Biomedical Engineering Reference
In-Depth Information
grow over several days. At several points along the microtunnels, the loor of the tunnels contain
metal electrodes that allow for recording large (up to 200 μV) electrical signals, including their
direction and propagation speed.
MEAs, however, are passive signal detectors and require an extra amplifying step, which adds
noise because of the distance between the MEA and the ampliier. How about using the exquisite
voltage sensitivity of microelectronics components such as transistors (right below the cells) for
detecting small bioelectric signals? his was precisely the thinking behind the work by Jürgen
Weis' laboratory, then at the University of Ulm (Germany), who in 1991 pioneered the use of
ield-efect transistors
(
FETs
) for bioelectric recordings (
Figure 5.47
). In microelectronics, FETs
are used as switches; the voltage applied to the “gate” electrode is used to modulate the passage
of current between the “source” and the “drain,” two adjacent regions of diferentially doped
silicon. Not surprisingly, a custom-fabricated FET (without metal on the gate, to allow for contact
between the neuron and the gate oxide) was able to detect changes in membrane potential of a
Retzius neuron. he work raised many hopes, as the sensors are readily fabricated and integrated
with other sensors in large microarrays. he fact that they are opaque did not seem to deter a
wealth of research and development on cellular applications of FETs in the following two decades.
he biggest obstacle seems to be the temporal stability of the recordings, which are impaired
by the leaky seal between neuron and gate and (more fundamentally) by transistor breakdown
(electrolyte ions difuse into the silicon bulk and irreversibly alter the conductance of the device).
Recent developments on FET technology may enable new modalities of neuronal recordings.
Charles Lieber's group at Harvard has shown that silicon nanowire FETs (Si-NWFETs) can be
built on transparent substrates and interfaced with brain slices (
Figure 5.48
). Unlike traditional
FETs or MEAs, Si-NWFETs allow for imaging cells in various optical microscopy modalities
and provides highly localized measurements (their active surface is ~0.06 μm
2
), at submillisec-
ond temporal resolution and 30 μm spatial resolution.
he biggest drawback of MEAs and FETs is, of course, that they are only capable of recording
extracellular signals, that is, membrane voltages. However, electrophysiologists are even more
interested in learning about the intracellular signals that give rise to those membrane voltages.
Traditionally, the intracellular signals can only be recorded with patch clamp probes or patch
clamp chips (see Section 5.8), but a new revolution in electrode technology has come to the
rescue. Micha Spira's group from the Hebrew University of Jerusalem, in Israel, has developed a
new type of MEA consisting of gold, mushroom-shaped microelectrodes (gMμEs) that allow for
recording and stimulating intracellularly from arrays of neurons, while
the electrodes maintain
Stimulating
current
I
St
a
b
c
I
D
V
M
Membrane
potential
FET
Neuron
5 mV
Electrolyte
Field
oxide
Neuron
100 ms
10 µm
ME
I
D
Source
Drain
-
0
G
n
-Si
30 mV
+
Bulk
silicon
Source-
drain
current
V
M
+
+
Gate
oxide
Current pulses
(
I
St
)
FIGURE 5.47
FETs. to. record. the. electrical. activity. of. cells.. (From. Peter. Fromherz,. Andreas.
Offenhäusser,.Thomas.Vetter,.and.Jürgen.Weis,.“A.neuron-silicon.junction:.A.Retzius.Cell.of.the.
leech.on.an.insulated-gate.ield-effect.transistor,”.
Science
.252,.1290-1293,.1991..Figure.contrib-
uted.by.Peter.Fromherz.)
