Objectives: Prior research has identified numerous genetic (including sex), education, health, and lifestyle factors that predict cognitive decline. Traditional model selection approaches (e.g., backward or stepwise selection) attempt to find one model that best fits the observed data, risking interpretations that only the selected predictors are important. In reality, several predictor combinations may fit similarly well but result in different conclusions (e.g., about size and significance of parameter estimates). In this study, we describe an alternative method, Information-Theoretic (IT) model averaging, and apply it to characterize a set of complex interactions in a longitudinal study on cognitive decline. Methods: Here, we used longitudinal cognitive data from 1256 late–middle aged adults from the Wisconsin Registry for Alzheimer’s Prevention study to examine the effects of sex, apolipoprotein E (APOE) ɛ4 allele (non-modifiable factors), and literacy achievement (modifiable) on cognitive decline. For each outcome, we applied IT model averaging to a set of models with different combinations of interactions among sex, APOE, literacy, and age. Results: For a list-learning test, model-averaged results showed better performance for women versus men, with faster decline among men; increased literacy was associated with better performance, particularly among men. APOE had less of an association with cognitive performance in this age range (∼40–70 years). Conclusions: These results illustrate the utility of the IT approach and point to literacy as a potential modifier of cognitive decline. Whether the protective effect of literacy is due to educational attainment or intrinsic verbal intellectual ability is the topic of ongoing work. (JINS, 2019, 25, 119–133)

Objectives: A major challenge in cognitive aging is differentiating preclinical disease-related cognitive decline from changes associated with normal aging. Neuropsychological test authors typically publish single time-point norms, referred to here as unconditional reference values. However, detecting significant change requires longitudinal, or conditional reference values, created by modeling cognition as a function of prior performance. Our objectives were to create, depict, and examine preliminary validity of unconditional and conditional reference values for ages 40–75 years on neuropsychological tests. Method: We used quantile regression to create growth-curve–like models of performance on tests of memory and executive function using participants from the Wisconsin Registry for Alzheimer’s Prevention. Unconditional and conditional models accounted for age, sex, education, and verbal ability/literacy; conditional models also included past performance on and number of prior exposures to the test. Models were then used to estimate individuals’ unconditional and conditional percentile ranks for each test. We examined how low performance on each test (operationalized as <7th percentile) related to consensus-conference–determined cognitive statuses and subjective impairment. Results: Participants with low performance were more likely to receive an abnormal cognitive diagnosis at the current visit (but not later visits). Low performance was also linked to subjective and informant reports of worsening memory function. Conclusions: The percentile-based methods and single-test results described here show potential for detecting troublesome within-person cognitive change. Development of reference values for additional cognitive measures, investigation of alternative thresholds for abnormality (including multi-test criteria), and validation in samples with more clinical endpoints are needed. (JINS, 2019, 25, 1–14)

Objectives: The purpose of this study was to investigate the longitudinal trajectory of self- and informant-subjective cognitive complaints (SCC), and to determine if SCC predict longitudinal changes in objective measures (OM) of cognitive function. Methods: The study included healthy and cognitively normal late middle-aged adults enriched with a family history of AD who were evaluated at up to three visits over a 4-year period. At each visit (Visit 1–3), self- and informant-SCC and OM were evaluated. Linear mixed models were used to determine if the longitudinal rate of change of self- and informant-SCC were associated with demographic variables, depressive symptoms, family history (FH), and apolipoprotein epsilon 4 (APOE4) status. The same modeling approach was used to examine the effect of Visit 1 SCC on longitudinal cognitive change after controlling for the same variables. Results: At Visit 1, more self-SCC were associated with fewer years of education and more depressive symptoms. SCC were also associated with poorer performance on cognitive measures, such that more self-SCC at Visit 1 were associated with poorer performance on memory and executive functioning measures at Visit 1, while more informant-SCC were associated with faster rate of longitudinal decline on a measure of episodic learning and memory. FH and APOE4 status were not associated with SCC. Discussion: Self- and informant-SCC showed an association with OM, albeit over different time frames in our late middle-aged sample. Additional longitudinal follow-up will likely assist in further clarifying these relationships as our sample ages and more pronounced cognitive changes eventually emerge. (JINS, 2017, 23, 617–626)

