13$B!%5!G=E*(BMRI$B8&5f2q(B

2002$BG/(B11$B7n(B28$BF|!](B11$B7n(B29$BF|(B
$BBeI=!&@$OC?M!'DjF#5,90(B
$B=jFbBP1~

$B!J#1!K(B

The real-time analysis of fMRI data: algorithm and system development

E. Bagarinao$B!$Cf0fIR@2!J;:Am8&!K(B

$B!J#2!K(B

fMRI Studies of Human Brain Functions at Columnar Resolution

Kang Cheng, Allen R. Waggoner$B!$(BKeiji Tanaka$B!JM}8&!K(B

$B!J#3!K(B

MRI$B;#A|2;6/EY$NJQ2=$KH<$&D03PLn$NIj3h(B

$B2,EDCN5W!J@8M}8&!K(B

$B!J#4!K(B

EPI$B;~7ONs?.9f$K$*$1$kDc<~GH%I%j%U%H$NMW0x$K$D$$$F(B

$BEgED0iW"!$F#K\0lO:!$@5LZ?.IW!J(BATR$BG>3hF0%$%a!<%8%s%0%;%s%?!

$B!J#5!K(B

fMRI/EEG$BF1;~7WB,$H(BBe-fMRI$B!J(BBrain event-related fMRI$B!K(B

$B5\FbE/!$;0:j>-Li!$0K4]2,=S=(!$EDCfLw?M!JDLAm8&!K(B

$B!J#6!K(B

fMRI$B$H(BNIR$BF1;~7WB,(B

$BGpAR7r0l!$:4F#E/Bg!$JFAR5A@2!JJ!0f0eBg!K(B

$B!J#7!K(B

$B%F%s%=%k2hA|(B

$B:4F#E/Bg!$GpAR7r0l!$JFAR5A@2!JJ!0f0eBg!K(B

$B!J#8!K(B

$BG'CN5-212]Bj$K$*$1$kA0F,MU3hF0(B

$B5\2.@>@65.!JElBg!K(B

$B!J#9!K(B

A Comparison of the Temporal Characteristics of the BOLD Responses in V1, MT, and the Primary Motor Cortex$B!J(BM1$B!K(Bto a Variety of Stimuli

R. Allen Waggoner, Kang Cheng, Keiji Tanaka$B!JM}8&!K(B

$B!J(B10$B!K(B

$B;R6!$HO7?M$N(BfMRI

$BBYMe2mEP!JF|K\Bg3X!K(B

$B!J(B11$B!K(B

activation study$B$K$*$1$k8D?M:9$K$D$$$F$N8!F$(B

$B4dED0l1,$B?y1:85N<(B¹,$BEOn5>fIW(B^1,2,Jorge Riera¹,$B;01:D>1,
$B=)7nM4;R(B^1,2,3,$B:4!9M4;R(B^1,4,$BEOItK'I'(B⁵,$B@8EDF`Jf(B^1,4,
$B2,K\1Q9T(B^1,2,$BA0EDBY90(B⁵,$B>>9>9nI'(B⁵,$B@nEgN4B@(B¹
$B!J(B¹$BElKLBgL$Mh2J3X5;=Q6&F18&5f%;%s%?!<(B,²$BElKLBg2CNp8&(B,
³$BElKLBg1!0e(B,⁴$BElKLBg1!9q:]J82=(B,⁵$BElKLJ!;cBg!K(B

