MAYBERG, Helen, Deep Brain Stimulation for Depression, Emory, Oct 29, 2015
<p class="MsoNormal" >MAYBERG, Helen, Deep Brain Stimulation for Depression, Emory, Oct 29, 2015</p>
<p>
This is Psych Grand Rounds at 4 PM in LKS 130 at Stanford on Thursday.
<b>This is beyond SRO</b>! (the overflow room is also SRO!)</p>
<p>
She is from Emory Univ.</p>
<p>
she begins...</p>
<p>
shows disclosures</p>
<p>
<<Amit Etkin (my seatmate) ... we chat (was a grad student at Columbia With <b>Eric Kandel</b>...
<b> Eric was v serious abt diet and life style</b>... never strays! no cookies!
<b>Eric swims everyday at age 85....>></b></p>
<p>
shows Emory Depression DBS Team...</p>
<p>
Bob Gross her neurosurg partner... has a huge group (these are the people who rx the patients)</p>
<p>
Genl Strategy...</p>
<p>
<b>modeling disorders as circuitopathies</b>...</p>
<p>
try to build system models...</p>
<p>
connectomics...e tc.</p>
<p>
this may work for movement disorders (shows Cortex to striatum to SNc SNr
and to GPe to STN and back to thalamus etc.</p>
<p>
Mood Disorders... shows circuits: mood regulation; mood monitoring, etc; etc.</p>
<p>
how did she get here... as a neuro resident?</p>
<p>
dev of a depression Network Model... (<b>MDD = Major Depressive Disorder)</b></p>
<p>
her inspiration... was post stroke depression eg RG Robinson in Brain 1983...</p>
<p>
paralimbics areas for depression (was our focus)</p>
<p>
then cortico-limbic (lizard brain + thinking brain)</p>
<p>
but gradually refined model </p>
<p>
then how does drug (via whatever her latest model was work as explained by that model)</p>
<p>
rostral cingulate.. then insula (discrim btwn drug rx etc)</p>
<p>
area 25...</p>
<p>
<b>people kill themselves to escape this pain</b></p>
<p>
(find me a rat that will kill itself... it is so depressed)</p>
<p>
"<b>a gnawing agony... a painful self loathing</b>...</p>
<p>
Toronto DBS #7 2004</p>
<p>
a gnawing agony, a painful self-loathing that consumes all your energy and attention</p>
<p>
akinetic; extreme of depression...</p>
<p>
the will to move is not there...</p>
<p>
When I am depressed (she quotes) everything ... standing, stepping,
<b>speaking, moving, pursuing a train of thoughts... gets hung up </b>on that
loop of hesitation that ends up feeling like paralysis
I can't; I want to... I can't</p>
<p>
I wanna make this cloud above me disappear; <b>I don't want to hurt anymore</b>...</p>
<p>
above<b>... v dramatic recitation</b> *</p>
<p>
so, what is the source of the negative mood?</p>
<p>
memory evoked sadness; CBF cortical blood flow by PET</p>
<p>
shows <b>R anter insula</b>... and motor cingulate;
<b>R attentional network turns off</b>...</p>
<p>
a tug of war btwn feeling and thought....</p>
<p>
see <b>Albert Compte Lab in Barcelona</b>... he models it... get that paper...</p>
<p>
foundation for computat MDD model...</p>
<p>
so, want to isolate the common changes with recovery...</p>
<p>
so <b>SCC25 might be the biggest activation</b> but... (subcallosal cingulate SCC25)</p>
<p>
something abt this region is important... it is OVERactive...
do ablation... eg ablation... do depolarization block at high frequency...</p>
<p>
did this in Toronto in 2006...
see Neuron v 45 page 1 to 10 in Mar 3, 2006.</p>
<p>
subcallosal white matter.</p>
<p>
4 people got better; 2 did not...
and the 4 got much better over 6 months...</p>
<p>
get thatpaper...</p>
<p>
next expanded toronto cohort...</p>
<p>
set it at 130 hz for 90 seconds... we did not fiddle...</p>
<p>
made decisions w/o the psychiatrists...</p>
<p>
did not change the setting... we were rewarded.</p>
<p>
60% response rate in these 20 patients (the next Toronto cohort)</p>
<p>
was either all or none... (worked and stayed working or not)</p>
<p>
then did cohort at Emory....</p>
<p>
nobody got hypomanic...
see Holtzheimer in Arch Gen Psych 2012...</p>
<p>
but what to do with non-responders...</p>
<p>
so... who do you NOT operate on...</p>
<p>
she cites a # of trials that were halted for futility (later eg habenula, Med Forebrain bundle)</p>
<p>
are we delusionall? she says... if I had that much influence, I would have a difft job...</p>
<p>
so how to think abt variance in these results...</p>
<p>
Toronto cohort... by 1 month there were better...</p>
<p>
but rough patch at 12 weeks...</p>
<p>
<b>we have a 90% response rate at 2 years...</b></p>
<p>
where is the active contact at 6 months...</p>
<p>
simple healthy is NOT helpful... Bob Gross is way better... he can hit this spot...l</p>
<p>
<b>it is not abt areas 25.. it's abt things that are connected to 25</b>...</p>
<p>
so... must consider full network not just areas 25.</p>
<p>
eg why is this partic lady not all better...</p>
<p>
the lead was off by one mm!!! (we reimplanted here and went anterior by 1 mm and it worked...!!!)</p>
<p>
we say... non responders may be non-responding for difft reasons...</p>
<p>
we do tractography with probabilsitic modoeling</p>
<p>
all 6 responders shared the same voxels...
