A research method for investigating delusional beliefs is outlined by adopting the delusional belief that one is dead (the Cotard delusion) as a model delusion. Detailed analyses of published case reports of the Cotard delusion demonstrate that the term 'syndrome' as it is currently applied to the belief that one is dead is not helpful in terms of our understanding of the delusion. Four new case studies of the Cotard delusion suggest that preoccupation with belief may play a role in the formation and maintenance of delusions. Preoccupation with delusional belief was investigated using a variant of the 'emotional' Stroop paradigm, commonly used in investigating anxiety disorders. Three individuals with the Cotard delusion, and diagnosed as suffering from depression, showed attentional biases toward words related to the theme of death. Two of the individuals had no history of anxiety and showed no bias toward words related to generalised anxiety. It was therefore suggested that the locus of attentional biases in delusions may be preoccupation with delusional belief, rather than anxiety per se. Consistent with this, a patient with fixed grandiose delusional beliefs, diagnosed with schizophrenia, and not suffering from anxiety, showed attentional bias toward words related to his delusional beliefs. Attentional bias failed to be demonstrated in a group of people with delusions arising in the context of schizophrenia, and reasons for this are discussed. Finally, three groups of individuals, who were free form any form of psychopathology, each showed a trend towards longer colour-naming times towards words related to their respective interests, but none of these were significant. It is concluded that attentional biases in delusions serve to reinforce delusional beliefs by constantly focusing die individual's attention onto delusion- relevant material. Implications for further research are discussed.