Stroke and neurocognitive deficits may follow cardiac surgery and have been linked to perioperative cerebral embolism. Alteration of cardiopulmonary bypass (CPB) or surgical technique to reduce embolism is, therefore, a rational neuroprotective strategy. Pharmacological cerebral protection has been advocated as an ideal "back-stop" to such "physical" interventions. A series of relevant studies conducted at Green Lane Hospital, Auckland, New Zeatand is described. Doppler ultra sound was used to monitor cerebral embolism during left heart valve surgery. Subsequently, salvaged CPB circuits were used to investigate several unexpected sources of emboli. The efficacy of a novel left heart deairing technique was audited using the Doppler de vice. Finally, a randomized double-blind trial of lidocaine in cerebral protection during cardiac surgery was con ducted. Most cerebral emboli were recorded after aortic declamping. However, cerebral emboli counts increased during stable CPB when the hard shell venous reservoir (HSVR) was operated at lower blood volumes and when air was seen in the venous return line. In vitro 2 HSVRs were found to generate bubbles when operated at blood volumes in excess of the manufacturer's recom mended minimum. Air in the venous return line was found to readily transit the CPB circuit and vacuum- assisted venous drainage markedly exacerbated this phenomenon. The novel deairing technique was mark edly superiorto conventional methods. Lidocaine admin istered in a standard antiarrhythmic dose for 48 hours from induction of anesthesia reduced the incidence of neuropsychological deficits at 10 days and 10 weeks postoperatively.