Christopher G. Schwarz

YNIMG Journal 2025 Journal Article

Measuring the potential risk of re-identification of imaging research participants from open-source automated face recognition software

• Carl M. Prakaashana
• Marios Savvides
• Jeffrey L. Gunter
• Matthew L. Senjem
• Prashanthi Vemuri
• Kejal Kantarci
• Johnanthan Graff-Radford
• Ronald C. Petersen

In recent years facial recognition software has gone from an area of research to widespread adoption and broad public availability. Open-source face recognition packages are freely available on the internet for anyone to download, and several public websites allow users to run facial recognition on photos without needing any technical knowledge or equipment beyond internet access, making facial recognition accessible for anyone to use for any purpose. Previous research has demonstrated the ability of commercial software to identify a person based on facial content in brain imaging. In this study we tested two commercial facial recognition programs and a variety of popular open-source computer vision and facial recognition software packages to measure how accurately they could be used for reidentification of research participants in brain imaging studies. We tested a "population to sample" threat model, measuring the rates of success for which face recognition software selected the correct MRI-based face reconstruction from a set of 182 participants as its top-scoring match for input facial photographs. We found that the freely available open-source software packages we tested can reidentify a research participant with up to 59 % accuracy. This was less than the commercial packages, which were able to achieve much higher accuracies in the ranges of 92 % and 98 % in identical testing scenarios, but it demonstrates the feasibility of re-identifying faces in research MRI even by individuals with access to only freely available software. As the trust and confidence of potential participants is essential to brain imaging research, especially with widespread and mandated data-sharing of brain scans, this further supports the need to replace identifiable face imagery in brain images to protect the privacy of research participants.

YNICL Journal 2024 Journal Article

Speech-language within and between network disruptions in primary progressive aphasia variants

• Neha Singh-Reilly
• Hugo Botha
• Joseph R. Duffy
• Heather M. Clark
• Rene L. Utianski
• Mary M. Machulda
• Jonathan Graff-Radford
• Christopher G. Schwarz

Primary progressive aphasia (PPA) variants present with distinct disruptions in speech-language functions with little known about the interplay between affected and spared regions within the speech-language network and their interaction with other functional networks. The Neurodegenerative Research Group, Mayo Clinic, recruited 123 patients with PPA (55 logopenic (lvPPA), 44 non-fluent (nfvPPA) and 24 semantic (svPPA)) who were matched to 60 healthy controls. We investigated functional connectivity disruptions between regions within the left-speech-language network (Broca, Wernicke, anterior middle temporal gyrus (aMTG), supplementary motor area (SMA), planum temporale (PT) and parietal operculum (PO)), and disruptions to other networks (visual association, dorsal-attention, frontoparietal and default mode networks (DMN)). Within the speech-language network, multivariate linear regression models showed reduced aMTG-Broca connectivity in all variants, with lvPPA and nfvPPA findings remaining significant after Bonferroni correction. Additional loss in Wernicke-Broca connectivity in nfvPPA, Wernicke-PT connectivity in lvPPA and greater aMTG-PT connectivity in svPPA were also noted. Between-network connectivity findings in all variants showed reduced aMTG-DMN and increased aMTG-dorsal-attention connectivity, with additional disruptions between aMTG-visual association in both lvPPA and svPPA, aMTG-frontoparietal in lvPPA, and Wernicke-DMN breakdown in svPPA. These findings suggest that aMTG connectivity breakdown is a shared feature in all PPA variants, with lvPPA showing more extensive connectivity disruptions with other networks.

YNIMG Journal 2023 Journal Article

A face-off of MRI research sequences by their need for de-facing

• Christopher G. Schwarz
• Walter K. Kremers
• Arvin Arani
• Marios Savvides
• Robert I. Reid
• Jeffrey L. Gunter
• Matthew L. Senjem
• Petrice M. Cogswell

It is now widely known that research brain MRI, CT, and PET images may potentially be re-identified using face recognition, and this potential can be reduced by applying face-deidentification ("de-facing") software. However, for research MRI sequences beyond T1-weighted (T1-w) and T2-FLAIR structural images, the potential for re-identification and quantitative effects of de-facing are both unknown, and the effects of de-facing T2-FLAIR are also unknown. In this work we examine these questions (where applicable) for T1-w, T2-w, T2*-w, T2-FLAIR, diffusion MRI (dMRI), functional MRI (fMRI), and arterial spin labelling (ASL) sequences. Among current-generation, vendor-product research-grade sequences, we found that 3D T1-w, T2-w, and T2-FLAIR were highly re-identifiable (96-98%). 2D T2-FLAIR and 3D multi-echo GRE (ME-GRE) were also moderately re-identifiable (44-45%), and our derived T2* from ME-GRE (comparable to a typical 2D T2*) matched at only 10%. Finally, diffusion, functional and ASL images were each minimally re-identifiable (0-8%). Applying de-facing with mri_reface version 0.3 reduced successful re-identification to ≤8%, while differential effects on popular quantitative pipelines for cortical volumes and thickness, white matter hyperintensities (WMH), and quantitative susceptibility mapping (QSM) measurements were all either comparable with or smaller than scan-rescan estimates. Consequently, high-quality de-facing software can greatly reduce the risk of re-identification for identifiable MRI sequences with only negligible effects on automated intracranial measurements. The current-generation echo-planar and spiral sequences (dMRI, fMRI, and ASL) each had minimal match rates, suggesting that they have a low risk of re-identification and can be shared without de-facing, but this conclusion should be re-evaluated if they are acquired without fat suppression, with a full-face scan coverage, or if newer developments reduce the current levels of artifacts and distortion around the face.

