Background: Most of the current tests that evaluate episodic memory in geriatric population don’t always consider factors like native language, social-cultural origin, visual capacity or attentional availability of the patients. The “Self-initiated Items Memory” test (SIM), tries to answer these constraints by proposing to the patients to identify the items they will have to memorize. Objective: to test the benefits of the Self-initiated items procedure, we compared the performance of the original test to an alternative version with imposed items. Methodology: Sixteen older participants (77.21 ± 6.8 years ; MMSE 28.35 ± 1.5) without cognitive impairments (control participants) and seven older participants with a diagnosis of Alzheimer's disease (89.29 ± 4.4 years; MMSE 17.86 ± 2.4) were recruited for the study. Results: The number of recalled items was significantly higher with the self-initiated procedure than with the imposed-items procedure in both groups. Furthermore, in both tests, the performance of participants with diagnosed memory impairment remained lower than that of participants in the control group. The subjective ratings of “agreeableness” was also higher when the items were self-initiated, and this effect was similar in both groups. Conclusion: The Self-initiated Item Memory test remains discriminating. Individuals without diagnosis performed better than individuals diagnosed with memory impairment. The fact that the test is easier, fun, short and enjoyable doesn’t alter its ability to detect disorders. These results underline the relevance of our test especially in the clinic when other traditional tests cannot be used.