The complex interrelationships among tricyclic antidepressant (TCA) effects, depression, and dementia in elderly patients are reviewed. Although results are mixed, some studies have shown TCA efficacy in elderly patients with organic disorders and associated depressive symptoms, such as forced weeping in pseudobulbar states. Available data, although not clear-cut, indicate that memory deficits or overt confusional states directly attributable to TCAs are relatively rare in patients without preexisting organicity. Studies in normal nonelderly subjects suggest an interaction wherein cognitive improvement due to relief of depression competes at higher doses with drug-induced memory impairment. In older depressed patients, significant relationships have not been found between degree of response to TCA and either baseline cognitive impairment or cognitive impairment during drug trials. Results from an ongoing study comparing imipramine and trimipramine in older depressives suggest that clinical improvement outweighs the anticholinergic effect, resulting in improved memory function.