Electrocortical and behavioral arousal are separate phenomena subserved by different neural substrata operating in parallel. A comprehensive theory of 'activation' must take into account the relationships between the electrical and behavioral activating systems. In pathological or experimentally induced states paradoxes, resolvable by a theory positing functional interaction between these systems, arise. EEG arousal is directly mediated, in both the waking and sleeping state, by cholinergic mechanisms. Antidepressant withdrawal precipitates cholinergic overdrive; this would account for the apparent disturbances of REM sleep occurring when antidepressants are stopped. Generally, cholinergic overdrive would produce behavioral inhibition but in particular instances it triggers marked psychomotor arousal by mobilizing a 'limbic activating system'. The existence of a monoaminergic 'limbic activating system', system 'A', with the properties attributed to it in this paper, is supported by both clinical and laboratory observations. System 'A' theory provides a parsimonious means of adequately explaining many phenomena. This theory also has in its favor explanatory power and scope. The Cholinergic-Monoaminergic Interaction Theory of antidepressant withdrawal induced activation and of rapidly-cycling manic-depressive illness maintains that system 'A' and a cholinergic inhibitory system interact dynamically, and that excessive monoaminergic function can precipitate excessive cholinergic function and a dearth of monoaminergic function (due to autoregulation) and hence depression. Likewise, excessive cholinergic function is posited to activate monoaminergic systems and hence to secondarily cause behavioral activation. Rapidly-cycling manic-depressive patients, according to the model, develop alternating cholinergic and monoaminergic overdrive states because the homeostatic mechanisms which should serve to maintain, within normal limits, the composite of cholinergic inhibitory and monoaminergic activating influences are defective. Consequently, rather than reaching a reasonable balance compatible with adaptive function there is oscillation between extremes. Each oscillatory movement is actually a move towards the 'golden mean' and is induced by deviation from this ideal but the defective homeostatic mechanisms promote ' perpetual ' overshooting. Lithium and ECT may be useful in the treatment of rapidly-cycling patients as both treatments may down-regulate muscarinic receptors, and otherwise modify cholinergic and monoaminergic systems in ways promoting homeostasis.(ABSTRACT TRUNCATED AT 400 WORDS)