$B!J(B12$B!K(B

$B5!G=E*(BMRI$B$r$b$A$$$??7@8;yG>Be

$BDjF#5,90!J@8M}8&!K(B

$B!Z;22C![(B
$B>.@n@?Fs!J>.@nG>5!G=8&5f=j!K!$4dED0lfIW!$=)7nM4;R!JElKLBg!K!$BYMe2mEP!JF|K\Bg!K!$EDCf7<<#!$(BKangCheng$B!JDx9/!K!$(BR Allen Waggoner$B!$(BPei Sun$B!$>eLn8-0l!$EAM%;R!$(BJustin Gardner$B!$KY9>N>Hx9aLo;R!$:XF#$b$h$3!J;:Am8&!K!$@5LZ?.IW!$EgED0iW"!$F#K\0lO:!J(BATR$B!K!$5\FbE/!JDLAm8&!K!$5\2AmM:!J@8M}8&!&E}9g@8M}!K!$>.>>1QI'!J@8M}8&!&9b.@n@5!$EDJUM4M|!$8qED9'OB!J@8M}8&!&9b

$B!Z35MW![(B
$B!!5!G=E*<'5$6&LD2hA|K!!J0J2<(BfMRI$B!K$O6aG/Cx$7$/Ia5Z$7!$%R%H$r4^$`NnD9N`$NG>5!G=$rHs?/=1E*$KC5$k>e$GBn1[$7$?2DG=@-$r$b$D$3$H$,L@$i$+$H$J$C$F$-$?!#$=$N0lJ}$GE,@Z$J;#A|J}K!$dE}7WE*2r@OK!$H$$$C$?5;=QLL$N$_$J$i$:!$5-O?$5$l$k?.9fJQ2=$N$b$D@8M}3XE*0UL#!JB($A6I=jG>7lN.$"$k$$$OEE5$E*3hF0$H$N4X78$J$I!K$rM}2r$9$k>e$G$b8!F$$9$Y$-2]Bj$,;3@Q$7$F$$$k!#K\8&5f2q$G$O!$(BfMRI$B$K4X$9$k5;=QE*$J$i$S$K@8M}3XE*$J=t2]Bj$K$D$$$F3hH/$J5DO@$r$*$3$J$&$H$H$b$K!$>pJs8r49$N>l$rDs6!$9$k$3$H$rL\E*$H$9$k!#6qBNE*$K$O!$8=:_!$(BfMRI$B8&5f$r@9$s$K

 

$B!J(B1$B!K(BThe real-time analysis of fMRI data: algorithm and system development

E. Bagarinao and T. Nakai$B!J(BLife Electronics Laboratory
National Institute of Advanced Industrial Science and Technology$B!K(B

$B!!(BThe real-time analysis of functional magnetic resonance imaging$B!J(BfMRI$B!K(Bdata advances the efficacy of fMRI as a flexible tool for neurological investigations for both basic research purposes and clinical applications. MRI systems with real-time statistical analysis capabilities have been used to provide immediate confirmation of activation, to assess subject performance and data quality, and to locate regions of interest, among others. In this work, we developed a system for the real-time analysis of functional magnetic resonance imaging$B!J(BfMRI$B!K(Btime series. The system is composed of an MR scanner for data acquisition and paradigm control, a computational server consisting of a cluster of readily available personal computers$B!J(BPC cluster$B!K(Bfor real-time fMRI data analysis, and a storage device for storing data. The highly parallel, voxel-wise processing of fMRI data motivated the use of PC clusters for parallel computation. Advantages of using a PC cluster include a significant increase in computational speed as well as the provision of the much needed storage requirement for processing fMRI data by pooling the resources, such as physical memory, of the different PCs in the cluster. To support real-time statistical analysis, we developed an algorithm to estimate the coefficients of general linear models$B!J(BGLM$B!K(Bin real-time. Results of the real-time analysis of fMRI data of a normal subject performing a simple finger-tapping task demonstrate the capabilities of the described system. For a basic analysis without image preprocessing such as realignment, smoothing, or normalization, the time required to estimate the coefficients of a GLM with 15 terms, compute the t-statistics, and update the statistical map for each volume acquisition is about 0.175 s for a 64 x 64 x 30 image data and 0.27 s for a 128 x 128 x 30 image data, which is much less than the TR which was set to 5 s.

