A Novel Approach to Clustering Accelerometer Data for Application in Passive Predictions of Changes in Depression Severity.

The treatment of mood disorders, which can become a lifelong process, varies widely in efficacy between individuals. Most options to monitor mood rely on subjective self-reports and clinical visits, which can be burdensome and may not portray an accurate representation of what the individual is experiencing. A passive method to monitor mood could be a useful tool for those with these disorders. Some previously proposed models utilized sensors from smartphones and wearables, such as the accelerometer. This study examined a novel approach of processing accelerometer data collected from smartphones only while participants of the open-science branch of the BiAffect study were typing. The data were modeled by von Mises-Fisher distributions and weighted networks to identify clusters relating to different typing positions unique for each participant. Longitudinal features were derived from the clustered data and used in machine learning models to predict clinically relevant changes in depression from clinical and typing measures. Model accuracy was approximately 95%, with 97% area under the ROC curve (AUC). The accelerometer features outperformed the vast majority of clinical and typing features, which suggested that this new approach to analyzing accelerometer data could contribute towards unobtrusive detection of changes in depression severity without the need for clinical input.

rest2vec: Vectorizing the resting-state functional connectome using graph embedding.

Resting-state functional magnetic resonance imaging (rs-fMRI) is widely used in connectomics for studying the functional relationships between regions of the human brain. rs-fMRI connectomics, however, has inherent analytical challenges, such as how to properly model negative correlations between BOLD time series. In addition, functional relationships between brain regions do not necessarily correspond to their anatomical distance, making the functional topology of the brain less well understood. Recent machine learning techniques, such as word2vec, have used embedding methods to map high-dimensional data into vector spaces, where words with more similar meanings are mapped closer to one another. Inspired by this approach, we have developed the graph embedding pipeline rest2vec for studying the vector space of functional connectomes. We demonstrate how rest2vec uses the phase angle spatial embedding (PhASE) method with dimensionality reduction to embed the connectome into lower dimensions, where the functional definition of a brain region is represented continuously in an intrinsic "functional space." Furthermore, we show how the "functional distance" between brain regions in this space can be applied to discover biologically-relevant connectivity gradients. Interestingly, rest2vec can be conceptualized in the context of the recently proposed maximum mean discrepancy (MMD) metric, followed by a double-centering approach seen in kernel PCA. In sum, rest2vec creates a low-dimensional representation of the rs-fMRI connectome where brain regions are mapped according to their functional relationships, giving a more informed understanding of the functional organization of the brain.

Baseline connectome modular abnormalities in the childhood phase of a longitudinal study on individuals with chromosome 22q11.2 deletion syndrome.

Occurring in at least 1 in 3,000 live births, chromosome 22q11.2 deletion syndrome (22q11DS) produces a complex phenotype that includes a constellation of medical complications such as congenital cardiac defects, immune deficiency, velopharyngeal dysfunction, and characteristic facial dysmorphic features. There is also an increased incidence of psychiatric diagnosis, especially intellectual disability and ADHD in childhood, lifelong anxiety, and a strikingly high rate of schizophrenia spectrum disorders, which occur in around 30% of adults with 22q11DS. Using innovative computational connectomics, we studied how 22q11DS affects high-level network signatures of hierarchical modularity and its intrinsic geometry in 55 children with confirmed 22q11DS and 27 Typically Developing (TD) children. Results identified 3 subgroups within our 22q11DS sample using a K-means clustering approach based on several midline structural measures-of-interests. Each subgroup exhibited distinct patterns of connectome abnormalities. Subtype 1, containing individuals with generally healthy-looking brains, exhibited no significant differences in either modularity or intrinsic geometry when compared with TD. By contrast, the more anomalous 22q11DS Subtypes 2 and 3 brains revealed significant modular differences in the right hemisphere, while Subtype 3 (the most anomalous anatomy) further exhibited significantly abnormal connectome intrinsic geometry in the form of left-right temporal disintegration. Taken together, our findings supported an overall picture of (a) anterior-posteriorly differential interlobar frontotemporal/frontoparietal dysconnectivity in Subtypes 2 and 3 and (b) differential intralobar dysconnectivity in Subtype 3. Our ongoing studies are focusing on whether these subtypes and their connnectome signatures might be valid biomarkers for predicting the degree of psychosis-proneness risk found in 22q11DS. Hum Brain Mapp 39:232-248, 2018. © 2017 Wiley Periodicals, Inc.