YNIMG Journal 2023 Journal Article

Cross–scanner harmonization methods for structural MRI may need further work: A comparison study

• Robel K. Gebre
• Matthew L. Senjem
• Sheelakumari Raghavan
• Christopher G. Schwarz
• Jeffery L. Gunter
• Ekaterina I. Hofrenning
• Robert I. Reid
• Kejal Kantarci

The clinical usefulness MRI biomarkers for aging and dementia studies relies on precise brain morphological measurements; however, scanner and/or protocol variations may introduce noise or bias. One approach to address this is post-acquisition scan harmonization. In this work, we evaluate deep learning (neural style transfer, CycleGAN and CGAN), histogram matching, and statistical (ComBat and LongComBat) methods. Participants who had been scanned on both GE and Siemens scanners (cross-sectional participants, known as Crossover (n = 113), and longitudinally scanned participants on both scanners (n = 454)) were used. The goal was to match GE MPRAGE (T1-weighted) scans to Siemens improved resolution MPRAGE scans. Harmonization was performed on raw native and preprocessed (resampled, affine transformed to template space) scans. Cortical thicknesses were measured using FreeSurfer (v.7.1.1). Distributions were checked using Kolmogorov-Smirnov tests. Intra-class correlation (ICC) was used to assess the degree of agreement in the Crossover datasets and annualized percent change in cortical thickness was calculated to evaluate the Longitudinal datasets. Prior to harmonization, the least agreement was found at the frontal pole (ICC = 0.72) for the raw native scans, and at caudal anterior cingulate (0.76) and frontal pole (0.54) for the preprocessed scans. Harmonization with NST, CycleGAN, and HM improved the ICCs of the preprocessed scans at the caudal anterior cingulate (>0.81) and frontal poles (>0.67). In the Longitudinal raw native scans, over- and under-estimations of cortical thickness were observed due to the changing of the scanners. ComBat matched the cortical thickness distributions throughout but was not able to increase the ICCs or remove the effects of scanner changeover in the Longitudinal datasets. CycleGAN and NST performed slightly better to address the cortical thickness variations between scanner change. However, none of the methods succeeded in harmonizing the Longitudinal dataset. CGAN was the worst performer for both datasets. In conclusion, the performance of the methods was overall similar and region dependent. Future research is needed to improve the existing approaches since none of them outperformed each other in terms of harmonizing the datasets at all ROIs. The findings of this study establish framework for future research into the scan harmonization problem.

YNICL Journal 2023 Journal Article

Effects of de-facing software mri_reface on utility of imaging biomarkers used in Alzheimer’s disease research

• Christopher G. Schwarz
• Walter K. Kremers
• Stephen D. Weigand
• Carl M. Prakaashana
• Matthew L. Senjem
• Scott A. Przybelski
• Val J. Lowe
• Jeffrey L. Gunter

Brain imaging research studies increasingly use "de-facing" software to remove or replace facial imagery before public data sharing. Several works have studied the effects of de-facing software on brain imaging biomarkers by directly comparing automated measurements from unmodified vs de-faced images, but most research brain images are used in analyses of correlations with cognitive measurements or clinical statuses, and the effects of de-facing on these types of imaging-to-cognition correlations has not been measured. In this work, we focused on brain imaging measures of amyloid (A), tau (T), neurodegeneration (N), and vascular (V) measures used in Alzheimer's Disease (AD) research. We created a retrospective sample of participants from three age- and sex-matched clinical groups (cognitively unimpaired, mild cognitive impairment, and AD dementia, and we performed region- and voxel-wise analyses of: hippocampal volume (N), white matter hyperintensity volume (V), amyloid PET (A), and tau PET (T) measures, each from multiple software pipelines, on their ability to separate cognitively defined groups and their degrees of correlation with age and Clinical Dementia Rating (CDR)-Sum of Boxes (CDR-SB). We performed each of these analyses twice: once with unmodified images and once with images de-faced with leading de-facing software mri_reface, and we directly compared the findings and their statistical strengths between the original vs. the de-faced images. Analyses with original and with de-faced images had very high agreement. There were no significant differences between any voxel-wise comparisons. Among region-wise comparisons, only three out of 55 correlations were significantly different between original and de-faced images, and these were not significant after correction for multiple comparisons. Overall, the statistical power of the imaging data for AD biomarkers was almost identical between unmodified and de-faced images, and their analyses results were extremely consistent.