 

$B!J(B2$B!K(BfMRI Studies of Human Brain Functions at Columnar Resolution

Kang Cheng, Allen R. Waggoner, Keiji Tanaka
$B!J(BLaboratory for Cognitive Brain Mapping, RIKEN Brain Science Institute$B!K(B

$B!!(BIn this presentation, we demonstrate that with optimal imaging parameters and proper experimental procedures, functional architectures such as ocular dominance columns in human primary visual cortex can be mapped with high-field fMRI. Possible sources contributing to the functional BOLD contrast at columnar resolution and limiting factors in conducting such experiments are discussed from both anatomical and imaging perspectives.

 

$B!J(B3$B!K(BMRI$B;#A|2;6/EY$NJQ2=$KH<$&D03PLn$NIj3h(B

$B2,EDCN5W!$K\ED!!3X!$DjF#5,90(B
$B!J@8M}3X8&5f=j?4M}@8M}3X8&5fItLg!K(B

$B!!5!G=E*(BMRI$B$O8@8l2]Bj$r4^$a$?B?$/$Nl$r0l;~E*$K;_$a$k$3$H$GAjBPE*@E2;;~4V$r@_$1!$D03P;I7cDh<($rMF0W$K$9$k;#A|J}K!$K$D$$$F4pACE*$J8!F$$r9T$C$?!#(B

$B!!]$K!$(B13-17$BICKh$K(B1$BIC4V79l$r;_$a$?!J(BOFF$B!K$H$3$m!$N>B&D03PLn$K(BOFF$B$KF14|$7$?9-HO$JIj3h$rG'$a$?!#$&$A(B11$B?M$rBP>]$Kl$r;_$a$F!$0z$-B3$-D03PLn$G5/$3$C$?7lN.JQ2=$re$,@bL@$5$l$?!#$9$J$o$A0ll9g$N$_$G$J$/!$$=$l$,8:>/$7$?>l9g$K$b!$Dj>o>uBV$h$j$bBg$-$JIj3h$r<($9$3$H$,H=L@$7$?!#(B

 

$B!J(B4$B!K(BEPI$B;~7ONs?.9f$K$*$1$kDc<~GH%I%j%U%H$NMW0x$K$D$$$F(B

$BEgED0iW"!$F#K\0lO:!$@5LZ?.IW(B
$B!J(BATR$BG>3hF0%$%a!<%8%s%0%;%s%?!

$B!!D9;~4V$N(BfMRI$B3hF0$K5/0x$9$k$b$N$G$O$J$$$H9M$($i$l$k!#J,@O;~$K%U%#%k%?$K$h$k=|5n=hM}Ey$r;\$9$,!$$+$($C$FM-32$J>l9g$b$"$k!#$3$N%I%j%U%H$NMW0x$r8!F$$7$?!#(B

$B!!!J(B1$B!K%U%!%s%H%`$GD9;~4V(BEPI$B%;%C%7%g%s$rO"B3$7$F9T$J$&$H!$?.9fJQF05Z$S2hA|%7%U%H$O(B2$B2sL\$N%9%-%c%s$G8:>/$7$?!#!J(B2$B!KAuCV$r%&%)!<%`%"%C%W8e$NO"B3%;%C%7%g%s$G$O!$?.9fJQF0!&2hA|%7%U%H$H$b$K2~A1$5$l$?!#!J(B3$B!KAuCV$N%&%)!<%`%"%C%W$J$7$G;iKCM^@)%Q%k%9$N(BFA$B$r>.$5$/$7$F%U%!%s%H%`$r%9%-%c%s$9$k$H?.9fJQF0$O8:>/$7$?$b$N$N!$2hA|%7%U%HNL$O(B2$B2sL\$G8:>/$9$k!#!J(B4$B!K!J(B3$B!K$HF1MM$N>r7o$G@8BN$N%9%-%c%s$r9T$&$H!$!J(B3$B!KF1MM$K?.9fJQF0$O8:>/$7$?$,!$2hA|%7%U%H$N798~$O(B2$B2sL\$K8:>/$7$?!#(B