Fractional anisotropy derived from the diffusion tensor distribution function boosts power to detect Alzheimer's disease deficits.

PURPOSE: In diffusion MRI (dMRI), fractional anisotropy derived from the single-tensor model (FADTI ) is the most widely used metric to characterize white matter (WM) microarchitecture, despite known limitations in regions with crossing fibers. Due to time constraints when scanning patients in clinical settings, high angular resolution diffusion imaging acquisition protocols, often used to overcome these limitations, are still rare in clinical population studies. However, the tensor distribution function (TDF) may be used to model multiple underlying fibers by representing the diffusion profile as a probabilistic mixture of tensors. METHODS: We compared the ability of standard FADTI and TDF-derived FA (FATDF ), calculated from a range of dMRI angular resolutions (41, 30, 15, and 7 gradient directions), to profile WM deficits in 251 individuals from the Alzheimer's Disease Neuroimaging Initiative and to detect associations with 1) Alzheimer's disease diagnosis, 2) Clinical Dementia Rating scores, and 3) average hippocampal volume. RESULTS: Across angular resolutions and statistical tests, FATDF showed larger effect sizes than FADTI , particularly in regions preferentially affected by Alzheimer's disease, and was less susceptible to crossing fiber anomalies. CONCLUSION: The TDF "corrected" form of FA may be a more sensitive and accurate alternative to the commonly used FADTI , even in clinical quality dMRI data. Magn Reson Med 78:2322-2333, 2017. © 2017 International Society for Magnetic Resonance in Medicine.

Investigating brain community structure abnormalities in bipolar disorder using path length associated community estimation.

In this article, we present path length associated community estimation (PLACE), a comprehensive framework for studying node-level community structure. Instead of the well-known Q modularity metric, PLACE utilizes a novel metric, Ψ(PL), which measures the difference between intercommunity versus intracommunity path lengths. We compared community structures in human healthy brain networks generated using these two metrics and argued that Ψ(PL) may have theoretical advantages. PLACE consists of the following: (1) extracting community structure using top-down hierarchical binary trees, where a branch at each bifurcation denotes a collection of nodes that form a community at that level, (2) constructing and assessing mean group community structure, and (3) detecting node-level changes in community between groups. We applied PLACE and investigated the structural brain networks obtained from a sample of 25 euthymic bipolar I subjects versus 25 gender- and age-matched healthy controls. Results showed community structural differences in posterior default mode network regions, with the bipolar group exhibiting left-right decoupling.

Using a Novel Digital Go/No-Go to Dissociate Intra-subject Temporal Fluctuations in Reaction Time and Accuracy.

Impulsivity can be a risk factor for serious complications for those with mood disorders. To understand intra-individual impulsivity variability, we analyzed longitudinal data of a novel gamified digital Go/No-Go (GNG) task in a clinical sample (n=43 mood disorder participants, n=17 healthy controls) and an open-science sample (n=121, self-reported diagnoses). With repeated measurements within-subject, we disentangled two aspects of GNG: reaction time and accuracy in response inhibition (i.e., incorrect No-Go trials) with respect to diurnal and potential learning effects. Mixed-effects models showed diurnal effects in reaction time but not accuracy, with a significant effect of hour on reaction time in the clinical sample and the open-science sample. Moreover, subjects improved on their response inhibition but not reaction time. Additionally, significant interactions emerged between depression symptom severity and time-of-day in both samples, supporting that repeated administration of our GNG task can yield mood-dependent circadian rhythm-aware biomarkers of neurocognitive function.

Temporal Alterations in White Matter in An App Knock-In Mouse Model of Alzheimer's Disease.

Alzheimer's disease (AD) is the most common form of dementia and results in neurodegeneration and cognitive impairment. White matter (WM) is affected in AD and has implications for neural circuitry and cognitive function. The trajectory of these changes across age, however, is still not well understood, especially at earlier stages in life. To address this, we used the AppNL-G-F/NL-G-F knock-in (APPKI) mouse model that harbors a single copy knock-in of the human amyloid precursor protein (APP) gene with three familial AD mutations. We performed in vivo diffusion tensor imaging (DTI) to study how the structural properties of the brain change across age in the context of AD. In late age APPKI mice, we observed reduced fractional anisotropy (FA), a proxy of WM integrity, in multiple brain regions, including the hippocampus, anterior commissure (AC), neocortex, and hypothalamus. At the cellular level, we observed greater numbers of oligodendrocytes in middle age (prior to observations in DTI) in both the AC, a major interhemispheric WM tract, and the hippocampus, which is involved in memory and heavily affected in AD, prior to observations in DTI. Proteomics analysis of the hippocampus also revealed altered expression of oligodendrocyte-related proteins with age and in APPKI mice. Together, these results help to improve our understanding of the development of AD pathology with age, and imply that middle age may be an important temporal window for potential therapeutic intervention.