Objectives: Intraindividual cognitive variability (IICV) has been shown to differentiate between groups with normal cognition, mild cognitive impairment (MCI), and dementia. This study examined whether baseline IICV predicted subsequent mild to moderate cognitive impairment in a cognitively normal baseline sample. Methods: Participants with 4 waves of cognitive assessment were drawn from the Wisconsin Registry for Alzheimer’s Prevention (WRAP; n=684; 53.6(6.6) baseline age; 9.1(1.0) years follow-up; 70% female; 74.6% parental history of Alzheimer’s disease). The primary outcome was Wave 4 cognitive status (“cognitively normal” vs. “impaired”) determined by consensus conference; “impaired” included early MCI (n=109), clinical MCI (n=11), or dementia (n=1). Primary predictors included two IICV variables, each based on the standard deviation of a set of scores: “6 Factor IICV” and “4 Test IICV”. Each IICV variable was tested in a series of logistic regression models to determine whether IICV predicted cognitive status. In exploratory analyses, distribution-based cutoffs incorporating memory, executive function, and IICV patterns were used to create and test an MCI risk variable. Results: Results were similar for the IICV variables: higher IICV was associated with greater risk of subsequent impairment after covariate adjustment. After adjusting for memory and executive functioning scores contributing to IICV, IICV was not significant. The MCI risk variable also predicted risk of impairment. Conclusions: While IICV in middle-age predicts subsequent impairment, it is a weaker risk indicator than the memory and executive function scores contributing to its calculation. Exploratory analyses suggest potential to incorporate IICV patterns into risk assessment in clinical settings. (JINS, 2016, 22, 1016–1025)

Objectives: The purpose of this study was to assess whether age-related differences in white matter microstructure are associated with altered task-related connectivity during episodic recognition. Methods: Using functional magnetic resonance imaging and diffusion tensor imaging from 282 cognitively healthy middle-to-late aged adults enrolled in the Wisconsin Registry for Alzheimer’s Prevention, we investigated whether fractional anisotropy (FA) within white matter regions known to decline with age was associated with task-related connectivity within the recognition network. Results: There was a positive relationship between fornix FA and memory performance, both of which negatively correlated with age. Psychophysiological interaction analyses revealed that higher fornix FA was associated with increased task-related connectivity amongst the hippocampus, caudate, precuneus, middle occipital gyrus, and middle frontal gyrus. In addition, better task performance was associated with increased task-related connectivity between the posterior cingulate gyrus, middle frontal gyrus, cuneus, and hippocampus. Conclusions: The findings indicate that age has a negative effect on white matter microstructure, which in turn has a negative impact on memory performance. However, fornix microstructure did not significantly mediate the effect of age on performance. Of interest, dynamic functional connectivity was associated with better memory performance. The results of the psychophysiological interaction analysis further revealed that alterations in fornix microstructure explain–at least in part–connectivity among cortical regions in the recognition memory network. Our results may further elucidate the relationship between structural connectivity, neural function, and cognition. (JINS, 2016, 22, 191–204)

