Index Home About Blog
From: Mike Darwin <>
Subject: Freezing and memory
Date: 03 Apr 1995

    ---- CryoNet Message Auto-Forwarded by <> ----

Date: 03 Apr 95 03:15:03 EDT
From: Mike Darwin <>
Subject: SCI.CRYONICS Freezing and memory

Bruce Zimov makes some very good comments about cerebral ischemia and refers to
a book by Lutz, et al.  Most of his statements are correct, however
excitotoxicity due to NMDA receptor activation is not, in our hands or in the
hands of most others, a signifigant cause of ischemic or ischemia-reperfusion
injury in models of global brain ischemia.  We have tried NMDA receptor
inhibitors in a canine cerebeal ischemia (cardiac arrest) model with not only no
success, but with worse results than no drugs.

Excitoxcicity does occur however, and we are currently looking hard at kainate
and quisqualate receptor blockers as well as others which, for priprietary
reasons, I cannot talk about publically.  I can tell you that we have recovered
one dog named Jake after 10 minutes of cardiac arrest at 37.5xC followed by 5
minutes of reperfusion (using bypass) at a mean arterial pressure (MAP) of 155
mmHg or less and ANOTHER 5 minutes of reperfusion at a MAP of 30 mmHg or less,
followed by hypertensive , (MAP =140 mmHg) hypothermic reperfusion.  Jake was
blind for about 2 weeks and is still a little clumsy 3 months later.  But, he
know his name( (and he knew it during ICU care), he can still learn (albeit more
slowly). His only visible defect is that he will run in circles to the left if
you put him in a small pen or stress him with confinement: otherwise he is
normal with a good overall performce catergory (OPC) score (we use Peter Safar's
system and Jake scores a 5).

Reperfusion injury has concerned me for a long time, and we administer over half
a dozen drugs to our cryopreservation patients after beginning CPR following
cardiac arrest.  We also put people on PREmedication to protect against cerebral
ischemia when they become terminally ill.  For instance, some drugs such as
Dilantin (phenytoin), are very cerebroprotective if given *before* before the
insult, but don't show much utility if given afterwards.  This a complex area
and really beyond the scope of discussion here.

However, I wish to make one more point: while cerebral ischemia and reperfusion
injury concern me greatly and mitigating them has occupied much of my career in
cryonics, I am far more concerned about cryoinjury.  Indeed, I now feel that in
patients who retain good cerebral perfusion up until the time of cardiac arrest
the ischemic intervals they are experiencing in an optimum setting (3-10
minutes) are not irreversible given the technology that we (Biopreservation) has
right now and is developing (Jake is living proof of this).

However, once you've seen some EMs from frozen-thawed brains, even brains
treated with high concentration of cryoprotectant, you will realize that insult
from ischemia is the LEAST of the patients' problems.  I've described
frozen-thawed cryoprotected human brains as chopped steak.  I wasn't kidding.
The injury is tremenous and one of its most disturbing features is the
reorganization of intracellular structures in membranous vesicles.  Indeed,
finding intact brain cell membranes becomes a challenge.  Long processes are cut
and chopped by ice and many axons are empty or are just feilds of debris.  Of
course, this is post-thaw, post fixation, post a lot of things... and much of
the stirring of the pieces may have occured on thawing.  However, we now have
freeze-substitution pictures of brains in the frozen state after 4M glycerol
perfusion (we're getting ready to do similar studies with 7.4M and 1.5M) and it
isn't a pretty picture.

My point here is that at an EM level things are a mess.  To put things in
perspective I'll use the following analogy:

Take a picture of Kobe Japan from space (say 200 miles up) BEFORE the
earthquake.  Take on afterwards (and after the *fires are out*).  It will be
very hard to see the changes, but you can, if you look closely notice that the
gridwork and order of things is a little fuzzy and what were once clean lines
(like the bullet train track) have litle breaks in them.  But on the whole
things look very good.  This is *exactly* what you see when you look at brains
at the light level.

Now overfly Kobe at say 10,000 feet. Damage is visible everywehere.  Building
are collapsed, communications lines cut, water is flowing from broken pipes and
roadways are disrupted.  None of this is visible at 200 miles up.  Finally, walk
around on the ground with a camera.  Now you can really see the havoc!