YNIMG Journal 2022 Journal Article

Deep learning identifies brain structures that predict cognition and explain heterogeneity in cognitive aging

• Krishnakant V. Saboo
• Chang Hu
• Yogatheesan Varatharajah
• Scott A. Przybelski
• Robert I. Reid
• Christopher G. Schwarz
• Jonathan Graff-Radford
• David S. Knopman

Specific brain structures (gray matter regions and white matter tracts) play a dominant role in determining cognitive decline and explain the heterogeneity in cognitive aging. Identification of these structures is crucial for screening of older adults at risk of cognitive decline. Using deep learning models augmented with a model-interpretation technique on data from 1432 Mayo Clinic Study of Aging participants, we identified a subset of brain structures that were most predictive of individualized cognitive trajectories and indicative of cognitively resilient vs. vulnerable individuals. Specifically, these structures explained why some participants were resilient to the deleterious effects of elevated brain amyloid and poor vascular health. Of these, medial temporal lobe and fornix, reflective of age and pathology-related degeneration, and corpus callosum, reflective of inter-hemispheric disconnection, accounted for 60% of the heterogeneity explained by the most predictive structures. Our results are valuable for identifying cognitively vulnerable individuals and for developing interventions for cognitive decline.

YNICL Journal 2022 Journal Article

Distinct brain iron profiles associated with logopenic progressive aphasia and posterior cortical atrophy

• Neha Atulkumar Singh
• Arvin Arani
• Jonathan Graff-Radford
• Matthew L. Senjem
• Peter R. Martin
• Mary M. Machulda
• Christopher G. Schwarz
• Yunhong Shu

Quantitative susceptibility mapping (QSM) can detect iron distribution in the brain by estimating local tissue magnetic susceptibility properties at every voxel. Iron deposition patterns are well studied in typical Alzheimer's disease (tAD), but little is known about these patterns in atypical clinical presentations of AD such as logopenic progressive aphasia (LPA) and posterior cortical atrophy (PCA). Seventeen PCA patients and eight LPA patients were recruited by the Neurodegenerative Research Group at Mayo Clinic, Rochester, MN, and underwent MRI that included a five-echo gradient echo sequence for calculation of QSM. Mean QSM signal was extracted from gray and white matter for regions-of-interest across the brain using the Mayo Clinic Adult Lifespan Template. Bayesian hierarchical models were fit per-region and per-hemisphere to compare PCA, LPA, 63 healthy controls, and 20 tAD patients. Strong evidence (posterior probability > 0.99) was observed for greater susceptibility in the middle occipital gyrus and amygdala in both LPA and PCA, and in the right inferior parietal, inferior temporal, and angular gyri in PCA and the caudate and substantia nigra in LPA compared to controls. Moderate evidence for greater susceptibility (posterior probability > 0.90) was also observed in the inferior occipital gyrus, precuneus, putamen and entorhinal cortex in both LPA and PCA, along with superior frontal gyrus in PCA and inferior temporal gyri, insula and basal ganglia in LPA, when compared to controls. Between phenotypic comparisons, LPA had greater susceptibility in the caudate, hippocampus, and posterior cingulate compared to PCA, while PCA showed greater susceptibility in the right superior frontal and middle temporal gyri compared to LPA. Both LPA and PCA showed moderate and strong evidence for greater susceptibility than tAD, particularly in medial and lateral parietal regions, while tAD showed greater susceptibility in the hippocampus and basal ganglia. This study proposes the possibility of unique iron profiles existing between LPA and PCA within cortical and subcortical structures. These changes match well with the disease-related changes of the clinical phenotypes, suggesting that QSM could be an informative candidate marker to study iron deposition in these patients.

YNIMG Journal 2022 Journal Article

Face recognition from research brain PET: An unexpected PET problem

• Christopher G. Schwarz
• Walter K. Kremers
• Val J. Lowe
• Marios Savvides
• Jeffrey L. Gunter
• Matthew L. Senjem
• Prashanthi Vemuri
• Kejal Kantarci

It is well known that de-identified research brain images from MRI and CT can potentially be re-identified using face recognition; however, this has not been examined for PET images. We generated face reconstruction images of 182 volunteers using amyloid, tau, and FDG PET scans, and we measured how accurately commercial face recognition software (Microsoft Azure's Face API) automatically matched them with the individual participants' face photographs. We then compared this accuracy with the same experiments using participants' CT and MRI. Face reconstructions from PET images from PET/CT scanners were correctly matched at rates of 42% (FDG), 35% (tau), and 32% (amyloid), while CT were matched at 78% and MRI at 97-98%. We propose that these recognition rates are high enough that research studies should consider using face de-identification ("de-facing") software on PET images, in addition to CT and structural MRI, before data sharing. We also updated our mri_reface de-identification software with extended functionality to replace face imagery in PET and CT images. Rates of face recognition on de-faced images were reduced to 0-4% for PET, 5% for CT, and 8% for MRI. We measured the effects of de-facing on regional amyloid PET measurements from two different measurement pipelines (PETSurfer/FreeSurfer 6.0, and one in-house method based on SPM12 and ANTs), and these effects were small: ICC values between de-faced and original images were > 0.98, biases were <2%, and median relative errors were < 2%. Effects on global amyloid PET SUVR measurements were even smaller: ICC values were 1.00, biases were <0.5%, and median relative errors were also <0.5%.

