Empirical assessment of the assumptions of ComBat with diffusion tensor imaging.

Purpose: Diffusion tensor imaging (DTI) is a magnetic resonance imaging technique that provides unique information about white matter microstructure in the brain but is susceptible to confounding effects introduced by scanner or acquisition differences. ComBat is a leading approach for addressing these site biases. However, despite its frequent use for harmonization, ComBat's robustness toward site dissimilarities and overall cohort size have not yet been evaluated in terms of DTI. Approach: As a baseline, we match N=358 participants from two sites to create a "silver standard" that simulates a cohort for multi-site harmonization. Across sites, we harmonize mean fractional anisotropy and mean diffusivity, calculated using participant DTI data, for the regions of interest defined by the JHU EVE-Type III atlas. We bootstrap 10 iterations at 19 levels of total sample size, 10 levels of sample size imbalance between sites, and 6 levels of mean age difference between sites to quantify (i) ßAGE, the linear regression coefficient of the relationship between FA and age; (ii) Î³/f*, the ComBat-estimated site-shift; and (iii) Î´/f*, the ComBat-estimated site-scaling. We characterize the reliability of ComBat by evaluating the root mean squared error in these three metrics and examine if there is a correlation between the reliability of ComBat and a violation of assumptions. Results: ComBat remains well behaved for ßAGE when N>162 and when the mean age difference is less than 4 years. The assumptions of the ComBat model regarding the normality of residual distributions are not violated as the model becomes unstable. Conclusion: Prior to harmonization of DTI data with ComBat, the input cohort should be examined for size and covariate distributions of each site. Direct assessment of residual distributions is less informative on stability than bootstrap analysis. We caution use ComBat of in situations that do not conform to the above thresholds.

Energetic Cost of Walking and Brain Atrophy in Mid-to-Late Life.

BACKGROUND: Higher energetic costs for mobility are associated with declining gait speed, and slow gait is linked to cognitive decline and Alzheimer's disease. However, the physiological underpinnings of gait and brain health have not been well explored. We examined the associations of the energetic cost of walking with brain volume in cognitively unimpaired adults from the Baltimore Longitudinal Study of Aging. METHODS: We used brain magnetic resonance imaging (MRI) data from 850 participants (mean baseline age 66.3 ± 14.5 years), of whom 451 had longitudinal MRI data (2.8 ± 1.0 MRI scans over 4.0 ± 2.0 years). The energetic cost of walking was assessed as the average energy expended (V̇O2) during 2.5 minutes of customary-paced overground walking. Multivariable linear mixed-effects models examined the associations between baseline energetic cost of walking and regional brain volumes adjusting for covariates. RESULTS: At baseline, higher energetic cost of walking was cross-sectionally associated with lower gray and white matter volumes within the frontal, parietal, and temporal lobes, as well as hippocampal, total brain, and larger ventricular volumes (all false-discovery rate [FDR] p < .05). A baseline energetic cost of walking × time interaction demonstrated that participants with higher energetic cost of walking had faster annual decline in hippocampal volume (FDR p = .02) and accelerated annual increase in ventricular volumes (FDR p = .02). CONCLUSIONS: The energetic cost of walking is associated with gray and white matter volumes and subsequent hippocampal atrophy and ventricular enlargement. Collectively, these data suggest the energetic cost of walking may be an early marker of neurodegeneration that contributes to the gait brain connection.

Disease Burden Affects Aging Brain Function.

BACKGROUND: Most older adults live with multiple chronic disease conditions, yet the effect of multiple diseases on brain function remains unclear. METHODS: We examine the relationship between disease multimorbidity and brain activity using regional cerebral blood flow (rCBF) 15O-water PET scans from 97 cognitively normal participants (mean baseline age 76.5) in the Baltimore Longitudinal Study of Aging (BLSA). Multimorbidity index scores, generated from the presence of 13 health conditions, were correlated with PET data at baseline and in longitudinal change (n = 74) over 5.05 (2.74 SD) years. RESULTS: At baseline, voxel-based analysis showed that higher multimorbidity scores were associated with lower relative activity in orbitofrontal, superior frontal, temporal pole and parahippocampal regions, and greater activity in lateral temporal, occipital, and cerebellar regions. Examination of the individual health conditions comprising the index score showed hypertension and chronic kidney disease individually contributed to the overall multimorbidity pattern of altered activity. Longitudinally, both increases and decreases in activity were seen in relation to increasing multimorbidity over time. These associations were identified in orbitofrontal, lateral temporal, brainstem, and cerebellar areas. CONCLUSION: Together, these results show that greater multimorbidity is associated with widespread areas of altered brain activity, supporting a link between health and changes in aging brain function.

Assessment of APOE in atypical parkinsonism syndromes.

