Longitudinal Scans Research Articles

Longitudinally consistent registration and parcellation of cortical surfaces using semi-supervised learning

Temporally consistent and accurate registration and parcellation of longitudinal cortical surfaces is of great importance in studying longitudinal morphological and functional changes of human brains. However, most existing methods are developed for registration or parcellation of a single cortical surface. When applying to longitudinal studies, these methods independently register/parcellate each surface from longitudinal scans, thus often generating longitudinally inconsistent and inaccurate results, especially in small or ambiguous cortical regions. Essentially, longitudinal cortical surface registration and parcellation are highly correlated tasks with inherently shared constraints on both spatial and temporal feature representations, which are unfortunately ignored in existing methods. To this end, we unprecedentedly propose a novel semi-supervised learning framework to exploit these inherent relationships from limited labeled data and extensive unlabeled data for more robust and consistent registration and parcellation of longitudinal cortical surfaces. Our method utilizes the spherical topology characteristic of cortical surfaces. It employs a spherical network to function as an encoder, which extracts high-level cortical features. Subsequently, we build two specialized decoders dedicated to the tasks of registration and parcellation, respectively. To extract more meaningful spatial features, we design a novel parcellation map similarity loss to utilize the relationship between registration and parcellation tasks, i.e., the parcellation map warped by the deformation field in registration should match the atlas parcellation map, thereby providing extra supervision for the registration task and augmented data for parcellation task by warping the atlas parcellation map to unlabeled surfaces. To enable temporally more consistent feature representation, we additionally enforce longitudinal consistency among longitudinal surfaces after registering them together using their concatenated features. Experiments on two longitudinal datasets of infants and adults have shown that our method achieves significant improvements on both registration/parcellation accuracy and longitudinal consistency compared to existing methods, especially in small and challenging cortical regions.

In Vivo Human Lacrimal Gland Imaging Using an Ultrasound Biomicroscopy.

In the present study, we introduce human lacrimal gland imaging using an ultrasound biomicroscopy (UBM) with a soft cover and show their findings. The representative UBM findings of palpebral lobes in seven subjects (four with non-Sjögren dry eye syndrome, one with Sjögren syndrome, and two healthy subjects) were described in this study. To prolapse the palpebral lobe, the examiner pulled the temporal part of the upper eyelid in the superotemporal direction and directed the subject to look in the inferonasal direction. We scanned the palpebral lobes longitudinally and transversely using UBM. We used an Aviso UBM with a 50 MHz linear probe and ClearScan. In UBM of two healthy subjects, the echogenicity of the lacrimal gland was lower than that of the sclera and homogeneous. But the parenchyma of a patient with Sjögren dry eye syndrome was quite inhomogeneous compared to the healthy subjects. In two patients with dry eye syndrome, we were able to observe some lobules in the parenchyma. We could find excretory ducts running parallel at the surface of the longitudinal section in some subjects. In the longitudinal UBM scan of a subject, we observed a tubular structure at a depth of 1,500 μm that was considered a blood vessel. It ran from the superonasal to the inferotemporal direction. In a subject, we observed a large cyst beneath the conjunctiva. Lacrimal gland imaging using UBM has both advantages of optical coherence tomography and sonography, and could be useful for evaluating dry eye syndrome.

Impact of serum leptin and adiponectin levels on brain infarcts in patients with mild cognitive impairment and Alzheimer's disease: a longitudinal analysis.

The adipokines leptin and adiponectin have been associated with atherosclerosis and the risk of cerebral infarcts. Pre-clinical studies, however, suggest a protective role against ischemic brain damage. In this study we analyzed the relationship between serum leptin and adiponectin levels and the onset or progression of brain infarcts in subjects with mild cognitive impairment (MCI) and Alzheimer's disease (AD). All data were extracted from the ADNI database. The final population included 566 subjects, with 58 healthy controls, 396 MCI and 112 AD. All patients with available serum leptin and adiponectin levels at baseline were selected. Demographics, neuropsychological test results, CSF biomarkers, regional brain metabolism with FDG-PET data and the number of brain infarcts on longitudinal MRI scans were extracted. Leptin levels were significantly lower in patients with MCI than controls at baseline, while adiponectin levels were not different between the groups. Multivariate logistic regression analysis at baseline for the presence of brain infarcts showed a predictive value for leptin but not for adiponectin. Multivariate longitudinal analysis showed that age was the only significant predictor of brain infarcts development at 15-year follow-up, while serum leptin and adiponectin levels did not play a role in this population. The evidence on the pathogenetic or protective role of adipokines on ischemic brain damage is mixed. In this MCI and AD population, serum leptin and adiponectin were not associated with the development of brain infarcts; therefore, these results do not support the use of adipokines as biomarkers of cerebrovascular pathology in this population.