YNICL Journal 2022 Journal Article

Tractography of supplementary motor area projections in progressive speech apraxia and aphasia

• Adrian Valls Carbo
• Robert I. Reid
• Nirubol Tosakulwong
• Stephen D. Weigand
• Joseph R. Duffy
• Heather M. Clark
• Rene L. Utianski
• Hugo Botha

Progressive apraxia of speech (AOS) is a motor speech disorder affecting the ability to produce phonetically or prosodically normal speech. Progressive AOS can present in isolation or co-occur with agrammatic aphasia and is associated with degeneration of the supplementary motor area. We aimed to assess breakdowns in structural connectivity from the supplementary motor area in patients with any combination of progressive AOS and/or agrammatic aphasia to determine which supplementary motor area tracts are specifically related to these clinical symptoms. Eighty-four patients with progressive AOS or progressive agrammatic aphasia were recruited by the Neurodegenerative Research Group and underwent neurological, speech/language, and neuropsychological testing, as well as 3 T diffusion magnetic resonance imaging. Of the 84 patients, 36 had apraxia of speech in isolation (primary progressive apraxia of speech, PPAOS), 40 had apraxia of speech and agrammatic aphasia (AOS-PAA), and eight had agrammatic aphasia in isolation (progressive agrammatic aphasia, PAA). Tractography was performed to identify 5 distinct tracts connecting to the supplementary motor area. Fractional anisotropy and mean diffusivity were assessed at 10 positions along the length of the tracts to construct tract profiles, and median profiles were calculated for each tract. In a case-control comparison, decreased fractional anisotropy and increased mean diffusivity were observed along the supplementary motor area commissural fibers in all three groups compared to controls. PPAOS also had abnormal diffusion in tracts from the supplementary motor area to the putamen, prefrontal cortex, Broca's area (frontal aslant tract) and motor cortex, with greatest abnormalities observed closest to the supplementary motor area. The AOS-PAA group showed abnormalities in the same set of tracts, but with greater involvement of the supplementary motor area to prefrontal tract compared to PPAOS. PAA showed abnormalities in the left prefrontal and frontal aslant tracts compared to both other groups, with PAA showing greatest abnormalities furthest from the supplementary motor area. Severity of AOS correlated with tract metrics in the supplementary motor area commissural and motor cortex tracts. Severity of aphasia correlated with the frontal aslant and prefrontal tracts. These findings provide insight into how AOS and agrammatism are differentially related to disrupted diffusivity, with progressive AOS associated with abnormalities close to the supplementary motor area, and the frontal aslant and prefrontal tracts being particularly associated with agrammatic aphasia.

YNIMG Journal 2021 Journal Article

Associations of quantitative susceptibility mapping with Alzheimer's disease clinical and imaging markers

• Petrice M. Cogswell
• Heather J. Wiste
• Matthew L. Senjem
• Jeffrey L. Gunter
• Stephen D. Weigand
• Christopher G. Schwarz
• Arvin Arani
• Terry M. Therneau

Altered iron metabolism has been hypothesized to be associated with Alzheimer's disease pathology, and prior work has shown associations between iron load and beta amyloid plaques. Quantitative susceptibility mapping (QSM) is a recently popularized MR technique to infer local tissue susceptibility secondary to the presence of iron as well as other minerals. Greater QSM values imply greater iron concentration in tissue. QSM has been used to study relationships between cerebral iron load and established markers of Alzheimer's disease, however relationships remain unclear. In this work we study QSM signal characteristics and associations between susceptibility measured on QSM and established clinical and imaging markers of Alzheimer's disease. The study included 421 participants (234 male, median age 70 years, range 34-97 years) from the Mayo Clinic Study of Aging and Alzheimer's Disease Research Center; 296 (70%) had a diagnosis of cognitively unimpaired, 69 (16%) mild cognitive impairment, and 56 (13%) amnestic dementia. All participants had multi-echo gradient recalled echo imaging, PiB amyloid PET, and Tauvid tau PET. Variance components analysis showed that variation in cortical susceptibility across participants was low. Linear regression models were fit to assess associations with regional susceptibility. Expected increases in susceptibility were found with older age and cognitive impairment in the deep and inferior gray nuclei (pallidum, putamen, substantia nigra, subthalamic nucleus) (betas: 0.0017 to 0.0053 ppm for a 10 year increase in age, p = 0.03 to <0.001; betas: 0.0021 to 0.0058 ppm for a 5 point decrease in Short Test of Mental Status, p = 0.003 to p<0.001). Effect sizes in cortical regions were smaller, and the age associations were generally negative. Higher susceptibility was significantly associated with higher amyloid PET SUVR in the pallidum and putamen (betas: 0.0029 and 0.0012 ppm for a 20% increase in amyloid PET, p = 0.05 and 0.02, respectively), higher tau PET in the basal ganglia with the largest effect size in the pallidum (0.0082 ppm for a 20% increase in tau PET, p<0.001), and with lower cortical gray matter volume in the medial temporal lobe (0.0006 ppm for a 20% decrease in volume, p = 0.03). Overall, these findings suggest that susceptibility in the deep and inferior gray nuclei, particularly the pallidum and putamen, may be a marker of cognitive decline, amyloid deposition, and off-target binding of the tau ligand. Although iron has been demonstrated in amyloid plaques and in association with neurodegeneration, it is of insufficient quantity to be reliably detected in the cortex using this implementation of QSM.

