o any specific nucleus o he halamus. P ojec ion o he spikes on o STIR image (C) demons a es a high deg ee o co ela ion be ween STIR signal cha ac e is ics and MER da a, whe e spikes a e p esen in ANT and VA bu p esumed whi e ma e lamina lacks spiking ac i i y. 827T. Mö önen e al. / Neu oImage: Clinical 7 (2015) 823–829 ha isualiza ion o he nucleus using MRI wi h he a o emen ioned somewha oblique appea ance, he le el o MRI image ela i e o he nucleus may ha e a ela i ely la ge impac on he obse ed mo phology and c oss-sec ional a ea o he nucleus. Howe e , ou cu en imaging me hod does no allow o mul iplana isualiza ion o he ANT wi hin a easonable scanning ime. Iso opic 3D-STIR migh p o e help ul, and we in end o es i . Ana omical a ia ion o he an e io nucleus has been p e iously de- sc ibedinanana omicals udy(Van Bu en and Bo ke, 1972). Ou s udy confi ms and ex ends his finding using mode n imaging echniques. Visualiza ion o ANT has been ecen ly demons a ed in 3 T T1 images using a igh -fi ing 32-channel head coil and 2 a e ages (Buen jen e al., 2014), allowing dema ca ion and di ec a ge ing o his nucleus. Ou esul s u he suppo he concep ha ANT can be di ec ly a ge ed using clinically a ailable 3 T MRI sequences. Howe e , he MRI sequence used by Buen jen3sg oupis15–20 minu es long and hus equi es a gua - an eed head s abiliza ion me hod o he whole scanning ime. In e sion eco e y sequences such as STIR imp o e issue con as by supp essing a pa icula issue acco ding o he selec ed in e sion ime. STIR images wi h a sho in e sion ime a he poin whe e he di - e ence in longi udinal magne iza ion be weenwhi e and g ay ma e is a i s maximum will gene a e images wi h a supe io con as be ween whi e and g ay ma e . Toge he hese s udies sugges ha isualiza ion o ANT is based on he con as be ween he g ay ma e o ANT and en eloping whi e ma e lamina. Impo an ly, we we e also able o demons a e his whi e ma e –g ay ma e bo de wi h ano he me hod using MER. Ou s udy also has mo e gene al implica ions. Ou da a clea ly demons a es he weaknesses o indi ec AC–PC based a ge ing. The AC–PC coo dina e sys em was ini ially used o a ge s uc u es du ing he en iculog aphy e a, bu i is s ill used in he cu en age o CT and MRI o some indica ions, such as epilepsy. In mo emen diso de su ge y, MRI images a e widely used in a ious deg ees o adjus he final a ge . Howe e , a ma e e en mo e impo an han he echnique used in a ge ing o he s uc u es, is ha he AC–PC based coo dina e sys em is s ill widely used o epo he ou come a e s e eo ac ic le- sions o he pos ope a i e loca ion o DBS elec odes. This mos likely causes excessi e inconsis encies in he ou come analyses and migh , in ac , p e en esea che s om iden i ying he mos op imal a ge . Only e y ew s udies desc ibe he use o “ a ge specific-coo dina e sys em”in epo ing ou come a e DBS in mo emen diso de s (Ashkan e al., 2007;Plaha e al., 2006;Vayssie e e al., 2004). Ou e- sul s also s ongly suppo his concep in ANT-DBS. In he SANTE- ial (Fishe e al., 2010), DBS lead posi ions we e e ified pos ope a i ely wi h magne ic esonance imaging. The mos cen ally loca ed con ac wi hin each ANT was selec ed as he ac i