CHICHILNISKY, EJ Retinal Processing Mar 9, 2015
<p class="MsoNormal" >CHICHILNISKY, EJ Retinal Processing Mar 9, 2015
(This is an outstanding lecture on a crucial topic and presents
a surprising new view of retina (and hence, of the brain.)</p>
<p>
EJ presents an intro for the bioengr students to the work of Fred Rieke
who will present his work on Thursday Mar 12th in Clark S360.</p>
<p>
EJ is from Depts of Neurosurg and Ophthalmology.
(He recently came to Stanford from UCSD.)</p>
<p>
This is a seminar in BioX at Clark S361 at noon.</p>
<p>
These talks start at 12:15 (and lunch is at noon.)</p>
<p>
he pron. Rieke REE-key. (a former UCSD colleague.)</p>
<p>
he begins...</p>
<p>
how to understand neural circuits....</p>
<p>
he shows a collection of them...</p>
<p>
fly visual system
superior colliculus for orienting toward space...
HC hippocampus for memory
Cadherin-defined sensory tracts
worm mechanosensory system
olfactory system...</p>
<p>
one v accessible system is the retina...</p>
<p>
vision... our most important sense modality...
and <b>uniquely approachable.</b>
<b>can dissect it and probe it in vivo</b>.</p>
<p>
EJ is in the HEPL Bldg (Hanson Appl Physics) with Daniel Palanker.</p>
<p>
visual processing... seems effortless...
how to navigate to catch a frisbee amidst a crowd of players.</p>
<p>
Illusions... two identical squares in visual intensity;
one on a light background and the other on a dark background.</p>
<p>
so, even brightness is entirely non-trivial.</p>
<p>
more illusions from Adelson in 2000... seeing squares thru a transparent strip.</p>
<p>
he says stimu LIE (not lee);</p>
<p>
about 1M nerve fibers from EACH eye.</p>
<p>
shows layers of retina... PRs = photoreceptors on the outside (away from the light)</p>
<p>
PRs connect to interneurons then to the RGCs which to to brain.</p>
<p>
the view he presents is quite difft than the textbook view...</p>
<p>
PR, horizontal, bipolar, amacrine, then RGCs...</p>
<p>
4 types of PRs; 2 horiz types; 15 bipolar types, 25 amacrine types
and about 20 RGC types.</p>
<p>
4PRs: 3 types of cones and 1 type of rod.</p>
<p>
horiz and amacrine are laterally oriented and
15 types of bipoalr cells connect front to back (in the visual stream)</p>
<p>
huge amt of visual processing in retina.</p>
<p>
105M photoreceptors:
5M cones and 100M rods
and 1M RGCs.</p>
<p>
so, massive convergence of info to RGCs...(100 to 1)</p>
<p>
graded potentials mainly exist in the retina (ie no spiking) (until you get to RGCs)</p>
<p>
(I ask abt graded potentials in the brain...
he thinks there are spikes in mammals simply because of size...
spikes are used for distance communication.
(graded potentials seem to be the case with worms, flies, etc... small brains)</p>
<p>
<b>RPE layer is dark black </b> (and will not xmit light... so, that's a reason why PRs are in the back (and not the front) but that's not the case in some animals.</p>
<p>
diversity of cells (as above) really matters!
<b>20 difft RGCs... shape, morphology, connectivity, function, patterns of projection in brain and its light response properties... all v distinct</b>...
and each type has a full image! (each RGC type entirely TILES the retina!)</p>
<p>
shows RGCs from monkey (Dacey 2004) versus
mouse (Volgi in 2009)</p>
<p>
RGCs are the output of the retina.</p>
<p>
(and they may have difft genetic markers)</p>
<p>
RGCs huge diff. in size, connectivity etc.</p>
<p>
difft patterns of connectivity of the RGCs to the interneurons (to the amacrine cells)</p>
<p>
IPL is inner plexiform layer (it is adjacent to the RGC layer)</p>
<p>
IPL has many strata where amacrines (specific ones) connect to specific RGCs.</p>
<p>
<b>retina is really 20 different, tightly interwoven parallel circuits</b>.</p>
<p>
projections are difft (in LGN and elsewhere)</p>
<p>
<b>each RGC uniformly tiles visual space</b>.</p>
<p>
shows ON alpha RGCs from cat (from wassle in 1983)</p>
<p>
he shows that the difft types of horiz/ amacrine etc. all have complete tiling....</p>
<p>
so, each of these 20 RGCs capture difft and multiple visual reprs of the world.
and they project to multiple targets.</p>
<p>
<b>each RGC type is a little machine...(with a reproducible output</b>, just like a circuit)</p>
<p>
temporal resol of 1.2 msec !</p>
<p>
and very reproducible...</p>
<p>
shows on to off etc.
and see highly repeatable spiking pattern
ON parasols are exactly the same but at a difft location</p>
<p>
and totally difft than an OFF parasol..
and different than an ON midget and OFF midget etc.</p>
<p>
all ON parasols are carbon copies of one another and different from OFF parasols...
and difft from small bistratified.</p>
<p>
Q are the visual fields fluid or determined ; A: it seems to be determined.</p>
<p>
Gauthier in 2009... from a jigsaz puzzle repr of visual space...</p>
<p>
now what happens to this info...</p>
<p>
from Kuffler in 1953...</p>
<p>
RGC... center surround... highlighting contrast....</p>
<p>
yes... has brought (Kuffler) into the modern age.. it IS the case... but there is much more.</p>
<p>
Barlow 1964...
DS = direction-selective RGCs...</p>
<p>
only specific cell types do this.
(and some respond to color)</p>
<p>
also object-motion sensitive RGCs... which
respond to object jittering around...</p>
<p>
may be sensit to object moving with respect to background...</p>
<p>
looming motion (eg prey animal... with looming owl)</p>
<p>
mouse... mainly navigates with whiskers and smelll.</p>
<p>
why use rabbit retina: easy to get retina out.</p>
<p>
monkey... v similar to us... front mounted eyes (macaques)</p>
<p>
Rieke studies macaque monkey intermediate layers (rare lab that studies this)</p>
<p>
intermediate layers are v hard to study... not accessible.
does retina tell us abt the brain...</p>
<p>
need source of retinas... primate retinas... greatest hope of studying human retina.</p>
<p>
* summary</p>
<p>
multiple parallel complete visual images sent to the brain...</p>
<p>
so, brain may be THAT specific also... (as opposed to willy-nilly connections)</p>
<p>
see this same lesson in fish tectum (specific layers)</p>
<p>
same in lamination of V1....</p>
<p>
could well replicate in the brain</p>
<p>
<b>midgets are half the RGCs</b> (but when driving down the highway, the motion cells are impt)</p>
<p>
I ask about the retinal implant: he says it will be in patients in abt a year Daniel Palanker... this is the leading prosthetic system but (even it) does not pay heed to difft RGC types.</p>
<p>
is the brain a universal learning machine or is it specific...</p>
<p>
we have much better access to retina than other guys have to the brain,
so we are 30 years ahead of the CNS guys.</p>
<p>
Q... plasticity in the retina
A some... adaptation (but less than you would expect)...</p>
<p>
</p>
<p>
</p>