YNIMG Journal 2021 Journal Article

Changing the face of neuroimaging research: Comparing a new MRI de-facing technique with popular alternatives

• Christopher G. Schwarz
• Walter K. Kremers
• Heather J. Wiste
• Jeffrey L. Gunter
• Prashanthi Vemuri
• Anthony J. Spychalla
• Kejal Kantarci
• Aaron P. Schultz

Recent advances in automated face recognition algorithms have increased the risk that de-identified research MRI scans may be re-identifiable by matching them to identified photographs using face recognition. A variety of software exist to de-face (remove faces from) MRI, but their ability to prevent face recognition has never been measured and their image modifications can alter automated brain measurements. In this study, we compared three popular de-facing techniques and introduce our mri_reface technique designed to minimize effects on brain measurements by replacing the face with a population average, rather than removing it. For each technique, we measured 1) how well it prevented automated face recognition (i.e. effects on exceptionally-motivated individuals) and 2) how it altered brain measurements from SPM12, FreeSurfer, and FSL (i.e. effects on the average user of de-identified data). Before de-facing, 97% of scans from a sample of 157 volunteers were correctly matched to photographs using automated face recognition. After de-facing with popular software, 28-38% of scans still retained enough data for successful automated face matching. Our proposed mri_reface had similar performance with the best existing method (fsl_deface) at preventing face recognition (28-30%) and it had the smallest effects on brain measurements in more pipelines than any other, but these differences were modest.

YNIMG Journal 2021 Journal Article

CSF dynamics as a predictor of cognitive progression

• Petrice M. Cogswell
• Stephen D. Weigand
• Heather J. Wiste
• Jeffrey L. Gunter
• Jonathan Graff-Radford
• David T. Jones
• Christopher G. Schwarz
• Matthew L. Senjem

Disproportionately enlarged subarachnoid-space hydrocephalus (DESH), characterized by tight high convexity CSF spaces, ventriculomegaly, and enlarged Sylvian fissures, is thought to be an indirect marker of a CSF dynamics disorder. The clinical significance of DESH with regard to cognitive decline in a community setting is not yet well defined. The goal of this work is to determine if DESH is associated with cognitive decline. Participants in the population-based Mayo Clinic Study of Aging (MCSA) who met the following criteria were included: age ≥ 65 years, 3T MRI, and diagnosis of cognitively unimpaired or mild cognitive impairment at enrollment as well as at least one follow-up visit with cognitive testing. A support vector machine based method to detect the DESH imaging features on T1-weighted MRI was used to calculate a "DESH score", with positive scores indicating a more DESH-like imaging pattern. For the participants who were cognitively unimpaired at enrollment, a Cox proportional hazards model was fit with time defined as years from enrollment to first diagnosis of mild cognitive impairment or dementia, or as years to last known cognitively unimpaired diagnosis for those who did not progress. Linear mixed effects models were fit among all participants to estimate annual change in cognitive z scores for each domain (memory, attention, language, and visuospatial) and a global z score. For all models, covariates included age, sex, education, APOE genotype, cortical thickness, white matter hyperintensity volume, and total intracranial volume. The hazard of progression to cognitive impairment was an estimated 12% greater for a DESH score of +1 versus -1 (HR 1.12, 95% CI 0.97-1.31, p = 0.11). Global and attention cognition declined 0.015 (95% CI 0.005-0.025) and 0.016 (95% CI 0.005-0.028) z/year more, respectively, for a DESH score of +1 vs -1 (p = 0.01 and p = 0.02), with similar, though not statistically significant DESH effects in the other cognitive domains. Imaging features of disordered CSF dynamics are an independent predictor of subsequent cognitive decline in the MCSA, among other well-known factors including age, cortical thickness, and APOE status. Therefore, since DESH contributes to cognitive decline and is present in the general population, identifying individuals with DESH features may be important clinically as well as for selection in clinical trials.

YNICL Journal 2021 Journal Article

FDG PET metabolic signatures distinguishing prodromal DLB and prodromal AD

• Kejal Kantarci
• Bradley F. Boeve
• Scott A. Przybelski
• Timothy G. Lesnick
• Qin Chen
• Julie Fields
• Christopher G. Schwarz
• Matthew L. Senjem

BACKGROUND AND PURPOSE: F-fluorodeoxyglucose (FDG) PET. Whether this pattern of hypometabolism is present as early as the prodromal stage of DLB is unknown. We investigated the pattern of hypometabolism in patients with mild cognitive impairment (MCI) who progressed to probable DLB compared to MCI patients who progressed to Alzheimer's disease (AD) dementia and clinically unimpaired (CU) controls. METHODS: Patients with MCI from the Mayo Clinic Alzheimer's Disease Research Center who underwent FDG PET at baseline and progressed to either probable DLB (MCI-DLB; n = 17) or AD dementia (MCI-AD; n = 41) during follow-up, and a comparison cohort of CU controls (n = 100) were included. RESULTS: Patients with MCI-DLB had hypometabolism in the parieto-occipital cortex extending into temporal lobes, substantia nigra and thalamus. When compared to MCI-AD, medial temporal and posterior cingulate metabolism were preserved in patients with MCI-DLB, accompanied by greater hypometabolism in the substantia nigra in MCI-DLB compared to MCI-AD. In distinguishing MCI-DLB from MCI-AD at the maximum value of Youden's index, CIS ratio was highly specific (90%) but not sensitive (59%), but a higher medial temporal to substantia nigra ratio was both sensitive (94%) and specific (83%). CONCLUSION: FDG PET is a potential biomarker for the prodromal stage of DLB. A higher medial temporal metabolism and CIS ratio, and lower substantia nigra metabolism have additive value in distinguishing prodromal DLB and AD.