Atypical parkinsonism syndromes are a heterogeneous group of neurodegenerative disorders that include corticobasal degeneration (CBD), Lewy body dementia (LBD), multiple system atrophy (MSA), and progressive supranuclear palsy (PSP). The APOE ε4 allele is a well-established risk factor for Alzheimer's disease; however, the role of APOE in atypical parkinsonism syndromes remains controversial. To examine the associations of APOE ε4 and ε2 alleles with risk of developing these syndromes, a total of 991 pathologically-confirmed atypical parkinsonism cases were genotyped using the Illumina NeuroChip array. We also performed genotyping and logistic regression analyses to examine APOE frequency and associated risk in patients with Alzheimer's disease (n = 571) and Parkinson's disease (n = 348). APOE genotypes were compared to those from neurologically healthy controls (n = 591). We demonstrate that APOE ε4 and ε2 carriers have a significantly increased and decreased risk, respectively, of developing Alzheimer's disease (ε4: OR: 4.13, 95% CI: 3.23-5.26, p = 3.67 × 10-30; ε2: OR: 0.21, 95% CI: 0.13-0.34; p = 5.39 × 10-10) and LBD (ε4: OR: 2.94, 95% CI: 2.34-3.71, p = 6.60 × 10-20; ε2: OR = OR: 0.39, 95% CI: 0.26-0.59; p = 6.88 × 10-6). No significant associations with risk for CBD, MSA, or PSP were observed. We also show that APOE ε4 decreases survival in a dose-dependent manner in Alzheimer's disease and LBD. Taken together, this study does not provide evidence to implicate a role of APOE in the neuropathogenesis of CBD, MSA, or PSP. However, we confirm association of the APOE ε4 allele with increased risk for LBD, and importantly demonstrate that APOE ε2 reduces risk of this disease.

Midlife and Late-Life Cardiorespiratory Fitness and Brain Volume Changes in Late Adulthood: Results From the Baltimore Longitudinal Study of Aging.

BACKGROUND: Higher cardiorespiratory fitness (CRF) is cross-sectionally associated with more conserved brain volume in older age, but longitudinal studies are rare. This study examined whether higher midlife CRF was prospectively associated with slower atrophy, which in turn was associated with higher late-life CRF. METHODS: Brain volume by magnetic resonance imaging was determined annually from 1994 to 2003 in 146 participants (M baseline age = 69.6 years). Peak oxygen uptake on a treadmill yielded estimated midlife CRF in 138 and late-life CRF in 73 participants. RESULTS: Higher midlife CRF was associated with greater middle temporal gyrus, perirhinal cortex, and temporal and parietal white matter, but was not associated with atrophy progression. Slower atrophy in middle frontal and angular gyri was associated with higher late-life CRF, independent of CRF at baseline magnetic resonance imaging. CONCLUSIONS: Higher midlife CRF may play a role in preserving middle and medial temporal volumes in late adulthood. Slower atrophy in middle frontal and angular gyri may predict late-life CRF.

Individual estimates of age at detectable amyloid onset for risk factor assessment.

INTRODUCTION: Individualized estimates of age at detectable amyloid-beta (Aß) accumulation, distinct from amyloid positivity, allow for analysis of onset age of Aß accumulation as an outcome measure to understand risk factors. METHODS: Using longitudinal Pittsburgh compound B (PiB) positron emission tomography data from Baltimore Longitudinal Study of Aging, we estimated the age at which each PiB+ individual began accumulating Aß. We used survival analysis methods to quantify risk of accumulating Aß and differences in onset age of Aß accumulation in relation to APOE ε4 status and sex among 36 APOE ε4 carriers and 83 noncarriers. RESULTS: Age at onset of Aß accumulation for the APOE ε4- and ε4+ groups was 73.1 and 60.7, respectively. APOE ε4 positivity conferred a threefold risk of accumulating Aß after adjusting for sex and education. DISCUSSION: Estimation of onset age of amyloid accumulation may help gauge treatment efficacy in interventions to delay symptom onset in Alzheimer's disease.

Analysis of spatio-temporal brain imaging patterns by Hidden Markov Models and serial MRI images.

