Biomedical Engineering Reference
In-Depth Information
We can record this activity from a population of motoneurons (the output
neurons projecting to the muscles) by suctioning a ventral root (axons from the
motoneurons) into an electrode. The activity recorded this way is a combination
of two signals. The slow component, illustrated in Figure 1A, represents the
depolarization (increase of membrane potential) of the motoneurons, propagat-
ing passively along the axons. Superimposed on this slow signal is a fast signal
(not visible in the figure) caused by the action potentials generated by the moto-
neurons. The slow signal shown in Figure 1A is a good indicator of the activity
in the whole network, since motoneuron depolarization is caused by synaptic
inputs from other neurons in the network. The activity is episodic, with episodes
lasting up to a minute separated by intervals of up to 20 minutes. Within an epi-
sode, the activity is rhythmic with a cycle frequency of ~0.2-1.0 Hz, markedly
decreasing toward the end of the episode. Each cycle can be seen as a network
"spike," representing the depolarization of the whole neuronal population. Al-
though the neurons are activated in synchrony, their action potentials are not
synchronized.
How is this activity generated? Several key experimental findings provide a
working hypothesis.
The first thing to note is that many developing networks can be considered
as purely excitatory circuits. This is because the inhibitory neurotransmitters
GABA and glycine have excitatory effects early in development (2,6).
1
Indeed,
blocking the action of excitatory neurotransmitters acetylcholine and glutamate
does not prevent the spinal cord from generating episodic activity (7). It is there-
fore easy to understand the presence of spontaneous activity in immature cir-
cuits. Any event such as a few neurons randomly firing can be amplified by
positive feedback through the recurrent excitatory projections, leading to mas-
sive activity in the whole network. This explains the activity, but not its episodic
pattern. How are the episodes terminated, in the absence of inhibitory connec-
tions?
One possibility is that the network "fatigues" during activity, until it is no
longer capable to sustain activity. In other words, there is an
activity-dependent
process that depresses the excitability of the network. To demonstrate the pres-
ence of such activity-dependent depression process, we have stimulated popula-
tions of interneurons through sensory or propriospinal afferents and observed the
evoked synaptic response on the motoneurons (11). The synaptic potentials re-
corded on the motoneurons are decreased after an episode of activity, and pro-
gressively increase during the interval between episodes. This suggests that
network excitability is depressed by the episodes of activity and that it recovers
in the interval between episodes. This depression may be synaptic, that is, activ-
ity decreases synaptic efficacy. An activity-dependent synaptic depression proc-
ess, by reducing the strength of the connections between neurons, would
effectively decrease the positive feedback. Another possibility is that this de-
pression acts on the neurons to decrease their excitability, making them less
