Computational aspects of anesthetic action in simple neural models.

THE VOLATILE and intravenous anesthetics modulate activity of a variety of voltage-gated and ligand-gated ion channels. However, the mechanism(s) by which one or more of these effects produces general anesthesia and the specific site(s) of action which produce this state are presently unknown. There is growing recognition of the need for integrative approaches in order to relate anesthetic effects on ion channels with systems level behavior. However, the difficulty of determining how modulation of ion channel activity can produce general anesthesia is compounded by the relative paucity of information as to what constitutes general anesthesia at the systems level. Here, neural models are used to link anesthetic modulation of ion channel activity with overall systems level behavior using computer simulation and the analysis of the mathematical structure of some of these models. This complements recent work focusing primarily on anesthetic modulation of network interactions in more abstract and more complex models. It has been hypothesized that the in vitro concentration-effect curve of the receptor responsible for general anesthesia should be similar to that seen in vivo at the systems level, where attention is generally paid to the slope and midpoint of such curves. An important feature of the in vivo concentration-effect curve that must be accounted for is its threshold-like behavior as anesthetic concentration is increased. Since anesthetic concentrations that lead to immobility lie well below the EC50 of the relatively shallow concentration-effect curves for volatile anesthetic suppression of voltagegated ion channel activity, the above hypothesis has been used to implicate the ligand-gated ion channels as the site of action for the volatile anesthetics. The ligand-gated ion channel thought to be most involved in the mechanism of action of the volatile anesthetics is the GABAA receptor which, with greater certainty, is the site of action for the intravenous anesthetics thiopental, propofol, etomidate, and the neurosteroids. Yet, volatile anesthetic action may not be completely understood in terms of anesthetic modulation of GABAA. Knockout mice lacking a component of the GABAA receptor are readily anesthetized with volatile anesthetics, although this may simply reflect that the component of the GABAA receptor necessary for volatile anesthetic action is still present in the knockout. Furthermore, the nonhalogenated anesthetics cyclopropane and butane have little effect on the GABAA receptor which may also be true for the gaseous anesthetic xenon. However, if voltage-gated ion channels are relevant in general anesthesia, it is necessary to appreciate how relatively modest anesthetic effects at the ion channel are integrated to produce general anesthesia. A model which links anesthetic action at the ion channel with systems level effects requires specific hypotheses as to the nature of general anesthesia at the systems level. Clinically, general anesthesia consists of at least loss of consciousness, amnesia, immobility, and blunting of autonomic reflexes. For inhalational anesthetics, anesthetic potency is generally quantified by using immobility as the endpoint. However, other attributes of general anesthesia appear at anesthetic concentrations that differ from those that lead to immobility. Moreover, immobility requires much greater anesthetic concentrations in preparations which provide anesthetic to only the brain and not the spinal cord. Thus, anesthetic effects at the systems level may be multimodal and associated with particular regions of the central nervous system. Quantifying anesthetic effects at the systems level has been limited to analysis of the electroencephalograph and functional imaging of brain activity. Use of the electroencephalograph to characterize anesthetic depth has been complicated because electroencephalographic signals are not invariant with respect to anesthetic agent, and are not readily associated with specific endpoints such as immobility, though some progress has been made. Nonetheless, several features of the electroencephalograph are generally observed. For the intravenous anesthetics (barbiturates, propofol, and etomidate) a rather characteristic pattern is observed. At lower, generally subanesthetic, concentrations there is an increase in the fundamental frequency of the raw electroencephalographic signal, followed by slowing and increasing amplitude of the electroencephalograph with increasing anesthetic depth and, ultimately, burst suppression and isoelectric behavior are observed. Administration of benzodiazepines leads to loss of the rhythm * Associate Professor, † Research Assistant.