Brain changes due to development and maturation, normal aging, or degenerative disease are continuous, gradual, and variable across individuals. To quantify the individual progression of brain changes, we propose a spatio-temporal methodology based on Hidden Markov Models (HMM), and apply it on four-dimensional structural brain magnetic resonance imaging series of older individuals. First, regional brain features are extracted in order to reduce image dimensionality. This process is guided by the objective of the study or the specific imaging patterns whose progression is of interest, for example, the evaluation of Alzheimer-like patterns of brain change in normal individuals. These regional features are used in conjunction with HMMs, which aim to measure the dynamic association between brain structure changes and progressive stages of disease over time. A bagging framework is used to obtain models with good generalization capability, since in practice the number of serial scans is limited. An application of the proposed methodology was to detect individuals with the risk of developing MCI, and therefore it was tested on modeling the progression of brain atrophy patterns in older adults. With HMM models, the state-transition paths corresponding to longitudinal brain changes were constructed from two completely independent datasets, the Alzheimer Disease Neuroimaging Initiative and the Baltimore Longitudinal Study of Aging. The statistical analysis of HMM-state paths among the normal, progressive MCI, and MCI groups indicates that, HMM-state index 1 is likely to be a predictor of the conversion from cognitively normal to MCI, potentially many years before clinical symptoms become measurable.

Alzheimer's disease risk assessment using large-scale machine learning methods.

The goal of this work is to introduce new metrics to assess risk of Alzheimer's disease (AD) which we call AD Pattern Similarity (AD-PS) scores. These metrics are the conditional probabilities modeled by large-scale regularized logistic regression. The AD-PS scores derived from structural MRI and cognitive test data were tested across different situations using data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) study. The scores were computed across groups of participants stratified by cognitive status, age and functional status. Cox proportional hazards regression was used to evaluate associations with the distribution of conversion times from mild cognitive impairment to AD. The performances of classifiers developed using data from different types of brain tissue were systematically characterized across cognitive status groups. We also explored the performance of anatomical and cognitive-anatomical composite scores generated by combining the outputs of classifiers developed using different types of data. In addition, we provide the AD-PS scores performance relative to other metrics used in the field including the Spatial Pattern of Abnormalities for Recognition of Early AD (SPARE-AD) index and total hippocampal volume for the variables examined.

Validation of a cognitive assessment battery administered over the telephone.

OBJECTIVES: To evaluate the validity and reliability of a cognitive test battery and questionnaires administered by telephone. DESIGN: Observational study; 110 participants randomly assigned to receive two administrations of the same cognitive test battery 6 months apart in one of four combinations (Time 1 administration/Time 2 administration): telephone/telephone, telephone/face to face, face to face/telephone, face to face/face to face. SETTING: Academic medical center. PARTICIPANTS: One hundred ten women aged 65 to 90 without dementia. MEASUREMENTS: The battery included tests of attention; verbal learning and memory; verbal fluency; executive function; working memory; global cognitive functioning; and self-reported measures of perceived memory problems, depressive symptoms, sleep disturbance, and health-related quality of life. Test-retest reliability, concurrent validity, relative bias associated with telephone administration, and change scores were evaluated. RESULTS: There were no statistically significant differences in scores on any of the cognitive tests or questionnaires between participants randomly assigned to telephone or face-to-face administration at the Time 1 assessment, indicating equivalence across administration modes. There was no significant bias for tests or questionnaires administered by telephone (P's > .01), nor was there a difference in mean change scores between administration modes except for Category Fluency (P = .01) and California Verbal Learning Test long-delay free recall (P = .004). Mean test-retest coefficients for the battery were not significantly different between groups, although individual test-retest correlation coefficients were generally higher within modes than between modes. CONCLUSION: Telephone administration of cognitive tests and questionnaires to older women is reliable and valid. Use of telephone batteries can substantially reduce the cost and burden of cognitive assessments and increase enrollment, retention, and data completeness, thereby improving study validity.

Correspondence between in vivo (11)C-PiB-PET amyloid imaging and postmortem, region-matched assessment of plaques.

The definitive Alzheimer's disease (AD) diagnosis requires postmortem confirmation of neuropathological hallmarks-amyloid-ß (Aß) plaques and neurofibrillary tangles (NFTs). The advent of radiotracers for amyloid imaging presents an opportunity to investigate amyloid deposition in vivo. The (11)C-Pittsburgh compound-B (PiB)-PET ligand remains the most widely studied to date; however, regional variations in (11)C-PiB binding and the extent of agreement with neuropathological assessment have not been thoroughly investigated. Sojkova and colleagues [35] reported variable agreement between CERAD-based neuropathologic diagnosis of AD lesions and mean cortical PiB, suggesting the need for a more direct quantification of regional Aß in relation to in vivo imaging. In the present study, we extend these findings by examining the correspondence among regional (11)C-PiB load, region-matched quantitative immunohistological assessments of Aß and NFTs, and brain atrophy (MRI) in six older Baltimore Longitudinal Study of Aging participants who came to autopsy (imaging-autopsy interval range 0.2-2.4 years). The total number of Aß plaques (6E10) and NFTs (PHF1) in paraffin sections from hippocampus, orbito-frontal cortex, anterior and posterior cingulate gyrus, precuneus and cerebellum was quantified using a technique guided by unbiased stereological principles. We report a general agreement between the regional measures of amyloid obtained via stereological assessment and imaging, with significant relationships evident for the anterior (r = 0.83; p = 0.04) and posterior (r = 0.94; p = 0.005) cingulate gyri, and the precuneus (r = 0.94; p = 0.005). No associations were observed between (11)C-PiB load and NFT count for any of the regions examined (p > 0.2 in all regions), or between regional Aß or NFT counts and corresponding brain volumes. The strong associations of PiB retention with region-matched, quantitative analyses of Aß in postmortem tissue offer support for the validity of (11)C-PiB-PET imaging as a method for evaluation of plaque burden in vivo.

