We investigated the earliest neuropsychological changes in Alzheimer's disease (AD) by comparing the baseline performance of 29 individuals who subsequently developed AD within an average of 7.91 ± 2.70 years with 29 pairwise-matched individuals who remained cognitively healthy (NC). We hypothesized that subtle, qualitative changes in cognition precede clinical AD by several years, and therefore examined subjective as well as standard quantitative measures of cognition, in addition to subjective estimates of mood and medical status. Participants were selected from the 825 members of the longitudinal BASEL study (BAsel Study on the ELderly), all of whom had been ApoE-genotyped and received comprehensive bi-annual neuropsychological assessments. Within 13 years, 29 were diagnosed with probable AD. Each individual who progressed to AD (AD-P) was pairwise matched to a NC participant based on age, education, demographic status, observation period, and, importantly, ApoE genotype. A regression analysis using the lasso technique identified which of 115 neuropsychological variables best discriminated baseline NC from baseline AD-P performance. This analysis yielded eleven neuropsychological variables that optimally discriminated the two groups (correct classification rate: 60.4%): 1) Intrusions and 2) response bias in verbal learning and memory tasks; 3) delayed figure recall; 4–6) three Wechsler Adult Intelligence Scale (WAIS) Block Design subtest variables; 7–8) number of errors and repetitions on letter fluency; and 9–11) self-report of memory problems, a feeling of sadness, and cardiac problems. These results suggest that the preclinical neuropsychological cascade to AD includes subtle but identifiable qualitative impairments in verbal and visual memory, visuospatial processing, error control, and subjective neuropsychological complaints.