Somatomotor-visual resting state functional connectivity increases after 2 years in the UK Biobank longitudinal cohort.

Functional magnetic resonance imaging (fMRI) and functional connectivity (FC) have been used to follow aging in both children and older adults. Robust changes have been observed in children, in which high connectivity among all brain regions changes to a more modular structure with maturation. We examine FC changes in older adults after 2 years of aging in the UK Biobank (UKB) longitudinal cohort. We process fMRI connectivity data using the Power264 atlas and then test whether the average internetwork FC changes in the 2722-subject longitudinal cohort are statistically significant using a Bonferroni-corrected -test. We also compare the ability of Power264 and UKB-provided, independent component analysis (ICA)-based FC to determine which of a longitudinal scan pair is older. Finally, we investigate cross-sectional FC changes as well as differences due to differing scanner tasks in the UKB, Philadelphia Neurodevelopmental Cohort, and Alzheimer's Disease Neuroimaging Initiative datasets. We find a 6.8% average increase in somatomotor network (SMT)-visual network (VIS) connectivity from younger to older scans (corrected ) that occurs in male, female, older subject ( years old), and younger subject ( years old) groups. Among all internetwork connections, the average SMT-VIS connectivity is the best predictor of relative scan age. Using the full FC and a training set of 2000 subjects, one is able to predict which scan is older 82.5% of the time using either the full Power264 FC or the UKB-provided ICA-based FC. We conclude that SMT-VIS connectivity increases with age in the UKB longitudinal cohort and that resting state FC increases with age in the UKB cross-sectional cohort.

Impact of amyloid and tau positivity on longitudinal brain atrophy in cognitively normal individuals.

Individuals on the preclinical Alzheimer's continuum, particularly those with both amyloid and tau positivity (A + T +), display a rapid cognitive decline and elevated disease progression risk. However, limited studies exist on brain atrophy trajectories within this continuum over extended periods. This study involved 367 ADNI participants grouped based on combinations of amyloid and tau statuses determined through cerebrospinal fluid tests. Using longitudinal MRI scans, brain atrophy was determined according to the whole brain, lateral ventricle, and hippocampal volumes and cortical thickness in AD-signature regions. Cognitive performance was evaluated with the Preclinical Alzheimer's Cognitive Composite (PACC). A generalized linear mixed-effects model was used to examine group × time interactions for these measures. In addition, progression risks to mild cognitive impairment (MCI) or dementia were compared among the groups using Cox proportional hazards models. A total of 367 participants (48 A + T + , 86 A + T - , 63 A - T + , and 170 A - T - ; mean age 73.8years, mean follow-up 5.1years, and 47.4% men) were included. For the lateral ventricle and PACC score, the A + T - and A + T + groups demonstrated statistically significantly greater volume expansion and cognitive decline over time than the A - T - group (lateral ventricle: β = 0.757 cm3/year [95% confidence interval 0.463 to 1.050], P < .001 for A + T - , and β = 0.889 cm3/year [0.523 to 1.255], P < .001 for A + T + ; PACC: β = - 0.19 /year [- 0.36 to - 0.02], P = .029 for A + T - , and β = - 0.59 /year [- 0.80 to - 0.37], P < .001 for A + T +). Notably, the A + T + group exhibited additional brain atrophy including the whole brain (β = - 2.782 cm3/year [- 4.060 to - 1.504], P < .001), hippocampus (β = - 0.057 cm3/year [- 0.085 to - 0.029], P < .001), and AD-signature regions (β = - 0.02mm/year [- 0.03 to - 0.01], P < .001). Cox proportional hazards models suggested an increased risk of progressing to MCI or dementia in the A + T + group versus the A - T - group (adjusted hazard ratio = 3.35 [1.76 to 6.39]). In cognitively normal individuals, A + T + compounds brain atrophy and cognitive deterioration, amplifying the likelihood of disease progression. Therapeutic interventions targeting A + T + individuals could be pivotal in curbing brain atrophy, cognitive decline, and disease progression.