YNICL Journal 2021 Journal Article

Neuroimaging correlates of gait abnormalities in progressive supranuclear palsy

• Irene Sintini
• Kenton Kaufman
• Hugo Botha
• Peter R. Martin
• Stacy R. Loushin
• Matthew L. Senjem
• Robert I. Reid
• Christopher G. Schwarz

Progressive supranuclear palsy is a neurodegenerative disorder characterized primarily by tau inclusions and neurodegeneration in the midbrain, basal ganglia, thalamus, premotor and frontal cortex. Neurodegenerative change in progressive supranuclear palsy has been assessed using MRI. Degeneration of white matter tracts is evident with diffusion tensor imaging and PET methods have been used to assess brain metabolism or presence of tau protein deposits. Patients with progressive supranuclear palsy present with a variety of clinical syndromes; however early onset of gait impairments and postural instability are common features. In this study we assessed the relationship between multimodal imaging biomarkers (i.e., MRI atrophy, white matter tracts degeneration, flortaucipir-PET uptake) and laboratory-based measures of gait and balance abnormalities in a cohort of nineteen patients with progressive supranuclear palsy, using univariate and multivariate statistical analyses. The PSP rating scale and its gait midline sub-score were strongly correlated to gait abnormalities but not to postural imbalance. Principal component analysis on gait variables identified velocity, stride length, gait stability ratio, length of gait phases and dynamic stability as the main contributors to the first component, which was associated with diffusion tensor imaging measures in the posterior thalamic radiation, external capsule, superior cerebellar peduncle, superior fronto-occipital fasciculus, body and splenium of the corpus callosum and sagittal stratum, with MRI volumes in frontal and precentral regions and with flortaucipir-PET uptake in the precentral gyrus. The main contributor to the second principal component was cadence, which was higher in patients presenting more abnormalities on mean diffusivity: this unexpected finding might be related to compensatory gait strategies adopted in progressive supranuclear palsy. Postural imbalance was the main contributor to the third principal component, which was related to flortaucipir-PET uptake in the left paracentral lobule and supplementary motor area and white matter disruption in the superior cerebellar peduncle, putamen, pontine crossing tract and corticospinal tract. A partial least square model identified flortaucipir-PET uptake in midbrain, basal ganglia and thalamus as the main correlate of speed and dynamic component of gait in progressive supranuclear palsy. Although causality cannot be established in this analysis, our study sheds light on neurodegeneration of brain regions and white matter tracts that underlies gait and balance impairment in progressive supranuclear palsy.

YNIMG Journal 2021 Journal Article

Relationships between β-amyloid and tau in an elderly population: An accelerated failure time model

• Terry M. Therneau
• David S. Knopman
• Val J. Lowe
• Hugo Botha
• Jonathan Graff-Radford
• David T. Jones
• Prashanthi Vemuri
• Michelle M. Mielke

Using positron emission tomography (PET)-derived amyloid and tau measurements from 1,495 participants, we explore the evolution of these values over time via an accelerated failure time (AFT) model. The AFT model assumes a shared pattern of progression, but one which is shifted earlier or later in time for each individual; an individual's time shift for amyloid and for tau are assumed to be linked. The resulting pattern for each outcome consists of an earlier indolent phase followed by sharp progression of the accumulation rate. APOE ε4 shifts the amyloid curve leftward (earlier) by 6.1 years, and the tau curve leftward by 2.6 years. Female sex shifts the amyloid curve leftward by 2.4 years and the tau curve leftward by 2.6 years. Per-person shifts (i.e., the individual's deviation from the population mean) for the onset of amyloid accumulation ranged from 13 years earlier to 13 years later (10th to 90th percentile) than average and 11 years earlier to 14 years later for tau, with an estimated correlation of 0.49. The average delay between amyloid increase and tau increase was 13.3 years.

YNIMG Journal 2021 Journal Article

Selecting software pipelines for change in flortaucipir SUVR: Balancing repeatability and group separation

• Christopher G. Schwarz
• Terry M. Therneau
• Stephen D. Weigand
• Jeffrey L. Gunter
• Val J. Lowe
• Scott A. Przybelski
• Matthew L. Senjem
• Hugo Botha

Since tau PET tracers were introduced, investigators have quantified them using a wide variety of automated methods. As longitudinal cohort studies acquire second and third time points of serial within-person tau PET data, determining the best pipeline to measure change has become crucial. We compared a total of 415 different quantification methods (each a combination of multiple options) according to their effects on a) differences in annual SUVR change between clinical groups, and b) longitudinal measurement repeatability as measured by the error term from a linear mixed-effects model. Our comparisons used MRI and Flortaucipir scans of 97 Mayo Clinic study participants who clinically either: a) were cognitively unimpaired, or b) had cognitive impairments that were consistent with Alzheimer's disease pathology. Tested methods included cross-sectional and longitudinal variants of two overarching pipelines (FreeSurfer 6.0, and an in-house pipeline based on SPM12), three choices of target region (entorhinal, inferior temporal, and a temporal lobe meta-ROI), five types of partial volume correction (PVC) (none, two-compartment, three-compartment, geometric transfer matrix (GTM), and a tau-specific GTM variant), seven choices of reference region (cerebellar crus, cerebellar gray matter, whole cerebellum, pons, supratentorial white matter, eroded supratentorial WM, and a composite of eroded supratentorial WM, pons, and whole cerebellum), two choices of region masking (GM or GM and WM), and two choices of statistic (voxel-wise mean vs. median). Our strongest findings were: 1) larger temporal-lobe target regions greatly outperformed entorhinal cortex (median sample size estimates based on a hypothetical clinical trial were 520-526 vs. 1740); 2) longitudinal processing pipelines outperformed cross-sectional pipelines (median sample size estimates were 483 vs. 572); and 3) reference regions including supratentorial WM outperformed traditional cerebellar and pontine options (median sample size estimates were 370 vs. 559). Altogether, our results favored longitudinally SUVR methods and a temporal-lobe meta-ROI that includes adjacent (juxtacortical) WM, a composite reference region (eroded supratentorial WM + pons + whole cerebellum), 2-class voxel-based PVC, and median statistics.

