The current study tested the hypothesis that attention mediates the impact of psychological disturbances (i.e. depression, loss of self-control, and bizarre thinking) on immediate memory. A hypothesized model, in which psychological disturbances indirectly impacted immediate memory via attention, was tested using structural equations modeling (EQS). The participants included heterogenous groups of brain-damaged and psychiatric subjects both with and without age-corrected observed variables. These groups did not vary with regard to the relationships within the model, and all models had good fit indices [comparative fit index (CFI)>0.98]. Competing models with paths directly from psychological disturbances to verbal, spatial, and general immediate memory were not significant, and these competing models did not result in a better fitting model. Therefore, the best fitting, most parsimonious model was the hypothesized one. These results corroborate the hypothesis of an indirect relationship between psychological disturbances and immediate memory via attention.