Functional impairment is an important aspect for early diagnosis of dementia, but it might be compromised by emotional factors of the informant. Objective: We aimed to compare the direct functional assessment with the indirect evaluation and to investigate whether caregiver/informant burden, anxiety and depression might interfere in this process. Method: This is a crosssectional study of activities of daily living of 40 Alzheimer´s disease (AD) patients compared to a matched 41 elderly controls. We assessed functionality by both, direct evaluations by a health professional and indirect instruments that were answered by the informant. We investigated the presence of depressive, anxiety and burden symptoms in the informant. Statistical analysis was performed based on single-factor analysis of variance (ANOVA) with multiple comparisons (Bonferroni test). For correlational analysis we used Pearson's correlation coefficient. Results: we observed between-groups significant differences in relation to the caregiver's burden (p < 0.001), however no other finding was observed related to depression, anxiety and cognition. In the analysis of correlations between emotional aspects and functionality evaluations we observed that in both controls and mild AD groups, IQCODE was the indirect assessment that most closely approximated to the direct evaluation and was less influenced by emotional symptoms. Conclusion: The presence of emotional symptoms influences the indirect evaluation of mild conditions of cognitive decline, in which case direct evaluation is preferable. In moderate cases, indirect evaluation is less influenced by caregiver burden because the impairment is more evident.