Unbiased tensor-based morphometry: improved robustness and sample size estimates for Alzheimer's disease clinical trials.

Various neuroimaging measures are being evaluated for tracking Alzheimer's disease (AD) progression in therapeutic trials, including measures of structural brain change based on repeated scanning of patients with magnetic resonance imaging (MRI). Methods to compute brain change must be robust to scan quality. Biases may arise if any scans are thrown out, as this can lead to the true changes being overestimated or underestimated. Here we analyzed the full MRI dataset from the first phase of Alzheimer's Disease Neuroimaging Initiative (ADNI-1) from the first phase of Alzheimer's Disease Neuroimaging Initiative (ADNI-1) and assessed several sources of bias that can arise when tracking brain changes with structural brain imaging methods, as part of a pipeline for tensor-based morphometry (TBM). In all healthy subjects who completed MRI scanning at screening, 6, 12, and 24months, brain atrophy was essentially linear with no detectable bias in longitudinal measures. In power analyses for clinical trials based on these change measures, only 39AD patients and 95 mild cognitive impairment (MCI) subjects were needed for a 24-month trial to detect a 25% reduction in the average rate of change using a two-sided test (α=0.05, power=80%). Further sample size reductions were achieved by stratifying the data into Apolipoprotein E (ApoE) ε4 carriers versus non-carriers. We show how selective data exclusion affects sample size estimates, motivating an objective comparison of different analysis techniques based on statistical power and robustness. TBM is an unbiased, robust, high-throughput imaging surrogate marker for large, multi-site neuroimaging studies and clinical trials of AD and MCI.

Angular versus spatial resolution trade-offs for diffusion imaging under time constraints.

Diffusion weighted magnetic resonance imaging (DW-MRI) are now widely used to assess brain integrity in clinical populations. The growing interest in mapping brain connectivity has made it vital to consider what scanning parameters affect the accuracy, stability, and signal-to-noise of diffusion measures. Trade-offs between scan parameters can only be optimized if their effects on various commonly-derived measures are better understood. To explore angular versus spatial resolution trade-offs in standard tensor-derived measures, and in measures that use the full angular information in diffusion signal, we scanned eight subjects twice, 2 weeks apart, using three protocols that took the same amount of time (7 min). Scans with 3.0, 2.7, 2.5 mm isotropic voxels were collected using 48, 41, and 37 diffusion-sensitized gradients to equalize scan times. A specially designed DTI phantom was also scanned with the same protocols, and different b-values. We assessed how several diffusion measures including fractional anisotropy (FA), mean diffusivity (MD), and the full 3D orientation distribution function (ODF) depended on the spatial/angular resolution and the SNR. We also created maps of stability over time in the FA, MD, ODF, skeleton FA of 14 TBSS-derived ROIs, and an information uncertainty index derived from the tensor distribution function, which models the signal using a continuous mixture of tensors. In scans of the same duration, higher angular resolution and larger voxels boosted SNR and improved stability over time. The increased partial voluming in large voxels also led to bias in estimating FA, but this was partially addressed by using "beyond-tensor" models of diffusion.

Brain growth rate abnormalities visualized in adolescents with autism.

Autism spectrum disorder is a heterogeneous disorder of brain development with wide ranging cognitive deficits. Typically diagnosed before age 3, autism spectrum disorder is behaviorally defined but patients are thought to have protracted alterations in brain maturation. With longitudinal magnetic resonance imaging (MRI), we mapped an anomalous developmental trajectory of the brains of autistic compared with those of typically developing children and adolescents. Using tensor-based morphometry, we created 3D maps visualizing regional tissue growth rates based on longitudinal brain MRI scans of 13 autistic and seven typically developing boys (mean age/interscan interval: autism 12.0 ± 2.3 years/2.9 ± 0.9 years; control 12.3 ± 2.4/2.8 ± 0.8). The typically developing boys demonstrated strong whole brain white matter growth during this period, but the autistic boys showed abnormally slowed white matter development (P = 0.03, corrected), especially in the parietal (P = 0.008), temporal (P = 0.03), and occipital lobes (P = 0.02). We also visualized abnormal overgrowth in autism in gray matter structures such as the putamen and anterior cingulate cortex. Our findings reveal aberrant growth rates in brain regions implicated in social impairment, communication deficits and repetitive behaviors in autism, suggesting that growth rate abnormalities persist into adolescence. Tensor-based morphometry revealed persisting growth rate anomalies long after diagnosis, which has implications for evaluation of therapeutic effects.

