The present study was designed to evaluate the integrity of cognitive fronto-temporal processes in drug naive patients with schizophrenia. The evaluation of drug naive patients discards the potential influence of medication, and may allow the specification of cognitive impairments that are truly illness-related. Subcomponents of long-term memory as well as several measures of attention were examined. A group of 16 patients who had never taken antipsychotics and a group of 20 normal controls underwent tests of alertness, information maintaining, and sustained and selective attention, as well as tests of explicit and implicit recall. The psychopathological manifestations of patients were also assessed with the BPRS, PANSS, ESRS clinical scales. Attention test performances revealed that drug naive patients presented a decrease in their ability to respond promptly to a stimulus, sustain their attention on a task, display normal selective attention strategies, and maintain information for on-line processing. The results also suggest that the drug naive patients are impaired when both strategic and associative processes must be triggered to explicitly recover information in long-term memory. In contrast, the results revealed that implicit access to perceptual mental representations is spared in schizophrenic patients. Finally, features of the patients' clinical symptomatology and some cognitive deficits were also shown to be correlated. Overall, results showed that, in relation to normals, drug naive patients were mildly impaired, with little intersubject variability, and that not all cognitive processes were equally disturbed in relation to the normal subjects' performances. Results support the idea that an important part of the impairments seen in schizophrenia is present before the introduction of neuroleptic medication and chronic illness.