Longitudinal CT-based finite element analyses provide objective fracture healing measures in an ovine tibia model.

Measuring the healing status of a bone fracture is important to determine the clinical care a patient receives. Implantable devices can directly and continuously assess the healing status of fracture fixation constructs, while subject-specific virtual biomechanical tests can noninvasively determine callus structural integrity at single time points. Despite their potential for objectification, both methods are not yet integrated into clinical practice with further evidence of their benefits required. This study correlated continuous data from an implantable sensor assessing healing status through implant load monitoring with computer tomography (CT) based longitudinal finite element (FE) simulations in a large animal model. Eight sheep were part of a previous preclinical study utilizing a tibial osteotomy model and equipped with such a sensor. Sensor signal was collected over several months, and CT scans were acquired at six interim time points. For each scan, two FE analyses were performed: a virtual torsional rigidity test of the bone and a model of the bone-implant construct with the sensor. The longitudinal simulation results were compared to the sensor data at corresponding time points and a cohort-specific empirical healing rule was employed. Healing status predicted by both in silico simulations correlated significantly with the sensor data at corresponding time points and correctly identified a delayed and a nonunion in the cohort. The methodology is readily translatable with the potential to be applied to further preclinical or clinical cohorts to find generalizable healing criteria. Virtual mechanical tests can objectively measure fracture healing progressing using longitudinal CT scans.

LONGITUDINAL ADAPTIVE OPTICS SCANNING LASER OPHTHALMOSCOPY REVEALS REGIONAL VARIATION IN CONE AND ROD PHOTORECEPTOR LOSS IN STARGARDT DISEASE.

To investigate the temporal sequence of changes in the photoreceptor cell mosaic in patients with Stargardt disease type 1, using adaptive optics scanning laser ophthalmoscopy. Two brothers with genetically confirmed Stargardt disease type 1 underwent comprehensive eye exams, spectral-domain optical coherence tomography, fundus autofluorescence, and adaptive optics scanning laser ophthalmoscopy imaging 3 times over the course of 28 months. Confocal images of the cones and rods were obtained from the central fovea to 10° inferiorly. Photoreceptors were counted in sampling windows at 100- µ m intervals of 200 µ m × 200 µ m for cones and 50 µ m × 50 µ m for rods, using custom cell marking software with manual correction. Photoreceptor density and spacing were measured and compared across imaging sessions using one-way analysis of variance. Adaptive optics scanning laser ophthalmoscopy revealed the younger brother had a 30% decline in foveal cone density after 8 months, followed by complete loss of foveal cones at 28 months; the older brother had no detectable foveal cones at baseline. In the peripheral macula, cone and rod spacings were greater than normal in both patients. The ratio of the cone spacing to rod spacing was greater than normal across all eccentricities, with a greater divergence closer to the foveal center. Cone cell loss may be an early pathogenetic step in Stargardt disease. Adaptive optics scanning laser ophthalmoscopy provides the capability to track individual photoreceptor changes longitudinally in Stargardt disease.

Pregnancy renders anatomical changes in hypothalamic substructures of the human brain that relate to aspects of maternal behavior

Animal studies have shown that pregnancy is associated with neural adaptations that promote maternal care. The hypothalamus represents a central structure of the mammalian maternal brain and hormonal priming of specific hypothalamic nuclei plays a key role in the induction and expression of maternal behavior. In humans, we have previously demonstrated that becoming a mother involves changes in grey matter anatomy, primarily in association areas of the cerebral cortex. In the current study, we investigated whether pregnancy renders anatomical changes in the hypothalamus. Using an advanced delineation technique, five hypothalamic substructures were defined in longitudinal MRI scans of 107 women extracted from two prospective pre-conception cohort studies, including 50 women who were scanned before and after pregnancy and 57 nulliparous control women scanned at a similar time interval. We showed that becoming a mother is associated with volume reductions in the anterior-superior, superior tuberal and posterior hypothalamus. In addition, these structural changes related to hormonal levels during pregnancy and specific aspects of self-reported maternal behavior in late pregnancy, including maternal-fetal attachment and nesting behavior. These findings show that pregnancy leads to changes in hypothalamic anatomy and suggest that these contribute to the development of maternal behavior in humans, supporting the conservation of key aspects of maternal brain circuitry and their role in maternal behavior across species.