YNICL Journal 2020 Journal Article

Brain volume and flortaucipir analysis of progressive supranuclear palsy clinical variants

• Jennifer L. Whitwell
• Nirubol Tosakulwong
• Hugo Botha
• Farwa Ali
• Heather M. Clark
• Joseph R. Duffy
• Rene L. Utianski
• Chase A. Stevens

BACKGROUND AND PURPOSE: F]flortaucipir uptake on PET across PSP variants. MATERIALS AND METHODS: 105 PSP patients (53 PSP-RS, 23 PSP-SL, 12 PSP-P, 8 PSP-CBS, 5 PSP-F and 4 PSP-PGF) underwent volumetric MRI, with 59 of these also undergoing flortaucipir PET. Voxel-level and region-level analyses were performed comparing PSP variants to 30 controls and to each other. Semi-quantitative tau burden measurements were also performed in 21 patients with autopsy-confirmed PSP. RESULTS: All variants showed evidence for atrophy or increased flortaucipir uptake in striatum, globus pallidus and thalamus. Superior cerebellar peduncle volume loss was only observed in PSP-RS, PSP-CBS and PSP-F. Volume loss in the frontal lobes was observed in PSP-SL, PSP-CBS and PSP-F, with these variants also showing highest cortical tau burden at autopsy. The PSP-P and PSP-PGF variants showed more restricted patterns of neurodegeneration predominantly involving striatum, globus pallidus, subthalamic nucleus and thalamus. The PSP-SL variant showed greater volume loss and flortaucipir uptake in supplementary motor area and motor cortex compared to all other variants, but showed less involvement of subthalamic nucleus and midbrain. Compared to PSP-RS, PSP-P had larger midbrain volume and greater flortaucipir uptake in putamen. CONCLUSION: The PSP variants have different patterns of involvement of subcortical circuitry, perhaps suggesting different patterns of disease spread through the brain. These findings will be important in the development of appropriate neuroimaging biomarkers for the different PSP variants.

YNICL Journal 2019 Journal Article

Automated detection of imaging features of disproportionately enlarged subarachnoid space hydrocephalus using machine learning methods

• Nathaniel B. Gunter
• Christopher G. Schwarz
• Jonathan Graff-Radford
• Jeffrey L. Gunter
• David T. Jones
• Neill R. Graff-Radford
• Ronald C. Petersen
• David S. Knopman

OBJECTIVE: Create an automated classifier for imaging characteristics of disproportionately enlarged subarachnoid space hydrocephalus (DESH), a neuroimaging phenotype of idiopathic normal pressure hydrocephalus (iNPH). METHODS: 1597 patients from the Mayo Clinic Study of Aging (MCSA) were reviewed for imaging characteristics of DESH. One core feature of DESH, the presence of tightened sulci in the high-convexities (THC), was used as a surrogate for the presence of DESH as the expert clinician-defined criterion on which the classifier was trained. Anatomical MRI scans were automatically segmented for cerebrospinal fluid (CSF) and overlaid with an atlas of 123 named sulcal regions. The volume of CSF in each sulcal region was summed and normalized to total intracranial volume. Area under the receiver operating characteristic curve (AUROC) values were computed for each region individually, and these values determined feature selection for the machine learning model. Due to class imbalance in the data (72 selected scans out of 1597 total scans) adaptive synthetic sampling (a technique which generates synthetic examples based on the original data points) was used to balance the data. A support vector machine model was then trained on the regions selected. RESULTS: Using the automated classification model, we were able to classify scans for tightened sulci in the high convexities, as defined by the expert clinician, with an AUROC of about 0.99 (false negative ≈ 2%, false positive ≈ 5%). Ventricular volumes were among the classifier's most discriminative features but are not specific for DESH. The inclusion of regions outside the ventricles allowed specificity from atrophic neurodegenerative diseases that are also accompanied by ventricular enlargement. CONCLUSION: Automated detection of tight high convexity, a key imaging feature of DESH, is possible by using support vector machine models with selected sulcal CSF volumes as features.