Regional brain volume differences in symptomatic and presymptomatic carriers of familial Alzheimer's disease mutations.

BACKGROUND: Mutations in the presenilin (PSEN1, PSEN2) and amyloid precursor protein (APP) genes cause familial Alzheimer's disease (FAD) in a nearly fully penetrant, autosomal dominant manner, providing a unique opportunity to study presymptomatic individuals who can be predicted to develop Alzheimer's disease (AD) with essentially 100% certainty. Using tensor-based morphometry (TBM), we examined brain volume differences between presymptomatic and symptomatic FAD mutation carriers and non-carrier (NC) relatives. METHODS: Twenty-five mutation carriers and 10 NC relatives underwent brain MRI and clinical assessment. Four mutation carriers had dementia (MUT-Dem), 12 had amnestic mild cognitive impairment (MUT-aMCI) and nine were cognitively normal (MUT-Norm). TBM brain volume maps of MUT-Norm, MUT-aMCI and MUT-Dem subjects were compared to NC subjects. RESULTS: MUT-Norm subjects exhibited significantly smaller volumes in the thalamus, caudate and putamen. MUT-aMCI subjects had smaller volumes in the thalamus, splenium and pons, but not in the caudate or putamen. MUT-Dem subjects demonstrated smaller volumes in temporal, parietal and left frontal regions. As non-demented carriers approached the expected age of dementia diagnosis, this was associated with larger ventricular and caudate volumes and a trend towards smaller temporal lobe volume. CONCLUSIONS: Cognitively intact FAD mutation carriers had lower thalamic, caudate and putamen volumes, and we found preliminary evidence for increasing caudate size during the predementia stage. These regions may be affected earliest during prodromal stages of FAD, while cortical atrophy may occur in later stages, when carriers show cognitive deficits. Further studies of this population will help us understand the progression of neurobiological changes in AD.

Patterns of brain atrophy in clinical variants of frontotemporal lobar degeneration.

BACKGROUND/AIMS: The clinical syndromes of frontotemporal lobar degeneration include behavioral variant frontotemporal dementia (bvFTD) and semantic (SV-PPA) and nonfluent variants (NF-PPA) of primary progressive aphasia. Using magnetic resonance imaging (MRI), tensor-based morphometry (TBM) was used to determine distinct patterns of atrophy between these three clinical groups. METHODS: Twenty-seven participants diagnosed with bvFTD, 16 with SV-PPA, and 19 with NF-PPA received baseline and follow-up MRI scans approximately 1 year apart. TBM was used to create three-dimensional Jacobian maps of local brain atrophy rates for individual subjects. RESULTS: Regional analyses were performed on the three-dimensional maps and direct comparisons between groups (corrected for multiple comparisons using permutation tests) revealed significantly greater frontal lobe and frontal white matter atrophy in the bvFTD relative to the SV-PPA group (p < 0.005). The SV-PPA subjects exhibited significantly greater atrophy than the bvFTD in the fusiform gyrus (p = 0.007). The NF-PPA group showed significantly more atrophy in the parietal lobes relative to both bvFTD and SV-PPA groups (p < 0.05). Percent volume change in ventromedial prefrontal cortex was significantly associated with baseline behavioral symptomatology. CONCLUSION: The bvFTD, SV-PPA, and NF-PPA groups displayed distinct patterns of progressive atrophy over a 1-year period that correspond well to the behavioral disturbances characteristic of the clinical syndromes. More specifically, the bvFTD group showed significant white matter contraction and presence of behavioral symptoms at baseline predicted significant volume loss of the ventromedial prefrontal cortex.

A commonly carried allele of the obesity-related FTO gene is associated with reduced brain volume in the healthy elderly.