High resolution dynamic ultrasound atlas of embryonic and fetal development of the common marmoset

PurposeThe common marmoset (Callithrix jacchus) provides an ideal model to study early development of primates, and an in vivo platform to validate conclusions from in vitro studies of human embryos and embryo models. Currently, however, no established staging atlas of marmoset embryonic development exists. Using high-resolution, longitudinal ultrasound scans on live pregnant marmosets, we present the first dynamic in vivo imaging of entire primate gestation beginning with attachment until the last day before birth.MethodsOur study unveils the first dynamic images of an in vivo attached mammalian embryo developing in utero, and the intricacies of the delayed development period unique to the common marmoset amongst primates, revealing a window for somatic interventions.ResultsEstablished obstetric and embryologic measurements for each scan were used comparatively with the standardized Carnegie staging of human development to highlight similarities and differences. Our study also allows for tracking the development of major organs. We focus on the ontogeny of the primate heart and brain. Finally, input ultrasound images were used to train deep neural networks to accurately determine the gestational age. All our ultrasounds and staging data recording are posted online so that the atlas can be used as a community resource toward monitoring and managing marmoset breeding colonies.ConclusionThe temporal and spatial resolution of ultrasound achieved in this study demonstrates the promise of noninvasive imaging in the marmoset for the in vivo study of primate-specific aspects of embryonic and fetal development.

Wasserstein HOG: Local Directionality Extraction via Optimal Transport.

Directionally sensitive radiomic features including the histogram of oriented gradient (HOG) have been shown to provide objective and quantitative measures for predicting disease outcomes in multiple cancers. However, radiomic features are sensitive to imaging variabilities including acquisition differences, imaging artifacts and noise, making them impractical for using in the clinic to inform patient care. We treat the problem of extracting robust local directionality features by mapping via optimal transport a given local image patch to an iso-intense patch of its mean. We decompose the transport map into sub-work costs each transporting in different directions. To test our approach, we evaluated the ability of the proposed approach to quantify tumor heterogeneity from magnetic resonance imaging (MRI) scans of brain glioblastoma multiforme, computed tomography (CT) scans of head and neck squamous cell carcinoma as well as longitudinal CT scans in lung cancer patients treated with immunotherapy. By considering the entropy difference of the extracted local directionality within tumor regions, we found that patients with higher entropy in their images, had significantly worse overall survival for all three datasets, which indicates that tumors that have images exhibiting flows in many directions may be more malignant. This may seem to reflect high tumor histologic grade or disorganization. Furthermore, by comparing the changes in entropy longitudinally using two imaging time points, we found patients with reduction in entropy from baseline CT are associated with longer overall survival (hazard ratio = 1.95, 95% confidence interval of 1.4-2.8, p = 1.65e-5). The proposed method provides a robust, training free approach to quantify the local directionality contained in images.

Similarities in the Electrographic Patterns of Delayed Cerebral Infarction and Brain Death After Aneurysmal and Traumatic Subarachnoid Hemorrhage.

While subarachnoid hemorrhage is the second most common hemorrhagic stroke in epidemiologic studies, the recent DISCHARGE-1 trial has shown that in reality, three-quarters of focal brain damage after subarachnoid hemorrhage is ischemic. Two-fifths of these ischemic infarctions occur early and three-fifths are delayed. The vast majority are cortical infarcts whose pathomorphology corresponds to anemic infarcts. Therefore, we propose in this review that subarachnoid hemorrhage as an ischemic-hemorrhagic stroke is rather a third, separate entity in addition to purely ischemic or hemorrhagic strokes. Cumulative focal brain damage, determined by neuroimaging after the first 2weeks, is the strongest known predictor of patient outcome half a year after the initial hemorrhage. Because of the unique ability to implant neuromonitoring probes at the brain surface before stroke onset and to perform longitudinal MRI scans before and after stroke, delayed cerebral ischemia is currently the stroke variant in humans whose pathophysiological details are by far the best characterized. Optoelectrodes located directly over newly developing delayed infarcts have shown that, as mechanistic correlates of infarct development, spreading depolarizations trigger (1) spreading ischemia, (2) severe hypoxia, (3) persistent activity depression, and (4) transition from clustered spreading depolarizations to a negative ultraslow potential. Furthermore, traumatic brain injury and subarachnoid hemorrhage are the second and third most common etiologies of brain death during continued systemic circulation. Here, we use examples to illustrate that although the pathophysiological cascades associated with brain death are global, they closely resemble the local cascades associated with the development of delayed cerebral infarcts.

