Identifying individuals at the earliest stages of Alzheimer’s Disease (AD) would enable development and study of interventions prior to onset of symptoms. However, differentiating age-related cognitive changes from subtle pathological changes remains a challenge in the field. Methods that would enable earlier detection of AD in elders with no subjective or objective cognitive concerns (i.e., individuals in the preclinical stage) would be of great interest. Community detection, a metric founded in graph theory, offers an alternative approach for characterizing subtle heterogeneity within aging samples and has the potential to inform cognitive variability in aging.

Using a hierarchical community detection, we examined whether cognitive subtypes could be identified in 226 cognitively normal older adults (from the Alzheimer’s Disease Neuroimaging Initiative [ADNI] study). Cognitive profiles of each community were characterized first using MANOVAs to examine the relationship between community membership and 12 age-, gender-, and education-corrected neuropsychological variables. Pair-wise comparisons were examined for significant main effects. We then examined whether these subtypes were related to biomarkers (cortical volumes, fluorodeoxyglucose (FDG)-positron emission tomography (PET) hypometabolism) or clinical progression. All p values were corrected for multiple comparisons.

Three communities (i.e., cognitive subtypes) were identified within the healthy aging sample. The first and largest community identified (N = 106) was characterized by a relative weakness on a single measure visuospatial executive function. Both the second (N = 76) and third community (N = 44) scored significantly lower on immediate, delayed, and recognition memory relative to the first community. The third community was characterized by a relative weakness in category fluency and speeded visual sequencing as well (p < .000). The three communities did not differ on age, gender, education, race, or ethnicity. Community membership was associated with entorhinal volume (with the second and third communities having significantly smaller entorhinal volumes than the first community), though community membership was not significantly associated with other biomarkers examined. Conversion rate reached trend level significance at 12 month follow up (more converters in the third community).

Hierarchical community detection is an alternative method for characterizing neuropsychological variation and it appears sensitive to relatively small differences that may be observed in a normal aging sample. While the sample size was relatively small, this approach shows promise for potentially leading to earlier detection of cognitive decline among individuals classified to be aging normally (e.g., community 3).

Most neuropsychological tests were developed without the benefit of modern psychometric theory. We used item response theory (IRT) methods to determine whether a widely used test – the 26-item Matrix Reasoning subtest of the WAIS-IV – might be used more efficiently if it were administered using computerized adaptive testing (CAT).

Data on the Matrix Reasoning subtest from 2197 participants enrolled in the National Neuropsychology Network (NNN) were analyzed using a two-parameter logistic (2PL) IRT model. Simulated CAT results were generated to examine optimal short forms using fixed-length CATs of 3, 6, and 12 items and scores were compared to the original full subtest score. CAT models further explored how many items were needed to achieve a selected precision of measurement (standard error ≤ .40).

The fixed-length CATs of 3, 6, and 12 items correlated well with full-length test results (with r = .90, .97 and .99, respectively). To achieve a standard error of .40 (approximate reliability = .84) only 3–7 items had to be administered for a large percentage of individuals.

This proof-of-concept investigation suggests that the widely used Matrix Reasoning subtest of the WAIS-IV might be shortened by more than 70% in most examinees while maintaining acceptable measurement precision. If similar savings could be realized in other tests, the accessibility of neuropsychological assessment might be markedly enhanced, and more efficient time use could lead to broader subdomain assessment.

Mild cognitive impairment (MCI) types may have distinct neuropathological substrates with hippocampal atrophy particularly common in amnestic MCI (aMCI). However, depending on the MCI classification criteria applied to the sample (e.g., number of abnormal test scores considered or thresholds for impairment), volumetric findings between MCI types may change. Additionally, despite increased clinical use, no prior research has examined volumetric differences in MCI types using the automated volumetric software, Neuroreader™.

The present study separately applied the Petersen/Winblad and Jak/Bondi MCI criteria to a clinical sample of older adults (N = 82) who underwent neuropsychological testing and brain MRI. Volumetric data were analyzed using Neuroreader™ and hippocampal volumes were compared between aMCI and non-amnestic MCI (naMCI).

T-tests revealed that regardless of MCI classification criteria, hippocampal volume z-scores were significantly lower in aMCI compared to naMCI (p’s < .05), and hippocampal volume z-scores significantly differed from 0 (Neuroreader™ normative mean) in the aMCI group only (p’s < .05). Additionally, significant, positive correlations were found between measures of delayed recall and hippocampal z-scores in aMCI using either MCI classification criteria (p’s < .05).

We provide evidence of correlated neuroanatomical changes associated with memory performance for two commonly used neuropsychological MCI classification criteria. Future research should investigate the clinical utility of hippocampal volumes analyzed via Neuroreader™ in MCI.

The National Neuropsychology Network (NNN) is a multicenter clinical research initiative funded by the National Institute of Mental Health (NIMH; R01 MH118514) to facilitate neuropsychology’s transition to contemporary psychometric assessment methods with resultant improvement in test validation and assessment efficiency.

The NNN includes four clinical research sites (Emory University; Medical College of Wisconsin; University of California, Los Angeles (UCLA); University of Florida) and Pearson Clinical Assessment. Pearson Q-interactive (Q-i) is used for data capture for Pearson published tests; web-based data capture tools programmed by UCLA, which serves as the Coordinating Center, are employed for remaining measures.

NNN is acquiring item-level data from 500–10,000 patients across 47 widely used Neuropsychology (NP) tests and sharing these data via the NIMH Data Archive. Modern psychometric methods (e.g., item response theory) will specify the constructs measured by different tests and determine their positive/negative predictive power regarding diagnostic outcomes and relationships to other clinical, historical, and demographic factors. The Structured History Protocol for NP (SHiP-NP) helps standardize acquisition of relevant history and self-report data.

NNN is a proof-of-principle collaboration: by addressing logistical challenges, NNN aims to engage other clinics to create a national and ultimately an international network. The mature NNN will provide mechanisms for data aggregation enabling shared analysis and collaborative research. NNN promises ultimately to enable robust diagnostic inferences about neuropsychological test patterns and to promote the validation of novel adaptive assessment strategies that will be more efficient, more precise, and more sensitive to clinical contexts and individual/cultural differences.

This study examined the relationship between patient performance on multiple memory measures and regional brain volumes using an FDA-cleared quantitative volumetric analysis program – Neuroreader™.

Ninety-two patients diagnosed with mild cognitive impairment (MCI) by a clinical neuropsychologist completed cognitive evaluations and underwent MR Neuroreader™ within 1 year of testing. Select brain regions were correlated with three widely used memory tests. Regression analyses were conducted to determine if using more than one memory measures would better predict hippocampal z-scores and to explore the added value of recognition memory to prediction models.

Memory performances were most strongly correlated with hippocampal volumes than other brain regions. After controlling for encoding/Immediate Recall standard scores, statistically significant correlations emerged between Delayed Recall and hippocampal volumes (rs ranging from .348 to .490). Regression analysis revealed that evaluating memory performance across multiple memory measures is a better predictor of hippocampal volume than individual memory performances. Recognition memory did not add further predictive utility to regression analyses.

This study provides support for use of MR Neuroreader™ hippocampal volumes as a clinically informative biomarker associated with memory performance, which is a critical diagnostic feature of MCI phenotype.