e s imula ion con ac . I no con ac s we e loca ed inside ANT, he lead was eplaced. The s e eo ac ic coo dina es used in he s udy o a ge defini ion we e p o ided a a la e ime. Taking in o conside a ion he in o ma ion p o ided by he p esen s udy, elec ode loca ion assess- men wi h a las compa isons and MRI sequences no clea ly delinea ing ANT likely con ains a ac o o unce ain y. The e o e, i is possible ha no all o he ac i e con ac s in he SANTE- ial we e in ANT. The exac knowledge o he loca ion o he elec odes wi h ega d o indi idual ana omy may lead o inc eased implan a ion accu acy and be e clinical esponse. T ial esul s should also be in e p e ed acco dingly. We in end o explo e his ma e in u he s udies. Afinding wi h significan clinical impo ance was ha he ANT models in ou pa ien s in MRI we e loca ed mo e supe io ly, la e ally and an e io ly compa ed o he ana omic a las o Schal enb and (Fig. 2). The s e eo ac ic coo dina es used in implan a ion p oce- du es epo ed in he li e a u e a e 12 mm supe io , 5–6mmla e al and 0–2 mm an e io o he mid-commissu al poin (Schal enb and and Wah en, 1998). This poin is loca ed a he in e io aspec o he ANT in MRI in majo i y o ou pa ien s. I is impo an o no ice ha he success o elec ode implan a ion o he ANT will depend no only on he selec ed a ge bu also on he selec ed ajec o y o he a ge poin . Indi ec a ge ing o he poin abo e using a ans en icula app oach esul s mos likely in a pene a ion o he main pa o he ANT. Howe e , i a mo e la e al anspa enchymal app oach is selec ed o a oid pene a ing he en icle, he ajec o y may only each an in- e io aspec o he ANT. This may ha e clinical significance wi h espec o he apeu ic e ec s. In he p esen s udy, we ha e used he isualized bounda ies o he ANT o s udy he localiza ion o MER samples ela i e o ANT. Simila a ge specific coo dina e sys ems ha e been desc ibed p e iously o he globus pallidus (Vayssie e e al., 2004) and pos e io sub halamic a ea/caudal zona ince a (Ashkan e al., 2007;Plaha e al., 2006). As he Schal enb and a las is composed o b ain slices ob ained om h ee di e en indi iduals, e o s ha e been made o p oduce a lases Fig. 5. The spa ial ela ionship be ween samples wi h spikes and no spikes in he ANT-STIR based coo dina e sys em (A) and AC–PC based coo dina e sys em (B). The spa ial densi y o samples wi h spikes co ela ing wi h ANT ( ed) and VA (blue) in he spike equency analysis is p esen ed oge he wi h he densi y o spike-nega i e samples. The mos supe io MER samples om pa ien s wi h ans en icula ajec o ies (and MER signal wi hou any neu al ac i i y) we e ega ded as samples ep esen ing he CSF, and a e p esen ed sepa a ely (ligh blue). Spike-nega i e samples wi h somewha highe backg ound ac i i y o mino spiking ac i i y a e hypo hesized o ep esen whi e ma e laminae (g ay). In he ANT-no mal- ized coo dina e sys em, wo g oups o spikes sugges ed o o igina e om ANT ( ed) and VA (blue) a e sepa a ed by spike-nega i e samples ep esen ing mos likely he whi e ma e lamina (B). Thus, he dis ibu ion o spikes and spike-nega i e samples om di e en indi iduals in ANT-no malized coo dina e sys em shows a co ela ion wi h ana omic a las