YNICL Journal 2019 Journal Article

Longitudinal tau-PET uptake and atrophy in atypical Alzheimer's disease

• Irene Sintini
• Peter R. Martin
• Jonathan Graff-Radford
• Matthew L. Senjem
• Christopher G. Schwarz
• Mary M. Machulda
• Anthony J. Spychalla
• Daniel A. Drubach

F]AV-1451 PET and PiB PET. Region- and voxel-level rates of tau accumulation and grey matter atrophy relative to cognitively unimpaired individuals, and the influence of age on such rates, were assessed. Univariate and multivariate analyses were performed between baseline measurements and rates of change, between baseline tau and atrophy, and between the two rates of change. Regional patterns of change in tau and volume differed, with highest rates of tau accumulation in frontal lobe and highest rates of atrophy in temporoparietal regions. Age had a negative effect on disease progression, predominantly on tau, with younger patients having a more rapid accumulation. Baseline tau uptake and regions of tau accumulation were disconnected, with high baseline tau uptake across the cortex correlated with high rates of tau accumulation in frontal and sensorimotor regions. In contrast, baseline volume and atrophy were locally related in the occipitoparietal regions. Higher tau uptake at baseline was locally related to higher rates of atrophy in frontal and occipital lobes. Tau accumulation rates positively correlated with rates of atrophy. In summary, our study showed that tau accumulation and atrophy presented different regional patterns in atypical AD, with tau spreading into the frontal lobes while atrophy remains in temporoparietal and occipital cortex, suggesting a temporal disconnect between protein deposition and neurodegeneration.

YNICL Journal 2017 Journal Article

Comparison of [18F]Flutemetamol and [11C]Pittsburgh Compound-B in cognitively normal young, cognitively normal elderly, and Alzheimer's disease dementia individuals

• Val J. Lowe
• Emily Lundt
• David Knopman
• Matthew L. Senjem
• Jeffrey L. Gunter
• Christopher G. Schwarz
• Bradley J. Kemp
• Clifford R. Jack

BACKGROUND: C]PiB (PiB) PET in the same people. METHODS: -test, respectively. We also compared the ability of the two tracers to discriminate between diagnostic groups using AUROC estimates. The effect of using different normalization regions on SUVr values was also evaluated. RESULTS: Both FMT and PiB showed greater uptake throughout GM structures in AD vs. eCN or yCN. In all dual-modality group comparisons (FMT vs. PiB in yCN, eCN, and AD), greater WM uptake was seen with FMT vs. PiB. In yCN and eCN greater diffuse GM uptake was seen with FMT vs. PiB. When comparing yCN to eCN within each tracer, greater WM uptake was seen in eCN vs yCN. CONCLUSIONS: Flutemetamol and PiB show similar topographical GM uptake in AD and CN participants and the tracers show comparable group discrimination. Greater WM accumulation with FMT suggests that quantitative differences vs. PiB will be apparent when using WM or GM as a reference region. Both imaging tracers demonstrate increased WM uptake in older people. These findings suggest that using different amyloid tracers or different methods of analyses in serial brain imaging in an individual may result in artifactual amyloid change measurements. Clinical use of several amyloid tracers in the same patient will have challenges that need to be carefully considered.

YNIMG Journal 2017 Journal Article

Optimizing PiB-PET SUVR change-over-time measurement by a large-scale analysis of longitudinal reliability, plausibility, separability, and correlation with MMSE

• Christopher G. Schwarz
• Matthew L. Senjem
• Jeffrey L. Gunter
• Nirubol Tosakulwong
• Stephen D. Weigand
• Bradley J. Kemp
• Anthony J. Spychalla
• Prashanthi Vemuri

YNICL Journal 2016 Journal Article

A large-scale comparison of cortical thickness and volume methods for measuring Alzheimer's disease severity

• Christopher G. Schwarz
• Jeffrey L. Gunter
• Heather J. Wiste
• Scott A. Przybelski
• Stephen D. Weigand
• Chadwick P. Ward
• Matthew L. Senjem
• Prashanthi Vemuri

Alzheimer's disease (AD) researchers commonly use MRI as a quantitative measure of disease severity. Historically, hippocampal volume has been favored. Recently, “AD signature” measurements of gray matter (GM) volumes or cortical thicknesses have gained attention. Here, we systematically evaluate multiple thickness- and volume-based candidate-methods side-by-side, built using the popular FreeSurfer, SPM, and ANTs packages, according to the following criteria: (a) ability to separate clinically normal individuals from those with AD; (b) (extent of) correlation with head size, a nuisance covariatel (c) reliability on repeated scans; and (d) correlation with Braak neurofibrillary tangle stage in a group with autopsy. We show that volume- and thickness-based measures generally perform similarly for separating clinically normal from AD populations, and in correlation with Braak neurofibrillary tangle stage at autopsy. Volume-based measures are generally more reliable than thickness measures. As expected, volume measures are highly correlated with head size, while thickness measures are generally not. Because approaches to statistically correcting volumes for head size vary and may be inadequate to deal with this underlying confound, and because our goal is to determine a measure which can be used to examine age and sex effects in a cohort across a large age range, we thus recommend thickness-based measures. Ultimately, based on these criteria and additional practical considerations of run-time and failure rates, we recommend an AD signature measure formed from a composite of thickness measurements in the entorhinal, fusiform, parahippocampal, mid-temporal, inferior-temporal, and angular gyrus ROIs using ANTs with input segmentations from SPM12.

YNIMG Journal 2014 Journal Article

Improved DTI registration allows voxel-based analysis that outperforms Tract-Based Spatial Statistics

• Christopher G. Schwarz
• Robert I. Reid
• Jeffrey L. Gunter
• Matthew L. Senjem
• Scott A. Przybelski
• Samantha M. Zuk
• Jennifer L. Whitwell
• Prashanthi Vemuri