Biomedical Engineering Reference
In-Depth Information
with synapses subject to a fast depression (a small fraction of inhibitory cells
didn't disrupt this activity). These population spikes are generated through a
similar mechanism as the episodes described here, but on a much faster time
scale. During a population spike, most neurons fire only once, but they become
transiently synchronized (41). Such synchrony among neuronal populations
could play an important role in cortical information processing, and Loebel and
Tsodyks have suggested some applications of the transient synchronization al-
lowed by population spikes, by considering whether transient input signals elic-
ited a response (i.e., a population spike) or not from the network. Success in
triggering a population spike depends on the level of depression of the synapses
in the network at the time of the stimulus. In other words, the response of the
network to an input depends on the (short-term) history of network activity (23).
On the other hand, it is possible that the episodic activity, which is charac-
teristic of developing networks, is an abnormal mode of activity in mature net-
works. For example, the episodic activity in cortical slabs mentioned above is
due to the much decreased number of synaptic connections once the slab is iso-
lated (39). Also, some forms of epileptic activity resemble the spontaneous epi-
sodic activity of developing systems. Indeed, epileptic events can be caused by
an impairment of inhibitory synapses, or by an anomalous level of excitation.
An example of "epileptic activity in vitro" was shown by Staley et al. (33), who
recorded spontaneous episodic bursts of activity in hippocampal slices that were
disinhibited or rendered hyperexcitable. They showed that these bursts were
regulated by synaptic depression, as in the s -model. Alternatively, a model by
Traub and Dingledine (40) of "epileptic" bursts in hyperexcitable hippocampal
networks proposed that bursts are terminated by a slow hyperpolarizing current,
that is, a cellular type of depression.
It is critical to know whether these bursts are terminated by a synaptic or
cellular process, if one wants to choose an appropriate pharmacological treat-
ment. As a thought experiment, suppose that we want to suppress "epileptic
bursts" in a hyperexcitable hippocampal network. Should we target cellular ex-
citability or synaptic connections? If, for example, the bursts are terminated by
synaptic depression as in the Staley et al. (33) experiments, our results with the
s -model suggest that we should use a pharmacological agent that blocks excita-
tory synapses, as this will increase the interval between each burst. Decreasing
cellular excitability would also increase the interburst interval, but it would in-
crease burst duration as well. Similarly, the effectiveness of a drug potentiating
inhibitory synapses in order to stop the bursts would depend on the type of de-
pression mechanism that terminates the bursts and on the effects of the inhibi-
tory connections—phasic or tonic.
Much recent work has been aimed at understanding the role of inhibition in
neuronal networks. Although our approach has emphasized excitatory networks
in vertebrates, it should be pointed out that circuits of inhibitory neurons may
also produce oscillations (31,42,43). A great deal of future work should be con-
Search WWH ::




Custom Search