Biomedical Engineering Reference
In-Depth Information
aCC motoneuron
Stage reference
a
b
Probe ref
Probe
Backbone
A
Axotomy
B
GFP
< 5% of intact length
5th axon severed
c
Intact axons for comparison
FIGURE 6.62
Neurotransmitter. clustering. induced. by. mechanical. forces.. (From. Scott. Siechen,.
Shengyuan.Yang,.Akira.Chiba,.and.Taher.Saif,.“Mechanical.tension.contributes.to.clustering.of.neuro-
transmitter.vesicles.at.presynaptic.terminals,”.
Proc. Natl. Acad. Sci. U. S. A.
.106,.12611-12616, 2009.)
6.5.4 Emergent Properties of Neuronal Networks
One of the paramount goals of neuroscience is to understand how brain behavior emerges
from the summed electrophysiological activity of each neuron. Unfortunately, no technology
is within reach to record (or make sense of) all neuronal activity in a mammalian brain, so
researchers are forced to study simpler systems, such as the intact nervous systems of slugs or
worms, brain slices or ganglia, or neuronal cultures. Each system has its technical advantages,
presents its own technical challenges, and has a horde of supporters and detractors that ight
about their biological legitimacy with a vehemence that is probably unnecessary, because each
system has a well-established trade-of between tractability and complexity.
6.5.4.1 Bottom-Up Approach: Neuronal Cultures
Dissociated neuronal cultures represent a “bottom-up” approach to try to understand the very
basic properties of neuronal networks, with the advantage that cells can be accessed for record-
ing by surface microelectrodes. However, the architecture of the networks lacks the complex-
ity found in vivo and the cell's physiology is likely irreversibly damaged by the isolation and
culture protocols. Early work in 1975 by Paul Letourneau, then at Stanford University (
Figure
2.25
), showed that dissociated neurons were able to grow along micropatterned proteins, which
promised chips that would wire up a “brain on a chip.” In 1988, Bruce Wheeler's group at the
University of Illinois at Urbana-Champaign used MEAs to record from an
Aplysia
ganglion and,
later, applied correlation algorithms to record from neuronal cultures. Jerry Pine's laboratory
at Caltech has developed “neuro-cages” since the 1990s to facilitate trapping neurons on MEA
arrays (
Figure 5.16
).
he extreme diiculty of correlating the information being recorded with a real neurophysi-
ological process has tempered the early feeling that these cell culture systems could be used to
understand the whole brain (“Brain on a Chip”) and has changed the focus of the ield to more
practical goals. Steve Potter's group, then at Caltech, using rat cortical neurons randomly seeded
on a MEA, claimed that the signals recorded by the MEA could be used to control an artiicial
animal or “Animat.” Special sotware recognizes, in approximately 8 minutes, patterns of activ-
ity that arise spontaneously from the neuronal network and uses them to tell the Animat which
way to move. he Animat sends a signal back to designated electrodes in the MEA when, for
example, it collides with a wall—a mimicry of a sensory input into the “brain” on the MEA.
Eventually, it is hoped that such cultures could be used to control “intelligent” robotic devices.
However, the randomness of the connections that arise in the dish make it diicult to envision