$B!!0J>e$+$i!$Dc<~GH?.9fJQF0$O79l$NFC@-JQF0$K$h$kCf?4<~GH?tJQF0$K5/0x$7$F$*$j!$;iKCM^@)$,IT40A4$K$J$k$?$a$H9M$($i$l$k!#(B

 

$B!J(B5$B!K(BfMRI/EEG$BF1;~7WB,$H(BBe-fMRI$B!J(BBrain event-related fMRI$B!K(B

$B5\Fb!!E/!$;0:j>-Li!$0K4]2,=S=(!$EDCfLw?M(B
$B!JDL?.Am9g8&5f=j4X@>@hC<8&5f%;%s%?!

$B!!(BfMRI$B$H(BEEG$B$NF1;~7WB,$O$3$l$^$G$K$bJs9p$5$l$F$$$k$,!$M-8z$J1~MQ;vNc$O>/$J$$!#$o$l$o$l$O0J2<$NFsE@$rL\E*$H$7$F(BfMRI$B$H(BEEG$B$NF1;~7WB,%7%9%F%`$r9=C[$7$?!#(B

$B!!!J(B1$B!K(BfMRI$B$NGH%Q%?%s$O82Cx$KJQ2=$9$k$K$b$+$+$o$i$:!$$3$l$^$G3P@C?e=`$NJQF0$KH<$&G>3hF0$NJQ2=$K4X$9$k8!F$$O$[$H$s$I9T$J$o$l$F$$$J$$!#!J(B2$B!K$5$i$K!$=>Mh$NFCDj$N%?%9%/$KBP$9$k(BfMRI$B?.9fCM$NJQ2=$r7WB,$9$kBe$o$j$K!$(BEEG$B>e$N(Bphasic event$B$KH<$&(BfMRI$B?.9fCM$NJQ2=$r7WB,$9$k!$(BBrain event-related fMRI$B!J(BBe-fMRI$B!K$r?7$?$KDs>'$9$k!#(B

$B!!:#2s$O!$(BfMRI/EEG$BF1;~7WB,%7%9%F%`$N>\:Y$H!J(B2$B!K$N(BBe-fMRI$B$N7WB,7k2LNc$rCf?4$KJs9p$9$k(B

 

$B!J(B6$B!K(BfMRI$B$H(BNIR$BF1;~7WB,(B

$BGpAR7r0l!$:4F#E/Bg!$JFAR5A@2(B
$B!JJ!0f0e2JBg3X9b%(%M%k%.!<0e3X8&5f%;%s%?!

$B!!6a@V307WB,AuCV!J(BNIRS$B!K5Z$S5!G=E*(BMRI$B!J(BfMRI$B!K$GF@$i$l$k>pJs$NE}9g$,2DG=$+$I$&$+8!F$$9$kL\E*$GF1;~7WB,$r;n$_$?!#8w%H%]%0%i%U%#!Ij3h8!::$r9T$C$?!#@53N$J0LCV$NF1Dj$r9T$&$?$a!$8w%H%]%0%i%U%#AuCV$N3F%W%m!<%V$K$OD>@~%^!<%+!<$rG[CV$7$?!#Ho83o@.?MCK=w7W(B12$BL>$G$"$k!#6k7A%A%'%C%+!<%\!<%I;I7c$rMQ$$(B4$Bu%A%'%C%+!<%\!<%I$rMQ$$;I7c;~4V$r(B2-20$BIC$HJQ2=$5$;!$;~4VE*Hf3S$r9T$C$?!J(Bn=5$B!K!#%?%$%_%s%0$N@53N@-$r7W$k$?$a(BMRI$B$H;I7cDs<(%=%U%H5Z$S(BNIRS$BAuCV4V$G%2!<%F%#%s%0$r9T$C$?!#$3$N7k2L!$3F?.9f!J(BBOLD$B?.9f5Z$S(Boxy-, deoxy-Hb$B!K$O6u4VE*!$;~4VE*$KNI9%$JAj4X$r<($7$?!#$^$?!$;~4VITJQ@~7A%7%9%F%`$+$i$NP*N%$b82Cx$G$"$C$?!#=>$C$F!$N>%b%@%j%F%#$H$b$[$\F10l$N@8M}3XE*8=>]$r4Q;!$7$F$$$k$H9M$($i$l$?!#(B