Development and Validation of Deep Learning-Based Automated Detection of Cervical Lymphadenopathy in Patients with Lymphoma for Treatment Response Assessment: A Bi-institutional Feasibility Study.

The purpose is to train and evaluate a deep learning (DL) model for the accurate detection and segmentation of abnormal cervical lymph nodes (LN) on head and neck contrast-enhanced CT scans in patients diagnosed with lymphoma and evaluate the clinical utility of the DL model in response assessment. This retrospective study included patients who underwent CT for abnormal cervical LN and lymphoma assessment between January 2021 and July 2022. Patients were grouped into the development (n = 76), internal test 1 (n = 27), internal test 2 (n = 87), and external test (n = 26) cohorts. A 3D SegResNet model was used to train the CT images. The volume change rates of cervical LN across longitudinal CT scans were compared among patients with different treatment outcomes (stable, response, and progression). Dice similarity coefficient (DSC) and the Bland-Altman plot were used to assess the model's segmentation performance and reliability, respectively. No significant differences in baseline clinical characteristics were found across cohorts (age, P = 0.55; sex, P = 0.13; diagnoses, P = 0.06). The mean DSC was 0.39 ± 0.2 with a precision and recall of 60.9% and 57.0%, respectively. Most LN volumes were within the limits of agreement on the Bland-Altman plot. The volume change rates among the three groups differed significantly (progression (n = 74), 342.2%; response (n = 8), - 79.2%; stable (n = 5), - 8.1%; all P < 0.01). Our proposed DL segmentation model showed modest performance in quantifying the cervical LN burden on CT in patients with lymphoma. Longitudinal changes in cervical LN volume, as predicted by the DL model, were useful for treatment response assessment.

Early left ventricular microvascular dysfunction in diabetic pigs: a longitudinal quantitative myocardial perfusion CMR study

BackgroundMicrovascular pathology is one of the main characteristics of diabetic cardiomyopathy; however, the early longitudinal course of diabetic microvascular dysfunction remains uncertain. This study aimed to investigate the early dynamic changes in left ventricular (LV) microvascular function in diabetic pig model using the cardiac magnetic resonance (CMR)-derived quantitative perfusion technique.MethodsTwelve pigs with streptozotocin-induced diabetes mellitus (DM) were included in this study, and longitudinal CMR scanning was performed before and 2, 6, 10, and 16 months after diabetic modeling. CMR-derived semiquantitative parameters (upslope, maximal signal intensity, perfusion index, and myocardial perfusion reserve index [MPRI]) and fully quantitative perfusion parameters (myocardial blood flow [MBF] and myocardial perfusion reserve [MPR]) were analyzed to evaluate longitudinal changes in LV myocardial microvascular function. Pearson correlation was used to analyze the relationship between LV structure and function and myocardial perfusion function.ResultsWith the progression of DM duration, the upslope at rest showed a gradually increasing trend (P = 0.029); however, the upslope at stress and MBF did not change significantly (P > 0.05). Regarding perfusion reserve function, both MPRI and MPR showed a decreasing trend with the progression of disease duration (MPRI, P = 0.001; MPR, P = 0.042), with high consistency (r = 0.551, P < 0.001). Furthermore, LV MPR is moderately associated with LV longitudinal strain (r = − 0.353, P = 0.022), LV remodeling index (r = − 0.312, P = 0.033), fasting blood glucose (r = − 0.313, P = 0.043), and HbA1c (r = − 0.309, P = 0.046). Microscopically, pathological results showed that collagen volume fraction increased gradually, whereas no significant decrease in microvascular density was observed with the progression of DM duration.ConclusionsMyocardial microvascular reserve function decreased gradually in the early stage of DM, which is related to both structural (but not reduced microvascular density) and functional abnormalities of microvessels, and is associated with increased blood glucose, reduced LV deformation, and myocardial remodeling.