A recently identified variant within the fat mass and obesity-associated (FTO) gene is carried by 46% of Western Europeans and is associated with an approximately 1.2 kg higher weight, on average, in adults and an approximately 1 cm greater waist circumference. With >1 billion overweight and 300 million obese persons worldwide, it is crucial to understand the implications of carrying this very common allele for the health of our aging population. FTO is highly expressed in the brain and elevated body mass index (BMI) is associated with brain atrophy, but it is unknown how the obesity-associated risk allele affects human brain structure. We therefore generated 3D maps of regional brain volume differences in 206 healthy elderly subjects scanned with MRI and genotyped as part of the Alzheimer's Disease Neuroimaging Initiative. We found a pattern of systematic brain volume deficits in carriers of the obesity-associated risk allele versus noncarriers. Relative to structure volumes in the mean template, FTO risk allele carriers versus noncarriers had an average brain volume difference of approximately 8% in the frontal lobes and 12% in the occipital lobes-these regions also showed significant volume deficits in subjects with higher BMI. These brain differences were not attributable to differences in cholesterol levels, hypertension, or the volume of white matter hyperintensities; which were not detectably higher in FTO risk allele carriers versus noncarriers. These brain maps reveal that a commonly carried susceptibility allele for obesity is associated with structural brain atrophy, with implications for the health of the elderly.

A Review on Smartphone Keystroke Dynamics as a Digital Biomarker for Understanding Neurocognitive Functioning.

Can digital technologies provide a passive unobtrusive means to observe and study cognition outside of the laboratory? Previously, cognitive assessments and monitoring were conducted in a laboratory or clinical setting, allowing for a cross-sectional glimpse of cognitive states. In the last decade, researchers have been utilizing technological advances and devices to explore ways of assessing cognition in the real world. We propose that the virtual keyboard of smartphones, an increasingly ubiquitous digital device, can provide the ideal conduit for passive data collection to study cognition. Passive data collection occurs without the active engagement of a participant and allows for near-continuous, objective data collection. Most importantly, this data collection can occur in the real world, capturing authentic datapoints. This method of data collection and its analyses provide a more comprehensive and potentially more suitable insight into cognitive states, as intra-individual cognitive fluctuations over time have shown to be an early manifestation of cognitive decline. We review different ways passive data, centered around keystroke dynamics, collected from smartphones, have been used to assess and evaluate cognition. We also discuss gaps in the literature where future directions of utilizing passive data can continue to provide inferences into cognition and elaborate on the importance of digital data privacy and consent.

Signed graph representation learning for functional-to-structural brain network mapping.

MRI-derived brain networks have been widely used to understand functional and structural interactions among brain regions, and factors that affect them, such as brain development and diseases. Graph mining on brain networks can facilitate the discovery of novel biomarkers for clinical phenotypes and neurodegenerative diseases. Since brain functional and structural networks describe the brain topology from different perspectives, exploring a representation that combines these cross-modality brain networks has significant clinical implications. Most current studies aim to extract a fused representation by projecting the structural network to the functional counterpart. Since the functional network is dynamic and the structural network is static, mapping a static object to a dynamic object may not be optimal. However, mapping in the opposite direction (i.e., from functional to structural networks) are suffered from the challenges introduced by negative links within signed graphs. Here, we propose a novel graph learning framework, named as Deep Signed Brain Graph Mining or DSBGM, with a signed graph encoder that, from an opposite perspective, learns the cross-modality representations by projecting the functional network to the structural counterpart. We validate our framework on clinical phenotype and neurodegenerative disease prediction tasks using two independent, publicly available datasets (HCP and OASIS). Our experimental results clearly demonstrate the advantages of our model compared to several state-of-the-art methods.

Smartphone accelerometer data as a proxy for clinical data in modeling of bipolar disorder symptom trajectory.

Being able to track and predict fluctuations in symptoms of mental health disorders such as bipolar disorder outside the clinic walls is critical for expanding access to care for the global population. To that end, we analyze a dataset of 291 individuals from a smartphone app targeted at bipolar disorder, which contains rich details about their smartphone interactions (including typing dynamics and accelerometer motion) collected everyday over several months, along with more traditional clinical features. The aim is to evaluate whether smartphone accelerometer data could serve as a proxy for traditional clinical data, either by itself or in combination with typing dynamics. Results show that accelerometer data improves the predictive performance of machine learning models by nearly 5% over those previously reported in the literature based only on clinical data and typing dynamics. This suggests it is possible to elicit essentially the same "information" about bipolar symptomology using different data sources, in a variety of settings.

A Hierarchical Graph Learning Model for Brain Network Regression Analysis.