 

$B!J(B7$B!K%F%s%=%k2hA|(B

$B:4F#E/Bg!$GpAR7r0l!$JFAR5A@2(B
$B!JJ!0f0e2JBg3X9b%(%M%k%.!<0e3X8&5f%;%s%?!

$B!!(BMRI$B$rMQ$$$?G>5!G=8&5f$NH/E8$K$7$?$,$C$F!$G>FbGrFb$N5!G=O"4X$r$O$+$kL\E*$+$i:#$^$G0J>e$K=EMW$H$J$C$F$-$?!#(BMRI$BAuCV$d;#A|5;=Q$N?JJb$K$H$b$J$C$F!$?eJ,;R$N3H;6MMBV$rMF0W$K2hA|2=$G$-$k$h$&$K$J$C$?!#3H;6MMBV$r$H$i$($?3H;66/D42hA|$O!$?eJ,;R$N3H;6$NDxEY$r51EY$K$h$C$FI=8=$9$k$3$H$,$G$-!$3H;66/D42hA|$rJ#?tKgMQ$$$k$3$H$G!$3H;6%F%s%=%k2hA|$N;;=P$,2DG=$H$J$k!#G>FbGrl8{G[$NJ}8~$H$NAj8_:nMQ$K$h$k1F6A$r]$O3H;6$NITEyJ}@-$H8F$P$l!$3H;66/D4<'>l8{G[$NJ}8~$KJ?9T$JGr$J@~0]B+$G$O>.$5$J3H;678?t$,7WB,$5$l$k!#K\8&5f$G$OG>FbGrH%\%/%;%k$H6a@\%\%/%;%k4V$NJ}8~$KFbGrFbGrNB$*$h$SFbJq$N%;%0%a%s%F!<%7%g%s$,2DG=$G$"$C$?!#Ds0FMh$N8GM-Bh0l%Y%/%H%k$N6JN($d1_?m3Q$rMxMQ$7$?uFCD'$rH?1G$7$?5wN%4X?t$NMxMQ$K$h$C$F6aK5$NC5:w;~$K<+M3EY$,9b$$FCD'$r$b$D!#K\8&5f$OG>FbGr5!G=$H$NO"4X2rL@$K4sM?$G$-$k$b$N$G$"$k!#(B

 

$B!J(B8$B!KG'CN5-212]Bj$K$*$1$kA0F,MU3hF0(B

$B5\2.@>@65.(B
$B!JEl5~Bg3X0e3XIt@8M}3X65<

$B!!(BMRI$B!J(BfMRI$B!K$NH/E8$K$h$C$F!$B?$/$N5-212]Bj!JFC$K%(%T%=!<%I5-212]Bj!K$K$*$$$F!$9-HO$JA0F,MUItJ,$N?@7P3hF0$,Js9p$5$l$k$h$&$K$J$C$?!#$7$+$7!$$3$&$7$??@7P3hF0$,$I$N$h$&$JG'CNE*>pJs=hM}$r$*$3$J$&4pAC$H$J$C$F$$$k$N$+$OI,$:$7$bL@$i$+$G$O$J$$!#?@7P3hF0$N6I:_@-!$0[$J$k2]Bj4V$N3hF0NN0h$N=EJ#!$2]Bj$N?@7P?4M}3XE*GX7JEy$r9MN8$7$F!$:G6a$$$/$D$+$N5!G=2hA|7WB,$r9T$C$?!#5;=QE*$K$O!$(Bcognitive subtraction$B7?$G$J$/!$2DG=$J8B$j(Bparametric modulation$B7?$N%G%6%$%s$r:NMQ$9$k$h$&$KEX$a$?!#$=$N$&$A!$7r>oHo83]$H$7$F9T$C$?6a;~5-21!J(Brecency memory$B!K2]Bj!$%U%#!<%j%s%0!&%*%V!&%N%&%$%s%0!J(Bfeeling-of-knowing$B!K2]BjEy$NCN8+$rCf?4$KJs9p$9$k!#(B