Effect of variable left ventricular ejection fraction assessed by equilibrium radionuclide angiocardiography using different software packages on the diagnosis of cardiotoxicity in patients with cancer

BackgroundThe equilibrium radionuclide angiocardiography (ERNA) scan is an established imaging modality for assessing left ventricular ejection fraction (LVEF) in oncology patients. This study aimed to explore the interchangeability of two commercially available software packages (MIM and JS) for LVEF measurement for a cancer-therapy-related cardiac dysfunction (CTRCD) diagnosis. MethodsThis is a single-center retrospective study among 322 patients who underwent ERNA scans. A total of 582 scans were re-processed using MIM and JS for cross-sectional and longitudinal LVEF measurements. ResultsThe median LVEF for MIM and JS were 56% and 66%, respectively (P < 0.001). LVEF processed by JS was 9.91% higher than by MIM. In 87 patients with longitudinal ERNA scans, serial studies processed by MIM were classified as having CTRCD in a higher proportion than serial studies processed by JS (26.4% vs 11.4%, P = 0.020). There were no significant differences in intra- or inter-observer LVEF measurement variability (R = 0.99, P < 0.001). ConclusionsSoftware packages for processing ERNA studies are not interchangeable; thus, reports of ERNA studies should include details on the post-processing software. Serial ERNA studies should be processed on the same software when feasible to avoid discrepancies in the diagnosis and management of CTRCD.

3DTINC: Time-Equivariant Non-Contrastive Learning for Predicting Disease Progression from Longitudinal OCTs.

Self-supervised learning (SSL) has emerged as a powerful technique for improving the efficiency and effectiveness of deep learning models. Contrastive methods are a prominent family of SSL that extract similar representations of two augmented views of an image while pushing away others in the representation space as negatives. However, the state-of-the-art contrastive methods require large batch sizes and augmentations designed for natural images that are impractical for 3D medical images. To address these limitations, we propose a new longitudinal SSL method, 3DTINC, based on non-contrastive learning. It is designed to learn perturbation-invariant features for 3D optical coherence tomography (OCT) volumes, using augmentations specifically designed for OCT. We introduce a new non-contrastive similarity loss term that learns temporal information implicitly from intra-patient scans acquired at different times. Our experiments show that this temporal information is crucial for predicting progression of retinal diseases, such as age-related macular degeneration (AMD). After pretraining with 3DTINC, we evaluated the learned representations and the prognostic models on two large-scale longitudinal datasets of retinal OCTs where we predict the conversion to wet-AMD within a six-month interval. Our results demonstrate that each component of our contributions is crucial for learning meaningful representations useful in predicting disease progression from longitudinal volumetric scans.

Prognostic enrichment for early-stage Huntington’s disease: An explainable machine learning approach for clinical trial

BackgroundIn Huntington’s disease clinical trials, recruitment and stratification approaches primarily rely on genetic load, cognitive and motor assessment scores. They focus less on in vivo brain imaging markers, which reflect neuropathology well before clinical diagnosis. Machine learning methods offer a degree of sophistication which could significantly improve prognosis and stratification by leveraging multimodal biomarkers from large datasets. Such models specifically tailored to HD gene expansion carriers could further enhance the efficacy of the stratification process. ObjectivesTo improve stratification of Huntington’s disease individuals for clinical trials. MethodsWe used data from 451 gene positive individuals with Huntington’s disease (both premanifest and diagnosed) from previously published cohorts (PREDICT, TRACK, TrackON, and IMAGE). We applied whole-brain parcellation to longitudinal brain scans and measured the rate of lateral ventricular enlargement, over 3 years, which was used as the target variable for our prognostic random forest regression models. The models were trained on various combinations of features at baseline, including genetic load, cognitive and motor assessment score biomarkers, as well as brain imaging-derived features. Furthermore, a simplified stratification model was developed to classify individuals into two homogenous groups (low risk and high risk) based on their anticipated rate of ventricular enlargement. ResultsThe predictive accuracy of the prognostic models substantially improved by integrating brain imaging features alongside genetic load, cognitive and motor biomarkers: a 24 % reduction in the cross-validated mean absolute error, yielding an error of 530 mm3/year. The stratification model had a cross-validated accuracy of 81 % in differentiating between moderate and fast progressors (precision = 83 %, recall = 80 %). ConclusionsThis study validated the effectiveness of machine learning in differentiating between low- and high-risk individuals based on the rate of ventricular enlargement. The models were exclusively trained using features from HD individuals, which offers a more disease-specific, simplified, and accurate approach for prognostic enrichment compared to relying on features extracted from healthy control groups, as done in previous studies. The proposed method has the potential to enhance clinical utility by: i) enabling more targeted recruitment of individuals for clinical trials, ii) improving post-hoc evaluation of individuals, and iii) ultimately leading to better outcomes for individuals through personalized treatment selection.