in o ma- ion. Using he same colo -coding o he samples in he AC–PC coo dina e sys em (A) shows he peaking o sample densi ies om di e en sou ces (ANT/VA/lamina) in a ela i ely small a ea wi h a poo spa ial esolu ion. 828 T. Mö önen e al. / Neu oImage: Clinical 7 (2015) 823–829 scalable o indi idual ana omy. The Talai ach a las o e s a concep o p opo ionali y, bu adap ing i o indi idual b ain dimensions makes an assump ion ha homo he ic ans o ma ion be ween wo b ains is possible, which appa en ly has no been p o en. We a e acu ely awa e o he small numbe o subjec s in ou s udy. Howe e , s e eo ac ic su - ge y is being pe o med all o e he wo ld using he Schal enb and a las ha is based on cada e samples om only h ee indi iduals. (Bajcsy, e al., 1983;Da a zikos, 1996;Vayssie e e al., 2002) I can be s a ed ha using a coo dina e sys em based on isible land- ma ks in he MRI ha a e as close as possible o he a ea o in e es should p o ide he mos accu a e esul s. In ou da a, o ins ance dis inguishing spikes o igina ing om VA and ANT was possible due o he ac ha he isible whi e ma e lamina be ween hese nuclei was used as a landma k o he coo dina e sys em. Mo e dis an ana om- ical landma ks such as AC and PC ailed o dis inguish hese samples. 5. Conclusion Ou da a sugges ha ANT can be di ec ly isualized using 3 T MRI due o en eloping whi e ma e s uc u es. Ana omical a ia ion in he an e io halamic a ea is ex ensi e, and he e o e di ec isualiza ion is supe io o indi ec a ge ing o accu a e implan a ion. Ex ensi e an- a omical a ia ion also calls o mo e de ailed analyses o ac i e con ac loca ions in in e p e ing esul s om bo h pas and cu en ea men ials. Fu he s udies wi h clinical esul s a e needed o p o e whe he di ec a ge ing also ansla es o supe io seizu e con ol. Acknowledgemen s We would like o exp ess ou g a i ude o all suppo ing pa ies, es- pecially o medical physicis Pe i Ryymin o his ex ensi e con ibu- ion in op imizing ou imaging sequences. Timo Mö önen would like o hank he Finnish Epilepsy Resea ch Founda ion o hei financial suppo . The s udy was also financially suppo ed by compe i i e EVO- unding o Pi kanmaa Hospi al Dis ic . Re e ences Ashkan,K.,Bloms ed ,P.,Z inzo,L.,Tisch,S.,Yous y,T.,Limousin-Dowsey,P., Ha iz, M.I., 2007. Va iabili y o he sub halamic nucleus: he case o di ec MRI guided a ge ing. B . J. Neu osu g. 21 (2), 197–200. h p://dx.doi.o g/10. 1080/0268869070127224017453788. Bajcsy, R., Liebe son, R., Rei ich, M., 1983. A compu e ized sys em o he elas ic ma ching o de o med adiog aphic images o idealized a las images. J Compu Assis Tomog 7 (4), 618–625. h p://dx.doi.o g/10.1097/00004728-198308000-000086602820. Buen jen,L.,Kopi zki,K.,Schmi ,F.C.,Voges,J.,Tempelmann,C.,Kau mann,J., Kanowski, M., 2014. Di ec a ge ing o he halamic an e o en al nucleus o deep b ain s imula ion by T1-weigh ed magne ic esonance imaging a 3 T. S e eo ac . Func . Neu osu g. 92 (1), 25–30. h p://dx.doi.o g/10.1159/ 00035152524216749. Child, N.D., Bena och, E.E., 2013. An e io nucleus o he halamus: unc ional o ganiza- ion and clinical implica ions. Neu ology 81 (21), 1869–1876. h p://dx.doi.o g/10. 