 

$B!J(B9$B!K(BA Comparison of the Temporal Characteristics of the BOLD Responses in V1, MT, and the Primary Motor Cortex$B!J(BM1$B!K(Bto a Variety of Stimuli

R. Allen Waggoner, Kang Cheng, Keiji Tanaka
Laboratory for Cognitive Brain Mapping, RIKEN Brain Science Institute

$B!!(BIn an effort to look beyond the basics of the BOLD response, and elucidate aspects which are specific to a given stimulus or to a given cortical area, we have compared BOLD responses in V1, MT, and M1 to a variety of stimuli. Stimuli that produce a strong response in either the visual cortex$B!J(Ba flickering checker board and moving dots$B!K(Bor the primary motor cortex$B!J(Bfinger tapping$B!K(Bwere used. The resulting time courses were compared, looking for common and unique features. The results in the visual areas show a two-stage rise for both stimuli, which is not seen in M1. Also, the flickering checker board caused the normal 0.1Hz vasomotionoscillations to become more pronounced during the sustained response in the visual areas, compared with the other stimuli both in the visual areas and M1.

 

$B!J(B10$B!K;R6!$HO7?M$N(BfMRI

$BBYMe2mEP(B
$B!JF|K\Bg3X!K(B

$B!!5!G=E*(BMRI$B!J(BfMRI$B!K$O(BPET$B$KHf$Y$FHs?/=1@-$,9b$$$?$a!$$3$l$^$G$OD4$Y$k$3$H$,Fq$7$+$C$?!$7r>o$J;R6!$dO7?M$NG>3hF0$r$3$l$rMQ$$$FD4$Y$k$3$H$,2DG=$K$J$C$?!%$3$l$^$G$K(B6$B!A(B14$B:P$N7r>o$J;R6!$rHo8!d2=$NH/C#$,Aa$/$*$3$kNN0h$G$O!$(BfMRI$B$rMQ$$$FG>$NIj3h$r@.?M$HF1MM$K4Q;!$9$k$3$H$,$G$-$k$3$H$,$o$+$C$?!#$7$+$7!$A0F,O"9gLn$J$I!$H/C#$NCY$$NN0h$G$O!$MD>/;y$G$OG>$NIj3h$,$&$^$/IA=P$G$-$:!$(BfMRI activation$B$N86M}$G$"$k(BBOLD$B8z2L$,!$?q>d2=$,L$H/C#$JNN0h$G$O4Q;!$G$-$J$$2DG=@-$r<(:9$9$k7k2L$rF@$?!#0lJ}!$O7?M$N(BfMRI$B$G$O!$A0F,MU2]Bj$r9T$o$;$?$H$-$K$O!$A0F,MU$N7ABV3XE*$J0`=L$,?J$s$G$$$k$K$b$+$+$o$i$:!$7r>o$J@DG/$HJQ$o$i$J$$3hF0$,F@$i$l$?!#:#2s$O$3$l$i$N%1!<%9$K$D$$$F9M;!$9$k!#(B

 