The aim of this study was to examine cross-sectionally whether higher cardiorespiratory fitness (CRF) might favorably modify amyloid-β (Aβ)-related decrements in cognition in a cohort of late-middle-aged adults at risk for Alzheimer’s disease (AD). Sixty-nine enrollees in the Wisconsin Registry for Alzheimer’s Prevention participated in this study. They completed a comprehensive neuropsychological exam, underwent 11C Pittsburgh Compound B (PiB)-PET imaging, and performed a graded treadmill exercise test to volitional exhaustion. Peak oxygen consumption (VO2peak) during the exercise test was used as the index of CRF. Forty-five participants also underwent lumbar puncture for collection of cerebrospinal fluid (CSF) samples, from which Aβ42 was immunoassayed. Covariate-adjusted regression analyses were used to test whether the association between Aβ and cognition was modified by CRF. There were significant VO2peak*PiB-PET interactions for Immediate Memory (p=.041) and Verbal Learning & Memory (p=.025). There were also significant VO2peak*CSF Aβ42 interactions for Immediate Memory (p<.001) and Verbal Learning & Memory (p<.001). Specifically, in the context of high Aβ burden, that is, increased PiB-PET binding or reduced CSF Aβ42, individuals with higher CRF exhibited significantly better cognition compared with individuals with lower CRF. In a late-middle-aged, at-risk cohort, higher CRF is associated with a diminution of Aβ-related effects on cognition. These findings suggest that exercise might play an important role in the prevention of AD. (JINS, 2015, 21, 841–850)

The relative influence of amyloid burden, neuronal structure and function, and prior cognitive performance on prospective memory decline among asymptomatic late middle-aged individuals at risk for Alzheimer's disease (AD) is currently unknown. We investigated this using longitudinal cognitive data from 122 middle-aged adults (21 “Decliners” and 101 “Stables”) enrolled in the Wisconsin Registry for Alzheimer's Prevention who underwent multimodality neuroimaging [11C-Pittsburgh Compound B (PiB), 18F-fluorodeoxyglucose (FDG), and structural/functional magnetic resonance imaging (fMRI)] 5.7 ± 1.4 years (range = 2.9–8.9) after their baseline cognitive assessment. Covariate-adjusted regression analyses revealed that the only imaging measure that significantly distinguished Decliners from Stables (p = .027) was a Neuronal Function composite derived from FDG and fMRI. In contrast, several cognitive measures, especially those that tap episodic memory, significantly distinguished the groups (p's<.05). Complementary receiver operating characteristic curve analyses identified the Brief Visuospatial Memory Test-Revised (BVMT-R) Total (.82 ± .05, p < .001), the BVMT-R Delayed Recall (.73 ± .06, p = .001), and the Reading subtest from the Wide-Range Achievement Test-III (.72 ± .06, p = .002) as the top three measures that best discriminated the groups. These findings suggest that early memory test performance might serve a more clinically pivotal role in forecasting future cognitive course than is currently presumed. (JINS, 2014, 20, 1–12)

After traumatic injury, the brain undergoes a prolonged period of degenerative change that is paradoxically accompanied by cognitive recovery. The spatiotemporal pattern of atrophy and the specific relationships of atrophy to cognitive changes are ill understood. The present study used tensor-based morphometry and neuropsychological testing to examine brain volume loss in 17 traumatic brain injury (TBI) patients and 13 controls over a 4-year period. Patients were scanned at 2 months, 1 year, and 4 years post-injury. High-dimensional warping procedures were used to create change maps of each subject's brain for each of the two intervals. TBI patients experienced volume loss in both cortical areas and white matter regions during the first interval. We also observed continuing volume loss in extensive regions of white matter during the second interval. Neuropsychological correlations indicated that cognitive tasks were associated with subsequent volume loss in task-relevant regions. The extensive volume loss in brain white matter observed well beyond the first year post-injury suggests that the injured brain remains malleable for an extended period, and the neuropsychological relationships suggest that this volume loss may be associated with subtle cognitive improvements. (JINS, 2012, 18, 1–13)