Chemically induced cone degeneration in the 13-lined ground squirrel.

Animal models of retinal degeneration are critical for understanding disease and testing potential therapies. Inducing degeneration commonly involves the administration of chemicals that kill photoreceptors by disrupting metabolic pathways, signaling pathways, or protein synthesis. While chemically induced degeneration has been demonstrated in a variety of animals (mice, rats, rabbits, felines, 13-lined ground squirrels (13-LGS), pigs, chicks), few studies have used noninvasive high-resolution retinal imaging to monitor the in vivo cellular effects. Here, we used longitudinal scanning light ophthalmoscopy (SLO), optical coherence tomography, and adaptive optics SLO imaging in the euthermic, cone-dominant 13-LGS (46 animals, 52 eyes) to examine retinal structure following intravitreal injections of chemicals, which were previously shown to induce photoreceptor degeneration, throughout the active season of 2019 and 2020. We found that iodoacetic acid induced severe pan-retinal damage in all but one eye, which received the lowest concentration. While sodium nitroprusside successfully induced degeneration of the outer retinal layers, the results were variable, and damage was also observed in 50% of contralateral control eyes. Adenosine triphosphate and tunicamycin induced outer retinal specific damage with varying results, while eyes injected with thapsigargin did not show signs of degeneration. Given the variability of damage we observed, follow-up studies examining the possible physiological origins of this variability are critical. These additional studies should further advance the utility of chemically induced photoreceptor degeneration models in the cone-dominant 13-LGS.

Toward hippocampal volume measures on ultra-high field magnetic resonance imaging: a comprehensive comparison study between deep learning and conventional approaches.

The hippocampus is a complex brain structure that plays an important role in various cognitive aspects such as memory, intelligence, executive function, and path integration. The volume of this highly plastic structure is identified as one of the most important biomarkers of specific neuropsychiatric and neurodegenerative diseases. It has also been extensively investigated in numerous aging studies. However, recent studies on aging show that the performance of conventional approaches in measuring the hippocampal volume is still far from satisfactory, especially in terms of delivering longitudinal measures from ultra-high field magnetic resonance images (MRIs), which can visualize more boundary details. The advancement of deep learning provides an alternative solution to measuring the hippocampal volume. In this work, we comprehensively compared a deep learning pipeline based on nnU-Net with several conventional approaches including Freesurfer, FSL and DARTEL, for automatically delivering hippocampal volumes: (1) Firstly, we evaluated the segmentation accuracy and precision on a public dataset through cross-validation. Results showed that the deep learning pipeline had the lowest mean (L = 1.5%, R = 1.7%) and the lowest standard deviation (L = 5.2%, R = 6.2%) in terms of volume percentage error. (2) Secondly, sub-millimeter MRIs of a group of healthy adults with test-retest 3T and 7T sessions were used to extensively assess the test-retest reliability. Results showed that the deep learning pipeline achieved very high intraclass correlation coefficients (L = 0.990, R = 0.986 for 7T; L = 0.985, R = 0.983 for 3T) and very small volume percentage differences (L = 1.2%, R = 0.9% for 7T; L = 1.3%, R = 1.3% for 3T). (3) Thirdly, a Bayesian linear mixed effect model was constructed with respect to the hippocampal volumes of two healthy adult datasets with longitudinal 7T scans and one disease-related longitudinal dataset. It was found that the deep learning pipeline detected both the subtle and disease-related changes over time with high sensitivity as well as the mild differences across subjects. Comparison results from the aforementioned three aspects showed that the deep learning pipeline significantly outperformed the conventional approaches by large margins. Results also showed that the deep learning pipeline can better accommodate longitudinal analysis purposes.

