Factors considered

We are interested in the effect of the following factors on the likelihood of MCI or D individuals moving to a less severe state in subsequent assessments:

Chronological age: exact age of individuals at the point of assessment.

Gender: male or female.

Education: we considered education as a factor that could increase resistance to cognitive decline. We focused on two different groups: one consisting of individuals that have completed 12 years of education and one of individuals that have more than 12 years of education.

Genetic background: absence or presence of one or two alleles of the apolipoprotein ɛ4 genotype (APOE ɛ4, homozygous or heterozygous), a well-known risk factor of AD.

Time between visits: time between clinical assessments (in years).

Cognitive performance: We are interested in a number of commonly used measures of cognition. We focused on the scores in an Alzheimer’s Disease Assessment Scale-Cognitive Subscale test (ADAS-Cog 11) and the scores in the MMSE test. The original ADAS-Cog 11 [27] includes 11 items assessing cognitive function (including memory, language, praxis, and orientation). Higher scores reflect greater cognitive impairment and the maximum score of the test is 70 points. Additional analysis including ADAS-Cog 13-item scale [28], which consists of all items of the original ADAS-Cog 11 test plus two additional tasks, yielded similar results and has not been considered further in this study. MMSE [29] is a 30-point questionnaire which also measures cognitive impairment, with a lower score indicating greater severity. ADAS-Cog 11 and ADAS-Cog 13 scores are available in ADNI. MMSE scores are available in both ADNI and NACC. It should be noted that MMSE score is one of the criteria used for the clinical diagnosis of participants in ADNI, and also one of the most widely used test of dementia severity in NACC [30].

Depression scores (Geriatric Depression Scale, GDS): These are scores from a screening test designed to identify depression symptoms in the elderly. A 15-item GDS bigger or equal to 6 was an exclusion criterion in ADNI. Individuals with history of major depression or bipolar disorder within the past year were also excluded from the study. In NACC, GDS is not recorded for individuals with severe dementia, but it is not used as an exclusion criterion for study entry. It should be noted, however, that GDS is not a required form in NACC, so it may not be recorded for any participant at any visit. Due to the inclusion/exclusion criteria and the reduced number of assessments of the GDS score in the datasets considered, the results and conclusions about the potential effect of depression should be interpreted with extra caution.