what contact were you doing in all the responders?.</p>
<p>
(must get the medial frontal area)</p>
<p>
we can drive around a tract map</p>
<p>
involves SCC25 ; Forceps minor cinglulum; and uncinate...</p>
<p>
75% response rate at 6 months.</p>
<p>
then went back and moved/ reimplanted trode and got another responder.</p>
<p>
how to look at this network... what is wrong with it...</p>
<p>
what is wrong in this network... their key question...</p>
<p>
shows <b>UF = uncinate fasciculus</b>...</p>
<p>
she posits.. this is a disconnection syndrome...</p>
<p>
recall... these people are paralyzed...</p>
<p>
the will to action (is gone... the area where you think to move... and it's gone)</p>
<p>
talking to Michelle Monje yesterday. (Helen is a "stalker " of her work)</p>
<p>
how abt role of white matter in TRS and DBS MOA</p>
<p>
thot to involve astroglia but NO... it is actually the oligodendroglia...
this is an adaptive myelination problem (and repair)</p>
<p>
Michelle: she works at 20 hz in beta band...</p>
<p>
MM does rodent work... stimualting and inducing myelin...</p>
<p>
please do this expt she pleads... (stimulate and check myelin in animals; hard to have trodes + fMRI)</p>
<p>
when you hit the right spot and stim in the OR.... you see the brake coming off and they know it...
there is change.</p>
<p>
early.. I feel lighter; I feel more connected;</p>
<p>
you didn't give me anything... you took something away (she compares it to driving with the emergency brakes on)</p>
<p>
eg one of her early patients... the void is gone... she said.</p>
<p>
we kept patients in hospital for a week... there were well at the end of a week.</p>
<p>
we had the clue early..</p>
<p>
seems like we were priming them with our stim in the OR...</p>
<p>
area 25 down and frontal up... that's what we want...</p>
<p>
we tried to measure our early effec tin the OR... but no place to do expt...</p>
<p>
we measure LFP... off the contact...
do frontal eeg; eCOG;</p>
<p>
<b>Characterizing the “Depression Switch</b>”</p>
<p>
<b>type 1 interoceptive change</b>... I feel lighter; I fel less heavy; I can breathe
the tension is gone;
the pain is gone (this is a <b>cingulate effect</b>)</p>
<p>
type 2: exteroceptive change </p>
<p>
medial frontal cortex... self reference...</p>
<p>
tracking maximal autonomic correlates...</p>
<p>
No patient said they liked it.. they say... brakes come off...</p>
<p>
in one patient , she had a wave of warmth... a pleasant rapid rush (but it does NOT hold...</p>
<p>
unsustained (like a hit of nicotine that you really need)</p>
<p>
we want to characterize the depression switch... its anatomy-physio-behav in the OR.</p>
<p>
what happens when the brain first sees the trode stim in the responders... it happens right away.</p>
<p>
tends to be in theta and alpha...</p>
<p>
we did parallel animal studies...</p>
<p>
stimulate infralimbic (in amygdala)</p>
<p>
we can measure LFPs in subcallosal cingulate several times a day... and track it...</p>
<p>
turn L side on then turn R side on...</p>
<p>
stim L the R later over an hour... see coupling effect...</p>
<p>
get longer and longer behav response everyday...</p>
<p>
contrast sick pt at 1 month post-op in the OFF state versus
when pt is WELL at 6 months... v responsive brain...</p>
<p>
want to figure out what the components of this circuit are...</p>
<p>
want to do pathway model dissection in model animals...</p>
<p>
eg forced swim test in rodents...</p>
<p>
back to the patient...</p>
<p>
so, can DBS do it all?</p>
<p>
she quotes a patient.. it didn’t give me anything... it took something away...(my depression)</p>
<p>
but you’ve got bad habits and now... you have to retrain habits and get unstruck...</p>
<p>
patients... they need to restructure their lives...</p>
<p>
patient to Helen: tell my husband that I’m still gonna have bad days...
cannot just turnup the juice to get thru bad days.
</p>
<p>
Q1 from audience: do u need to stim both L and R... A yes both sides!</p>
<p>
both cingulum bundles (and need forceps minor on one side)
if L crosses the midline that might be enough..</p>
<p>
from tractography I would be afraid to implant only one side.</p>
<p>
EEG is wildly asymmetric...
big effect is usually on the left...</p>
<p>
big asymmetry (even in gray/white junctions in cingulum)</p>
<p>
Q2 (Alan) shouldn’t we get investigators together on this and get some agreement...</p>
<p>
A some patients are badly implanted and some do NOT respond...</p>
<p>
people should not be walking around with this hardware if they do not respond...
it shoudl be taken out cuz (with it in) they will never be eligible for anything else...
could be just location of contact (of stim); did we miss the mark?</p>
<p>
Q3 Brian W: I work in tractography... your pic did not have great resolution...
you place a lot of emphasis on placement of trodes...
shouldnt the effort be to get the trodes in the right place...</p>
<p>
A: you’re exactly right; I went to Connectome folks...
is it forceps minor where it hits medial 10?
(I’m just a poor girl... don’t have high res in my shop)
do u find uncinate fasciculus first...
I sent a pt up to Minnes... it was a disaster.</p>
<p>
but I agree... we want to drive (to the correct place) in the OR to get to the right spot...
working with PRISMA scanner...
we need an algo for placement that will work thru the noise.
</p>