How does this relate to brain cryoinjury.  Well, my suspicion is that gross
personality traits, temperament and some skills (such as walking) are encoded on
a level comparable to that seen with light microscopy.  More specific procedural
memories are probably encoded at the level of complexity comparable to an
overflight at 10K feet.  And declarative memories are encoded at the ground
level (synapse/receptor/membrane morphology changes).  I think contemporary
cryopreservation patients' general coarse wiring diagram may be inferrable, but
not easily, since there are tears and major disruption of connection at 10-30
micron intervals (these are visible at even the light level) throughout all
three dimensions of the brain.  When we get down to the ultrastructural level of
seeing membranes as lines on an EM things are really messed up. As to the
integrity of boutons and vesicles involved in hypothetical models of learning
mediated changes of synapses: well we haven't looked, but I'd be very surprized
if the news was good; especially when I see amorphous granular debris scattered
in ice holes, major morpholohical changes in intracellular and plasma membranes,
and so on.  Synapses: yeah, you still see 'em, they're pretty rugged.  But thats
often all you see is a synapse just sitting out there in field of debris.  Empty
remanents of axons with unravelled, swollen mylein, ice holes, naked nucei with
fragments of endothial cells still hanging on them in capillaries are all par
for the course.

When you look at ice in a frozen brain, you come to understand WHY these changes
have occured.  Maybe we will be able to infer things in frozen patient by
dismantling them in the frozen state.  But this is not all clear to me and many
of the membrane changes may be going on during freezing when the membrane
architecture is being reshaped by cooling and by exposure to an extremely
hyperosmolar environment.  Inferring the normal state from those changes is a
situation I doubt is possible given our current understanding of physical law.

This why I am so focused on improving brain cryopreservation and in looking at
brains frozen with current techniques after thawing *and in the frozen state* in
order to tell exactly what's happening.  The only major modality we have *not*
been able to bring to bear is the use of freeze-fracture which we cannot seem to
find any lab competent to do.  This is very important because it would allow us
to look at membrane morphology closer to the "street level"  (and in three
dimension) than we have ever been able to do so far.  Anyone who can point me to
a good EM center that does free lance freeze-fracture work of high quality would
be greatly appreciated.

Mike Darwin

From: Mike Darwin <>
Subject: CRYONICS: dendritic spines
Date: 13 Oct 1995

    ---- CryoNet Message Auto-Forwarded by <> ----

Date: 12 Oct 95 23:14:55 EDT
From: Mike Darwin <>
Subject: CRYONICS: dendritic spines

I did not get a copy of the Cryonet with Mike Perry's article on dendritic
spines.  Joe Strout was kind enough to send it along.

I was aware of this work.  Furthermore, while it is nice systematic work which
*quantifies* what happens it is not a surprise.

Over a decade ago Jerry Leaf, Hugh Hixon and I looked at dog brains at post
mortem intervals out to 24 hours.  We observed the same phenomenon at an
accelerated rate since we were exposing the animals to room temperature ischemia
(24C) throughout the process.  The breakover point beyond which there was
largely hash (no synapses lots of debris) at the ultrastructural level was
betweenm 6-8 hours if I remember correctly.

There is lots of other literature out there on post-mortem ultramiscroscopic
changes.  I had a full file drawer on this when I left Alcor, and now have a
drawer and a half.

No one paid much attention to this work.  To his credit, I can remember talking
with Mike Perry about autolytic changes and cryoinjury and the questionable
inferrability of the healthy state of the tissue, in the crew room over at Alcor
when they were in Riverside and I was on staff, and he was very concerned.  He
should have been.  So should have a lot of others.

For years I have been preaching about how rapidly some changes occur postmortem
and how quickly phospholipases, complement and other degradative processes are
activated: not only in global ischemia (cardiac arrest) but in the long, pre
arrest shock phase.

The notion that the brain is getting adequate flow and not experiencing serious
injury during premortem shock is ridiculous and I have the data to prove it.

Furthermore, I have for years harped on the very troubling clinical phenomenon
observed in humans who sustain head injury with prolonged unconsciousness
usually in the presence of cerebral edema.  Most people injured in this fashion
do not recover.  Those that do are often profoundly compromised neurologically
running the spectum from near vegatative to severely "retarded" and child-like
often with serious motor and language deficits as well.

However, a SMALLER subset of these patients recover surprisingly well, in fact
the only way to determine their history of head injury is by giving them fairly
sophisticated neurological tests (here I speak of cognitive tests, not invasive
ones). Of the cohort that recover well a small but significant number have a
complete or nearly complete loss of declarative memory which does NOT resolve or
improve with time.  There is sometimes also frontal lobe disinhibition (I used
Roseanne Barr as an example of  frontal lobe injury related disinhibition) and
alterations in *character* and *personality*, but these are not universal. Such
people usually do not lose their skills (procedural) memories.  This suggests
that there are different mechanisms for the two kinds of memory.  A tired
example I trot out regularly is the one of piano playing.  The essentials of
playing the piano or driving a car can be easily inputed into declarative
memory.  However, mastery of these skills does not come with "book" knowledge or
"verbal knowledge" but rather with many repetitions and much time and effort.