Brain networks have attracted increasing attention due to the potential to better characterize brain dynamics and abnormalities in neurological and psychiatric conditions. Recent years have witnessed enormous successes in deep learning. Many AI algorithms, especially graph learning methods, have been proposed to analyze brain networks. An important issue for existing graph learning methods is that those models are not typically easy to interpret. In this study, we proposed an interpretable graph learning model for brain network regression analysis. We applied this new framework on the subjects from Human Connectome Project (HCP) for predicting multiple Adult Self-Report (ASR) scores. We also use one of the ASR scores as the example to demonstrate how to identify sex differences in the regression process using our model. In comparison with other state-of-the-art methods, our results clearly demonstrate the superiority of our new model in effectiveness, fairness, and transparency.

Hippocampal functional connectivity across age in an App knock-in mouse model of Alzheimer's disease.

Introduction: Alzheimer's disease (AD) is a progressive neurodegenerative disease. The early processes of AD, however, are not fully understood and likely begin years before symptoms manifest. Importantly, disruption of the default mode network, including the hippocampus, has been implicated in AD. Methods: To examine the role of functional network connectivity changes in the early stages of AD, we performed resting-state functional magnetic resonance imaging (rs-fMRI) using a mouse model harboring three familial AD mutations (App NL-G-F/NL-G-F knock-in, APPKI) in female mice in early, middle, and late age groups. The interhemispheric and intrahemispheric functional connectivity (FC) of the hippocampus was modeled across age. Results: We observed higher interhemispheric functional connectivity (FC) in the hippocampus across age. This was reduced, however, in APPKI mice in later age. Further, we observed loss of hemispheric asymmetry in FC in APPKI mice. Discussion: Together, this suggests that there are early changes in hippocampal FC prior to heavy onset of amyloid ß plaques, and which may be clinically relevant as an early biomarker of AD.

Differentiating prenatal exposure to methamphetamine and alcohol versus alcohol and not methamphetamine using tensor-based brain morphometry and discriminant analysis.

Here we investigate the effects of prenatal exposure to methamphetamine (MA) on local brain volume using magnetic resonance imaging. Because many who use MA during pregnancy also use alcohol, a known teratogen, we examined whether local brain volumes differed among 61 children (ages 5-15 years), 21 with prenatal MA exposure, 18 with concomitant prenatal alcohol exposure (the MAA group), 13 with heavy prenatal alcohol but not MA exposure (ALC group), and 27 unexposed controls. Volume reductions were observed in both exposure groups relative to controls in striatal and thalamic regions bilaterally and in right prefrontal and left occipitoparietal cortices. Striatal volume reductions were more severe in the MAA group than in the ALC group, and, within the MAA group, a negative correlation between full-scale intelligence quotient (FSIQ) scores and caudate volume was observed. Limbic structures, including the anterior and posterior cingulate, the inferior frontal gyrus (IFG), and ventral and lateral temporal lobes bilaterally, were increased in volume in both exposure groups. Furthermore, cingulate and right IFG volume increases were more pronounced in the MAA than ALC group. Discriminant function analyses using local volume measurements and FSIQ were used to predict group membership, yielding factor scores that correctly classified 72% of participants in jackknife analyses. These findings suggest that striatal and limbic structures, known to be sites of neurotoxicity in adult MA abusers, may be more vulnerable to prenatal MA exposure than alcohol exposure and that more severe striatal damage is associated with more severe cognitive deficit.

Accurate measurement of brain changes in longitudinal MRI scans using tensor-based morphometry.

This paper responds to Thompson and Holland (2011), who challenged our tensor-based morphometry (TBM) method for estimating rates of brain changes in serial MRI from 431 subjects scanned every 6 months, for 2 years. Thompson and Holland noted an unexplained jump in our atrophy rate estimates: an offset between 0 and 6 months that may bias clinical trial power calculations. We identified why this jump occurs and propose a solution. By enforcing inverse-consistency in our TBM method, the offset dropped from 1.4% to 0.28%, giving plausible anatomical trajectories. Transitivity error accounted for the minimal remaining offset. Drug trial sample size estimates with the revised TBM-derived metrics are highly competitive with other methods, though higher than previously reported sample size estimates by a factor of 1.6 to 2.4. Importantly, estimates are far below those given in the critique. To demonstrate a 25% slowing of atrophic rates with 80% power, 62 AD and 129 MCI subjects would be required for a 2-year trial, and 91 AD and 192 MCI subjects for a 1-year trial.