Cognitive measures that distinguish frontotemporal Dementia (FTD) from Alzheimer's Disease (AD) have shown poor discriminability. Clinical diagnosis is based on behavioral features since cognitive studies show inconsistent differences. To compare frontal lobe processes deficits exhibited by patients with DFT, AD as well as subjects w/o dementia with similar cognitive disabilities as bipolar disorders (BD). Exploration of functions of frontal lobes of patients with pathology involving such area was accomplished. FTD (n=20), AD (n=20), and BD (n=20) were diagnosed by consensus (Geriatric, Neurology, Neuropsychology and Psychiatry) at the San Ignacio Hospital Memory Clinic as well as two Psychiatry Clinics. Six traditional frontal lobe tests were applied (London Tower, WCST, Stroop, TMT, A&B and FAS). Analysis of results was done through its cognitive components: abstraction, set failure, conceptualization, search, planning, inhibition, perseveration. K–means cluster analysis controlling demographic variables, level and disease progression, revealed that splitting components discriminated FTD from AD and BD. For instance, planning and percentage of abstraction (number of trails to achieve a category in the WCST against number of trails in London Tower) discriminate FTD groups from the rest of the sample, becoming an important cognitive marker. While phonological verbal search distinguishes FTD, semantic search scores are significantly dispersed in all subjects. Stroop errors, (inhibition) cluster FTD patients as well as BD subjects, while time performance does not, becoming an unique marker for BD subjects. Finally, failure to maintain the set (WCST) and number of categories achieved did also classify FTD patients. Analysis of cognitive components of frontal neuropsychological tests facilitates a better discrimination of AD and FTD. In the present study, abstraction, phonological search, planning and ability to inhibit irrelevant responses could be early cognitive markers of FTD patients. BD subjects were included assuming they have the same cognitive dysfunction (frontal lobe) in contrast with the results obtained. Cognitive components in BD could be prominent when the sample includes only BDs. Further studies with this modality of analysis may exhibit a better discrimination of these cortical dementias, allowing possible heterogeneous cognitive profiles of FTD identification and definition of early cognitive markers.