The diminished ability to perform everyday tasks is a salient problem for individuals diagnosed with a dementia. Recent research suggest that dementia patients detect significantly fewer action errors than age-matched controls; however, very little is known about the derivation of their error monitoring difficulties. The primary aims of my dissertation were to evaluate a novel, task training action intervention (TT-NAT) designed to increase error monitoring in dementia patients, and to pinpoint the relation between error monitoring and neuropsychological processes in participants who receive the task training intervention. Results indicated that dementia participants in the TT-NAT condition produced fewer total errors and detected significantly more of their errors than individuals in the Standard condition (z = 3.0 and t = 3.36, respectively; p < . 05). Error detection in the TT-NAT condition was strongly related to the language/semantic knowledge composite index only (r = .57, p = .00), whereas it was moderately related to both the language and executive composite indices in the Standard condition. No differences in error correction rates were noted, although patients in all groups corrected the majority of errors detected. The findings suggest that the TT-NAT may be a promising intervention for error monitoring deficits in dementia patients, and have considerable implications for neuropsychological rehabilitation. / Psychology