To investigate further the cholinergic specificity of the effects of basal forebrain lesion-induced disruption of attentional performance, the present study examined the efficacy of various pharmacological agents in improving performance of a five-choice serial reaction time task in rats that had received lesions of the cholinergic basal forebrain. Specifically, the effects of the novel 5-HT3 receptor antagonist, ondansetron (0.3, 1, 10 ng/kg), and of nicotine (0.03, 0.06, 0.1, 0.3 mg/kg) and the anticholinesterase, physostigmine (0.05, 0.1 mg/kg), on attentional function were examined in animals which had received AMPA-induced lesions of the nucleus basalis magnocellularis (nbM). The behavioural impairments observed immediately following the lesion were a reduction were choice accuracy and an increase in correct response latency. Although these impairments showed recovery over the course of the following weeks, the deficit in choice accuracy could be reinstated by reducing the duration of the visual stimulus and thus increasing the attentional load placed on the animals. This reduction in choice accuracy could be dose dependently improved by systemic administration of either physostigmine or nicotine, suggesting that this impairment in attentional function may be attributed to disruption of cholinergic function. The pharmacological specificity of these improvements was supported by the inability of d-amphetamine to improve task performance (0.2, 0.4, 0.8 mg/kg). Ondansetron was also unable to improve accuracy of performance in lesioned animals, but was effective in reducing the anticipatory or premature responding observed in both control and lesioned animals, even when elevated (in the case of controls) by treatment with systemic d-amphetamine. The results of the present study therefore suggest that cholinergic dysfunction can lead to attentional impairments which can be ameliorated by cholinergic treatments such as physostigmine and nicotine, but that ondansetron, despite its proposed ability to release cortical acetylcholine, was unable to restore choice accuracy at the doses employed. The results further suggest a double dissociation of effects on accuracy and the disinhibition of responding.