Significant anoxia may cause a variety of neuropathologic changes as well as cognitive deficits. We have recently seen 3 patients who have suffered severe anoxic episodes all with initial Glasgow Coma Scores (GCS) of 3 with sustained coma for 10–14 d. All 3 patients had extended hospitalizations and rehabilitation therapy. A neuropsychological test battery was administered and volumetric analyses of MRI scans were carried out in each case at least 6 mo postinjury. Two of the patients display distinct residual cognitive and neuropathologic changes while 1 patient made a remarkable recovery without evidence of significant morphological abnormality. These three cases demonstrate, that even with similar admission GCS, the outcome is variable and the degree of neuropsychological impairment appears to match the degree of morphologic abnormalities demonstrated by quantitative MR image analysis. An important finding of this study is that even though subjects with an anoxic insult exhibit severe cognitive and memory impairments along with concomitant morphologic changes, their attention/concentration abilities appear to be preserved. MR morphometry provides an excellent means by which neural structural changes can be quantified and compared to neuropsychological and behavioral outcomes. (JINS, 1995, I, 501–509.)

Morphometric analysis of magnetic resonance (MR) scans in 88 traumatic brain injury (TBI) patients demonstrated significantly larger ventricle-to-brain ratios (VBR) and temporal horn volumes, and significantly smaller fornix-to-brain ratios (FBR) and corpus callosum (CC) area measurements, compared to 73 controls. Additionally, TBI patients were grouped according to Glasgow Coma Scale (GCS) for a within-TBI sample comparison so that severity of injury on brain morphology could be examined. The severe TBI group (GCS = 3–6) differed from the mild and moderate injury groups on measures of the internal capsule, VBR, temporal horn volume, and CC. In a separate analysis wherein the TBI subjects were grouped by degree of fornix atrophy, the group with the smallest fornix size demonstrated the lowest memory performance. Furthermore, anatomic measures correlated with severity of injury, and tests of memory and motor function. Results demonstrate the diffuse nature of degeneration in TBI with more severe injury, and that quantified MR identified morphologic changes relate to neuropsychological outcome. (JINS, 1995, 1, 17–28.)

Awareness of cognitive dysfunction shown by individuals with Mild Cognitive Impairment (MCI), a condition conferring risk for Alzheimer's disease (AD), is variable. Anosognosia, or unawareness of loss of function, is beginning to be recognized as an important clinical symptom of MCI. However, little is known about the brain substrates underlying this symptom. We hypothesized that MCI participants' activation of cortical midline structures (CMS) during self-appraisal would covary with level of insight into cognitive difficulties (indexed by a discrepancy score between patient and informant ratings of cognitive decline in each MCI participant). To address this hypothesis, we first compared 16 MCI participants and 16 age-matched controls, examining brain regions showing conjoint or differential BOLD response during self-appraisal. Second, we used regression to investigate the relationship between awareness of deficit in MCI and BOLD activity during self-appraisal, controlling for extent of memory impairment. Between-group comparisons indicated that MCI participants show subtly attenuated CMS activity during self-appraisal. Regression analysis revealed a highly significant relationship between BOLD response during self-appraisal and self-awareness of deficit in MCI. This finding highlights the level of anosognosia in MCI as an important predictor of response to self-appraisal in cortical midline structures, brain regions vulnerable to changes in early AD. (JINS, 2007, 13, 450–461.)

The Hooper Visual Organization Test (VOT), a commonly applied neuropsychological test of visual spatial ability, is used for assessing patients with suspected right hemisphere, or parietal lobe involvement. A controversy has developed over whether the inferences of this test metric can be assumed to involve global, lateralized, or regional functionality. In this study, the characteristic visual organization and object naming aspects of the VOT task presentation were adapted to a functional MR imaging (fMRI) paradigm to probe the neuroanatomic correlates of this neuropsychological test. Whole brain fMRI mapping results are reported on a cohort of normal subjects. Bilateral fMRI responses were found predominantly in the posterior brain, in regions of superior parietal lobules, ventral temporal-occipital cortex, and posterior visual association areas, and to a lesser extent, the frontal eye fields bilaterally, and left dorsolateral prefrontal cortex. The results indicate a general brain region or network in which VOT impairment, due to its visuospatial and object identification demands, is possible to be detected. Discussion is made of interpretive limitations when adapting neuropsychological tests to fMRI analysis. (JINS, 2004, 10, 939–947.)