Characterization of dynamic patterns of human fetal to neonatal brain asymmetry with deformation-based morphometry.

Despite established knowledge on the morphological and functional asymmetries in the human brain, the understanding of how brain asymmetry patterns change during late fetal to neonatal life remains incomplete. The goal of this study was to characterize the dynamic patterns of inter-hemispheric brain asymmetry over this critically important developmental stage using longitudinally acquired MRI scans. Super-resolution reconstructed T2-weighted MRI of 20 neurotypically developing participants were used, and for each participant fetal and neonatal MRI was acquired. To quantify brain morphological changes, deformation-based morphometry (DBM) on the longitudinal MRI scans was utilized. Two registration frameworks were evaluated and used in our study: (A) fetal to neonatal image registration and (B) registration through a mid-time template. Developmental changes of cerebral asymmetry were characterized as (A) the inter-hemispheric differences of the Jacobian determinant (JD) of fetal to neonatal morphometry change and the (B) time-dependent change of the JD capturing left-right differences at fetal or neonatal time points. Left-right and fetal-neonatal differences were statistically tested using multivariate linear models, corrected for participants' age and sex and using threshold-free cluster enhancement. Fetal to neonatal morphometry changes demonstrated asymmetry in the temporal pole, and left-right asymmetry differences between fetal and neonatal timepoints revealed temporal changes in the temporal pole, likely to go from right dominant in fetal to a bilateral morphology in neonatal timepoint. Furthermore, the analysis revealed right-dominant subcortical gray matter in neonates and three clusters of increased JD values in the left hemisphere from fetal to neonatal timepoints. While these findings provide evidence that morphological asymmetry gradually emerges during development, discrepancies between registration frameworks require careful considerations when using DBM for longitudinal data of early brain development.

Data‐driven neuroanatomical subtype trajectories of primary progressive aphasia

AbstractBackgroundPrimary progressive aphasia (PPA) is an atypical neurodegenerative dementia with three main clinical variants: semantic (svPPA), non‐fluent/agrammatic (nfvPPA) and logopenic (lvPPA). While svPPA is typically associated with left anterior temporal lobe atrophy, neuroimaging findings in nfvPPA, lvPPA and PPA not otherwise specified (PPA‐nos) are more variable. We therefore applied unsupervised machine learning to one of the largest databases of PPA magnetic resonance imaging (MRI) scans ever acquired.MethodBaseline MRI scans were collected from 273 people with PPA participating in the longitudinal research programme in the Queen Square Dementia Research Centre (svPPA n = 97; nfvPPA n = 109; lvPPA n = 51; PPA‐nos n = 16). Longitudinal scans were collected from 138 participants. We used Subtype and Stage Inference (SuStaIn), a model that combines disease progression modelling with unsupervised machine learning to identify subgroups of patients following a similar trajectory. We standardised the data with respect to cognitively normal controls (n = 360). Based on clinicians’ input we chose 19 brain regions of interest to model from those generated by GIF (Cardoso et al. 2012). The model was fit using baseline data, selecting the appropriate number of subtypes using cross‐validation (Young et al. 2018).ResultCross‐validation supported four subtypes of spatiotemporal atrophy (Figure 1). Clinical variant representation per subtype assignment (1/2/3/4) was: svPPA (n = 74/15/4/4), nfvPPA (n = 3/35/21/44), lvPPA (n = 1/18/30/2) and PPA‐nos (n = 7/4/3/2) (Figure 2). Subtype 1 showed early decline in the left anterior temporal lobe and sparing of the left frontal lobe. Subtype 2 showed early decline of the left insula and sparing of the temporal pole. Subtype 3 showed initial atrophy of the left temporoparietal junction; Subtype 4 showed early atrophy in the left precentral gyrus. Statistical analysis of posterior distributions found low similarity between subtypes (Hellinger distance &gt; 0.66) (Figure 3), validating the presence of four distinct subtypes.ConclusionWhile the data‐driven atrophy subtypes did not align perfectly with clinical variants, we found a correspondence between Subtype 1 and svPPA, the most coherent PPA variant. Subtypes 2,3 and 4 were associated with nfvPPA and lvPPA, corroborating work suggesting these diagnostic terms encompass multiple clinical phenotypes. Future work will investigate the association between clinical diagnosis, longitudinal subtype, and clinical measures.