$B!J(B11$B!K(Bactivation study$B$K$*$1$k8D?M:9$K$D$$$F$N8!F$(B

$B4dED0l1,$B?y1:85N<(B¹$B!$EOn5>fIW(B^1,2$B!$(BJorge Riera¹$B!$;01:D>1$B!$=)7nM4;R(B^1,2,3$B!$(B
$B:4!9M4;R(B^1,4$B!$EOItK'I'(B⁵$B!$@8EDF`Jf(B^1,4$B!$2,K\1Q9T(B^1,2$B!$A0EDBY90(B⁵$B!$>>9>9nI'(B⁵$B!$@nEgN4B@(B¹
$B!J(B¹$BElKLBg3XL$Mh2J3X5;=Q6&F18&5f%;%s%?!<(B,²$BElKLBg3X2CNp0e3X8&5f=j(B,
³$BElKLBg3XBg3X1!0e3X8&5f2J(B,⁴$BElKLBg3XBg3X1!9q:]J82=8&5f2J(B,
⁵$BElKLJ!;cBg3X46@-J!;c8&5f=j!K(B

$B!!(BfMRI$B$rMQ$$$?G>$NIj3h8&5f$K$*$$$F$Or7o!&2]Bj$N0c$$$H$=$l$KH<$&(BMR$B?.9f$NJQ2=$H$N4V$N4XO"@-$,5DO@$5$l$k$,!$8D?MKh$N(BMR$B?.9f$NH?1~@-$N:90c$K$D$$$F$O==J,9MN8$5$l$F$3$J$+$C$?!#$=$3$GK\8&5f$G$O!$7r9/@.?M(B49$BL>$rHo832A$N0Y$NI8=`2]Bj$H$7$F(B1.5Hz$B$NIQEY$GH?E>$9$kGr9u%A%'%C%+!<%U%i%C%0$rDs<($7H?E>$N%Z!<%9$KF14|$7$F1&8$r3+JD$5$;$k;k3P;I7c!&1?F02]Bj$r2]$7!$(B5$BICKh$K(BEPI$B;#A|$r9T$C$?!#0B@E;~$r(Bbase line$B$H$7$?2]BjCf$N0l2A$7!$$=$NHo83.$5$/$J$$$3$H$K==J,Cm0U$r8~$1$kI,MW$,$"$k$H9M$($i$l$k!#(B

 

$B!J(B12$B!K5!G=E*(BMRI$B$r$b$A$$$??7@8;yG>Be

$BDjF#5,90(B¹$B!$;3ED902$B!$JFAR5A@2(B^$B#3(B
$B!J(B¹$B@8M}3X8&5f=j!$(B²$BJ!0f0e2JBg3XJ| ^$B#3(B$BJ!0f0e2JBg3X9b%(%M%k%.!<0e3X8&5f%;%s%?!

$B!!(BBOLD$B?.9f$NJQ2=$O!$G>7lN.$NJQ2=$N$_$J$i$:;@AGBe$O$=$NH/@8=i4|$K$*$$$F7ABV5!G=Bee>:$rG'$a!$$3$l$O@.?M$K$*$1$k%Q%?!<%s$H0lCW$7$F$$$?!#0lJ}!$(B8$B=50J9_$G$O5U$K(BMR$B?.9f$N8:>/$,8+$i$l$?!#0l/$,4Q;!$5$l$?!#(BMR$B?.9f8:>/$N8+$i$l$k;~4|$O$^$?%V%I%&E|Bej%7%J%W%93hF0$NA}2C$,!$;@AG>CHq$NA}Bg$R$$$F$O6I=j4T857?%X%b%0%m%S%sNL$rA}2C$5$;$F(BBOLD$B?.9f$N8:>/$K;j$k$b$N$H9M$($i$l$?!#5!G=E*(BMRI$B$O@5>oG>$K$*$1$kH/C#2aDx$rBed2=$H$J$i$s$GH/0i2aDx$N(Bmilestone$B$H$J$j$&$k$3$H$,<($5$l$?!#(B