We have recently reported (Saykin et al., 1999b) selective activation of left medial temporal lobe structures during processing of novel compared to familiar words using functional magnetic resonance imaging (fMRI). The current study describes the relationship between a widely used clinical test of verbal learning, the California Verbal Learning Test (CVLT), and the previously reported fMRI activations. Thirteen right-handed healthy adult participants were studied with whole brain echo-planar fMRI while listening to novel and recently learned (familiar) words intermixed pseudorandomly in an event-related design. These participants were also tested with the CVLT. Scores for CVLT Trial 1 (immediate encoding of novel words) and recognition discriminability (recognition of familiar vs. novel words) were correlated with fMRI signal change during processing of novel versus familiar words using a covariance model implemented in SPM96. For the novel words analysis, voxels in the right anterior hippocampus correlated significantly with Trial 1 (r = .76 at the maxima). For the recognition analysis, a significant cluster of voxels was found in the right dorsolateral prefrontal cortex (r = .88 at the maxima). Our prior results of separable left medial temporal activation to novel and familiar words, together with results of the covariance analyses reported here, suggest that in addition to the left medial temporal lobe (MTL) regions that are engaged during novel and familiar word processing, the right hippocampus and right frontal lobe are also involved, particularly in those participants with better memory ability. This positive relationship between fMRI activation and CVLT performance suggests a role for these right hemisphere regions in successful memory processing of verbal material, perhaps reflecting more efficient encoding and retrieval strategies that facilitate memory. (JINS, 2001, 7, 55–62.)

Few studies have examined the consequences of alcohol and drug abuse on TBI though they commonly co-occur. Both TBI and substance abuse independently result in neuropathological changes in the brain such as ventricular enlargement and cortical atrophy, thus it is reasonable to hypothesize that the combination of the two would result in more significant cerebral damage. In this study, 3 groups of patients—traumatically brain injured (TBI) with substance abuse (N = 19), TBI without substance abuse (N = 19), and substance abuse with no TBI (N = 16)—were compared with normal controls (N = 20) on several quantitative MRI (QMRI) measures. Since TBI most frequently occurs in older adolescents and young men, we examined only male participants between 16 and 30 years of age. Comparing young substance abusers to controls resulted in no QMRI differences. When controlling for head injury severity, the effects of substance abuse in combination with TBI resulted in greater atrophic changes than seen in any other group. TBI and substance abuse patients' neuropsychological test performances also were examined, and no differences were found among patient groups on any measures. These findings have implications for the deleterious interaction of substance abuse combining with TBI to result in greater neuropathological changes that can be detected by QMRI techniques. (JINS, 1999, 5, 593–608.)

Impairment in semantic processing occurs early in Alzheimer's disease (AD) and differential impact on subtypes of semantic relations have been reported, yet there is little data on the neuroanatomic basis of these deficits. Patients with mild AD and healthy controls underwent 3 functional MRI auditory stimulation tasks requiring semantic or phonological decisions (match–mismatch) about word pairs (category–exemplar, category–function, pseudoword). Patients showed a significant performance deficit only on the exemplar task. On voxel-based fMRI activation analyses, controls showed a clear activation focus in the left superior temporal gyrus for the phonological task; patients showed additional foci in the left dorsolateral prefrontal and bilateral cingulate areas. On the semantic tasks, predominant activation foci were seen in the inferior and middle frontal gyrus (left greater than right) in both groups but patients showed additional activation suggesting compensatory recruitment of locally expanded foci and remote regions, for example, right frontal activation during the exemplar task. Covariance analyses indicated that exemplar task performance was strongly related to signal increase in bilateral medial prefrontal cortex. The authors conclude that fMRI can reveal similarities and differences in functional neuroanatomical processing of semantic and phonological information in mild AD compared to healthy elderly, and can help to bridge cognitive and neural investigations of the integrity of semantic networks in AD. (JINS, 1999, 5, 377–392.)