Biomedical Engineering Reference
In-Depth Information
Because the surface charge redistribution (both in the electrolyte and in the CNT
meshwork) occurs in conducting media, a characterization of the physical system
also in terms of distributed parameters could be more appropriate. However, on the
basis of the electrochemical considerations employed in metal-electrolyte inter-
faces (Robinson
1968
), we decided to focus only on the frequency range that char-
acterizes (extracellular) neuronal signaling (i.e., 100 Hz-10 kHZ). Within such a
bandwidth, the model of Fig.
4c, d
is adequate to describe the signal transduction at
the interface, and we further consider the frequency-independent resistance and
capacitance as correct, lumped descriptions of our system.
3.2
The CNT-Neuron Junction
Based on the work of Grattarola and Martinoia (
1993
), we now introduce a method
useful to describe the biophysical and electrochemical properties of the interface
between CNTs and neurons. As already mentioned for the CNT-electrolyte interface
(Fig.
4c, d
), such an approach consists in the formulation of linear electrical circuit
models composed of capacitors and resistors, which are equivalent to the differential
equations describing the temporal evolution of the extracellular electrical variables
of interest. This description shares obvious similarities with the classic approach
employed for the electrical excitability in neurons (Hodgkin and Huxley
1952
;
Fig.
7b
). The exploration and applications of these concepts to the description of the
interface between metal and the neuronal membrane gained renewed interest in the
last few years, thanks to the increasing availability of multisite in vitro electrophysi-
ological chronic recordings from cultured networks of dissociated neurons in neuro-
electronic studies (Fig.
5a-b
) (Rutten
2002
and references therein). In fact, the specifi c
features of the voltage waveforms, detected extracellularly by means of substrate
metallic microelectrodes (Fig.
5d
), could be related to the intracellular action poten-
tials (Fig.
5c
) “fi ltered” by the biophysical and electrochemical properties of the cell-
electrode coupling and of the electrolyte (Grattarola and Martinoia
1993
; Martinoia
et al.
2004
) .
Figure
5
reports the intracellular and extracellular recordings of the spiking
activity, which emerges spontaneously in primary cultures of neurons (Marom and
Shahaf
2002
). Simultaneous patch recordings (Fig.
5a
) reveal the typical intracel-
lular signals expected during action potentials, and unveil subthreshold, network-
driven synaptic activity that underlies the spontaneous occurrence of bursts of action
potentials (Fig.
5c
). On the other hand, the same electrical activity, detected extra-
cellularly by thin-fi lm substrate microelectrodes (Fig.
5b
), allows enhanced spatial
resolution at the expenses of the temporal resolution: indeed, only suprathreshold
(i.e., spiking) activity is detected extracellularly (Fig.
5b, d
). On the basis of our
current knowledge, CNT-based substrate electrodes are expected to behave qualita-
tively as metallic extracellular electrodes, with unique and unparalleled physical
properties. Through our biophysical modeling approach (Fig.
8
), one can quantita-
tively appreciate how these features boost signal-to-noise ratio (Fig.
8c
) and record-
