Previously, a diagnosis of Alzheimer´s dementia could only be made with confidence when hallmark AD pathological agents were observed in post-mortem brain dissection. Over the past decades, methodological advancements have opened the gates for in-vivo diagnostics as neuropathological events can be detected through cerebrospinal fluid analyses and neuroimaging. The earliest pathological events take place decades before dementia, different pathologies are likely to start at different times, accumulate gradually and accompanied with subtle cognitive changes. Mild cognitive impairment (MCI) is considered a prodromal phase to dementia. In contrast to the more elusive phase of subjective cognitive decline (SCD), MCI is characterized by mildly impaired cognitive test results. One particular challenge is tied to the unreliability of MCI diagnostics. Many of those identified as MCI do not progress clinically and others revert back to normal neuropsychological functioning when followed longitudinally.
An objective cognitive deficit is needed for a MCI diagnosis. However, what a “mild cognitive impairment” specifically entails varies greatly. The traditional MCI criteria require one test score to lie one and a half standard deviation below average mean scores. But, recent investigations have found a propensity of the traditional neuropsychological criterion to make false-positive MCI diagnostic errors in as many as about 1/3 of MCI cases. When healthy participants take several neuropsychological tests, there is high probability of obtaining impaired scores by chance. By modifying the criteria with respect to what constitutes an objective cognitive deficit a dramatic reduction in making false-positive diagnostic errors have been reported.
In the first article conveyed in this thesis, we found that patients subtyped as amnestic MCI (aMCI) and multidomain MCI (mdMCI) displayed different levels of atrophy to substructures within the hippocampal complex. Only the perirhinal and entorhinal corticies were significantly thinner in the aMCI group, while additional significant volume reductions to subfields in the hippocampus proper were observed in the mdMCI group.
The second article investigated the extent to which sensitive neuropsychological tests of memory and executive functioning, predicted variation in metabolic and structural neuroimaging modalities. And, importantly, if single-tests are equally equipped to delineate associations with neurobiological properties as a composite test score. Results showed that while a composite test score was the strongest predictor of variation in cortical glucose PET metabolism, the single-tests were superior at predicting MRI cortical thickness. Neuropsychological testing was found to be sensitive, but variance in performance may differently relate to markers of neurobiological function and integrity. Establishing the extent to which neuropsychological performance converge with variation in hallmark neuroimaging underscores the utility of cognitive testing as a cost-effective and efficient clinical tool.
In the third article, we focused on how AD biomarker profiles differ in amnestic and non-amnestic MCI, and SCD. Results showed that amnestic MCI presented with the highest biomarker burden, while those identified as non-amnestic MCI and SCD displayed similar levels of pathological biomarkers. This was surprising as only the former had objective cognitive impairment. When biomarker data were analyzed continuously, results showed that the groups differed in types of biomarker pathology. For example, SCD displayed significantly thinner entorhinal cortex than healthy controls, which corroborate with subjective memory complaints. On the other hand, non-amnestic MCI did not show entorhinal cortex thinning. This study supports the clinical utility of neuropsychological profiling in the context of early AD and that SCD may represent a preclinical condition.