Additional evidence, albeit indirect, of this seperation of mechanisms of
encoding declarative and procedural memories is most dramatically demonstrated
in Alzheimer's (ALZ) disease.  In the early to mid-phases of the disease the
victims can still drive.  A careful analysis was just completed on the rate of
accidents in ALZ patients versus pair matched (by age, sex , education etc.)
normals.  The accident rate was the same for the ALZ group as for the normal
controls.  Driving is like walking or riding a bicycle.  However, while the
accident rate was the same (indicating conservation of skills memory) the ALZ
group got lost and ended up coming into police or social agency custody at many
times the rate of the normals.

Similarly, both Dr. Steve Harris (Steve is Gerontologist) and I have had the
experience of having a totally uncommunicative and demented ALZ  patients sat
down at the piano and their hands placed on the keys; they can often play
beautifully, usually only one or two pieces, but sometimes a wide range.  Here
too we see obliteration of declarative memory and relative conservation of
skills  memory (note they must of course have FIRST known how to play the
piano!).  A look at the frontal and temporal cortex of these patients (who have
entered the final phase of the disease) at autopsy shows massive cell loss,
plaque formation and neurofibrillary tangles.  There is, apparently, no
structure meaningfully preserved in advanced cases of this disease in these

There are of course, apologists for this phenomenon of selective loss of
declarative memory who posit that the memories are not really destroyed, but are
inaccessible.  This certainly may be true, but we have no compelling evidence
that this is the case.

If we look at various proposed mechanisms for storage of declarative memory and
procedural memory we are looking not only at synapse formation or elimination,
but at structural and biochemical changes within synapses and within brain cells
themselves (i.e., cell bodies).  I doubt seriously if these are very robust in
the face of autolytic degradation (self digestion) and in the face of highly
destructive free radicals which are unleashed in the shock and global ischemic

Brian Wowk has, mistakenly I believe, minimized the significance of this kind of
injury in "optimum" cases and has pointed out that a substantial number (maybe
1/3rd) of cryopreserved people were treated under "optimum conditions."  I take
issue with this.  A large fraction of the "optimum" cases are far from optimum
and they are exposed to many hours of shock, trickle flow and finally ischemia
and reperfusion injury.  The immune-inflammatory cascade and the accompanying
activation of structure degrading processes is well underway in most optimum
cases before legal death is ever pronounced.  Far from being better than the
guniea pig model, these cases may be worse because they are exposed to injury
exacerbating conditions and active autolysis many hours before legal death,
cooling and blood washout occur.  We know from our continued deep hypothermic
blooodless perfusion work on dogs that this cascade, once activated is not
STOPPED by cooling, although its speed and severity are reduced.

I have spent 15 years of my cryonics career trying to quantify and understand
qualitatively what is happening in ischemia (of which shock is one species).  I
am not reassured by what I have found in cases of prolonged insult, folllowed by
flush-ice store-followed by CPA perfusion.  I also spent that time trying to
block ischemic injury and I believe I and others working with me have made great

Nevertheless, the fact remains that the majority of cryonics cases will not be
done under conditions where prolonged ischemic insult ante- or post- mortem has
not occurred.  It is questionable whether such patients will be recoverable with
their declarative memory intact.

The take home message here is quite simple, but one which I doubt will be
carried off by many: garbage in = garbarge out.  Ten minutes of ischemia is not
trivial. Ten hours is a disaster.

I for one am through running around with rosy scenarios.  We need solid data and
we need to mitigate this injury.  That is exactly what we are doing here and the
end point must be reversible brain cryopreservation.  Otherwise, as a growing
statistical base has demonstrated, 1/3rd (minimum) of patients are just going to
get hashed up in various imaginative ways: by not being found for hours after
cardiac arrest, by being autopsied, by long delays, etc.  And, many of our
so-called BEST cases will be far from it in reality.

The agony of worry and uncertainty, the lack of social and medicolegal change
will destroy or gravely injure many cryonicists and, perhaps as importantly or
more so, will make cryonics an "unbelieveable" option for most.  This represents
a staggering death toll.  It can only be avoided or mitigated by clear, rigorous
procedures which result in feedback.  Reversible brain cryopreservation is
certainly one milestone along the way.  Perhaps the most importat one.

Mike Darwin

Index Home About Blog