1212/01.wnl.0000436078.95856.5624142476. Da a zikos, C., 1996. Spa ial no maliza ion o 3D b ain images using de o mable models. J Compu Assis Tomog 20 (4), 656–665. h p://dx.doi.o g/10.1097/00004728- 199607000-000318708076. Fishe ,R.,Salano a,V.,Wi ,T.,Wo h,R.,Hen y,T.,G oss,R.,Oommen,K.,Oso io,I., Nazza o,J.,Laba ,D.,Kapli ,M.,Spe ling,M.,Sandok,E.,Neal,J.,Hand o h,A., S e n,J.,DeSalles,A.,Chung,S.,She e ,A.,Be gen,D.,2010.Elec icals imula- ion o he an e io nucleus o halamus o ea men o e ac o y epilepsy. Epilepsia 51 (5), 899–908. h p://dx.doi.o g/10.1111/j.1528-1167.2010.02536. x20331461. Hodaie, M., Wennbe g, R.A., Dos o sky, J.O., Lozano, A.M., 2002. Ch onic an e io hala- mus s imula ion o in ac able epilepsy. Epilepsia 43 (6), 603–608. h p://dx.doi. o g/10.1046/j.1528-1157.2002.26001.x12060019. Ke igan, J.F., Li , B., Fishe , R.S., C ans oun, S., F ench, J.A., Blum, D.E., Dich e , M., She e , A., Bal uch, G., Jaggi, J., K one, S., B odie, M., Rise, M., G a es, N., 2004. Elec ical s imula ion o he an e io nucleus o he halamus o he ea men o in ac able epilepsy. Epilepsia 45 (4), 346–354. h p://dx.doi.o g/10.1111/j.0013-9580.2004. 01304.x15030497. Lee, K.J., Shon, Y.M., Cho, C.B., 2012. Long- e m ou come o an e io halamic nucleus s imula ion o in ac able epilepsy. S e eo ac . Func . Neu osu g. 90 (6), 379–385. h p://dx.doi.o g/10.1159/00033999122922474. Mai, J.K., Paxinos, G., Voss, T., 2007. A las o he Human B ain hi d edi ion. Academic P ess. Plaha, P., Ben-Shlomo, Y., Pa el, N.K., Gill, S.S., 2006. S imula ion o he caudal zona ince a is supe io o s imula ion o he sub halamic nucleus in imp o ing con ala e al Pa - kinsonism. B ain J. Neu ol. 129 (7), 1732–1747. h p://dx.doi.o g/10.1093/b ain/ awl12716720681. Schal enb and, G., Wah en, W., 1998. A las o S e eo axy o he Human B ain second edi- ion. Thieme, S u ga . S ypulkowski, P.H., Gi akis, J.E., Bills om, T.M., 2011. De elopmen o a la ge animal model o in es iga ion o deep b ain s imula ion o epilepsy. S e eo ac . Func . Neu osu g. 89 (2), 111–122. h p://dx.doi.o g/10.1159/00032334321336007. Van Bu en, J.M., Bo ke, R.C., 1972. The mesial empo al subs a um o memo y. Ana om- ical s udies in h ee indi iduals. B ain J. Neu ol. 95 (3), 599–6324655285. Vayssie e, N., Hemm, S., Ci , L., Pico , M.C., Diakono a, N., El Fe i , H., F e ebeau, P., Coubes, P., 2002. Compa ison o a las- and magne ic esonance imaging-based s e- eo ac ic a ge ing o he globus pallidus in e nus in he pe o mance o deep b ain s imula ion o ea men o dys onia. J. Neu osu g. 96 (4), 673–679. h p://dx.doi. o g/10.3171/jns.2002.96.4.067311990806. Vayssie e, N., an de Gaag, N., Ci , L., Hemm, S., Ve die , R., F e ebeau, P., Coubes, P., 2004. Deep b ain s imula ion o dys onia confi ming a soma o opic o ganiza ion in he globus pallidus in e nus. J. Neu osu g. 101 (2), 181–188. h p://dx.doi.o g/10.3171/ jns.2004.101.2.018115309906. Yamada, K., Akazawa, K., Yuen, S., Go o, M., Ma sushima, S., Takaha a, A., Nakagawa, M., Mineu a, K., Nishimu a, T., 2010. MR imaging o en al halamic nuclei. A.J.N.R.Am.J.Neu o adiol.31(4),732–735. h p://dx.doi.o g/10.3174/ajn . A187019926703. 829T. Mö önen e al. / Neu oImage: Clinical 7 (2015) 823–829