In vivo fibrillar beta-amyloid detected using [11C]PiB positron emission tomography and neuropathologic assessment in older adults.

BACKGROUND: In demented older adults, in vivo amyloid imaging shows agreement with diagnostic neuropathologic assessment of ß-amyloid (Aß). However, the extent of agreement in nondemented older adults remains unclear. OBJECTIVE: To compare Aß quantified using in vivo carbon 11-labeled Pittsburgh Compound B positron emission tomography and postmortem neuropathologic assessment of Aß in older adults. DESIGN: Case series. SETTING: Community-dwelling older adults who came to autopsy. PARTICIPANTS: Five nondemented and 1 demented participant from the Baltimore Longitudinal Study of Aging. MAIN OUTCOME MEASURE: Agreement between the mean cortical distribution volume ratio and the Consortium to Establish a Registry for AD (CERAD) neuritic plaque (NP) score used for pathologic diagnosis of Alzheimer disease. RESULTS: Of the 6 participants, 4 had moderate NPs, 2 had sparse or no detectable NPs, and 3 had microscopic findings of cerebral amyloid angiopathy at autopsy. On in vivo imaging, the mean cortical distribution volume ratio ranged from 0.96 to 1.59. Although there was agreement between in vivo amyloid imaging and CERAD NP scores in participants with either high or negligible Aß levels in vivo, only limited agreement was observed among those with intermediate levels of Aß. The best overall agreement was achieved at a distribution volume ratio of 1.2. CONCLUSIONS: In older adults, variable agreement between in vivo imaging and CERAD NP score was observed. The limited agreement may, in part, reflect differences in typical measurements of Aß using imaging compared with the CERAD neuropathologic protocol. Direct quantification of regional Aß in relation to in vivo imaging is necessary to further enhance our understanding of the imaging-pathologic assessment correlation.

Spatio-temporal analysis of brain MRI images using hidden Markov models.

A rapidly increasing number of medical imaging studies is longitudinal, i.e. involves series of repeated examinations of the same individuals. This paper presents a methodology for analysis of such 4D images, with brain aging as the primary application. An adaptive regional clustering method is first adopted to construct a spatial pattern, in which a measure of correlation between morphological measurements and a continuous patient's variable (age in our case) is used to group brain voxels into regions; Secondly, a dynamic probabilistic Hidden Markov Model (HMM) is created to statistically analyze the relationship between spatial brain patterns and hidden states; Thirdly, parametric HMM models under a bagging framework are used to capture the changes occurring with time by decoding the hidden states longitudinally. We apply this method to datasets from elderly individuals, and test the effectiveness of this spatio-temporal model in analyzing the temporal dynamics of spatial aging patterns on an individual basis. Experimental results show this method could facilitate the early detection of pathological brain change.

Longitudinal assessment of serum free testosterone concentration predicts memory performance and cognitive status in elderly men.

Circulating testosterone (T) levels have behavioral and neurological effects in both human and nonhuman species. Both T concentrations and neuropsychological function decrease substantially with age in men. The purpose of this prospective, longitudinal study was to investigate the relationships between age-associated decreases in endogenous serum T and free T concentrations and declines in neuropsychological performance. Participants were volunteers from the Baltimore Longitudinal Study of Aging, aged 50-91 yr at baseline T assessment. Four hundred seven men were followed for an average of 10 yr, with assessments of multiple cognitive domains and contemporaneous determination of serum total T, SHBG, and a free T index (FTI). We administered neuropsychological tests of verbal and visual memory, mental status, visuomotor scanning and attention, verbal knowledge/language, visuospatial ability, and depressive symptomatology. Higher FTI was associated with better scores on visual and verbal memory, visuospatial functioning, and visuomotor scanning and a reduced rate of longitudinal decline in visual memory. Men classified as hypogonadal had significantly lower scores on measures of memory and visuospatial performance and a faster rate of decline in visual memory. No relations between total T or the FTI and measures of verbal knowledge, mental status, or depressive symptoms were observed. These results suggest a possible beneficial relationship between circulating free T concentrations and specific domains of cognitive performance in older men.