Neuropathology and Applied Neurobiology

Foundation models (FMs) have rapidly become dominant in artificial intelligence and are increasingly being adopted in computational pathology. Numerous pathology-specific FMs have been developed and evaluated for a variety of downstream tasks, most commonly using frozen image embeddings with linear probes. While a pathology FM, at least implicitly suggests broad reusability across tasks, the transferability of these models to neurodegenerative disease-related tasks remains largely unexplored. In this study, we evaluated fourteen frozen feature extractors, including general-vision and pathology FMs, across four pathology image classification datasets spanning two specialized neurodegenerative disease immuno-histochemistry (IHC) tasks (Tau neurofibrillary tangle (NFT) and amyloid-{beta} plaque classification) and two cancer-related hematoxylin and eosin (H&E) tasks (the breast tissue BACH dataset and the multi-tissue TIL dataset). The cancer-related H&E datasets represent tasks more closely aligned with the predominant pretraining domain of many current pathology FMs, whereas the neurodegenerative IHC datasets represent a more specialized domain with limited representation in existing FM pretraining cohorts. Frozen embedding linear probes were compared against a conventional supervised ResNet-50 convolutional neural network (CNN) trained directly on image tiles using identical train, validation, and test splits. Across both neurodegenerative IHC datasets, the supervised CNN substantially outperformed all frozen FM linear probes. In contrast, pathology FMs achieved performance comparable to, and in some cases exceeding, the supervised CNN across the two cancer-related H&E datasets, with UNI2-h achieving the highest performance on BACH and several pathology FMs performing on par with the CNN on the larger TIL dataset. Furthermore, a supervised CNN trained using only 1% of the Tau NFT training data (1,938 tiles) still exceeded the performance of the best frozen FM linear probe trained on the complete dataset. Together, these results suggest the transferability of frozen pathology FMs may depend strongly on how well the downstream task is represented by their pretraining domain. These findings demonstrate frozen pathology FMs transfer effectively to the cancer-related H&E tasks evaluated here but may be less effective for specialized neurodegenerative IHC tasks less well represented in current FM pretraining cohorts. Expanding pathology FM pretraining datasets to include a broader range of disease domains and staining modalities may therefore improve transferability to specialized pathology applications. Our findings also highlight the continued importance of conventional supervised learning and motivate future work investigating nonlinear probes and end-to-end FM fine-tuning.

DNA methylation-based classification is increasingly a cornerstone of CNS tumor diagnostics. Recently,nanopore sequencing enables very rapid methylation profiling to support surgical decision making. Time-constrained sequencing yields sparse methylation profiles, necessitating specialized classifiers for accurate predictions. All publicly available sparse-profile classifiers rely on a single reference set of only 2,801 samples published in 2018 and its accompanying classification system. Here we present Sturgeon V2, trained on nearly 10,000 methylation profiles assembled from a range of public repositories and collaborations. To harmonize differences in tumor class labels, we developed an iterative semi-supervised relabeling strategy with a two-tiered concordance filter. We demonstrate that this filtering is essential for optimal performance. Compared to Sturgeon V1, we further redesigned the classifier architecture to provide more stable, and consistent confidence estimates while accumulating sequencing data and we introduce an updated class system including new classes, such as meningioma subtypes, PATZ- fusion neuroepithelial tumors and PLAG1/2-amplified embryonal tumors. In a retrospective intraoperative cohort of 94 pediatric cases, Sturgeon V2 was confidently correct in 85 cases (Sturgeon V1: 78) with one confident misclassification, and in simulated sequencing runs (n=1880) reduced the error rate among confident calls from 7.1% to 2.8%, outperforming publicly available alternatives. Sturgeon V2 is the first openly available, locally executable classifier for sparse methylation profiles trained beyond the 2018 reference era.

Tissue-based assays provide broad screening for neural autoantibodies but remain vulnerable to technical and reader-dependent variability. We evaluated an integrated workflow combining automated 3,3'-diaminobenzidine tissue-based assay processing, digital whole-slide imaging, and interpretable deep learning. Model development used 2,357 retrospective serum and cerebrospinal fluid images referred from more than 80 institutions and processed in a centralized laboratory, with reference labels established by expert-consensus tissue-based assay interpretation. A prospective validation cohort included 200 images from 200 unique patients (154 positive and 46 negative; 100 serum and 100 cerebrospinal fluid). The same images were evaluated by two expert and two novice readers in a fixed-order two-round study, with unaided interpretation followed by artificial-intelligence-assisted interpretation after a washout of at least two weeks. The standalone binary classifier achieved an area under the receiver operating characteristic curve of 0.922, accuracy of 86.5%, sensitivity of 85.7%, and specificity of 89.1% in prospective validation. Pooled reader accuracy was 90.4% in the unaided round and 97.4% in the artificial-intelligence-assisted round (McNemar P < 0.001), with specificity of 82.6% and 99.5%, respectively. Novice-reader accuracy was 83.3% unaided and 96.5% with artificial-intelligence assistance (P < 0.001), approaching unaided expert accuracy of 97.5%. Among tissue-based-assay-positive cases, reader accuracy for cytoplasmic/nuclear antigen-associated versus neural surface antigen-associated staining-pattern subclassification was 88.8% unaided and 94.0% with artificial-intelligence assistance (P < 0.001). These findings support further multicentre, cross-platform evaluation of artificial-intelligence-assisted neural autoantibody tissue interpretation.

Neurologic autoantibody cell-based assay images are difficult to interpret when fluorescence is weak, focal or dominated by background signal. We developed a visual artificial intelligence overlay that transfers detected cell boundaries from the structural channel to the antibody-staining image and evaluated standalone model concordance and its association with reader classifications. The analysis used 2 763 records for model development and 499 held-out records spanning 14 antibody targets. Four readers assessed the held-out records first without assistance and then with the boundary overlay and model category available. The selected model agreed with the reader-derived three-class reference in 85.2% of records, but recall was lowest for weak-positive or equivocal records. Reader agreement increased in the assisted session, and almost all revised ratings moved towards the model category. The combined display may support consistent visual review while also creating a risk of model-directed anchoring; its role as a second-read aid requires independent validation.

Frontotemporal lobar degeneration (FTLD) encompasses clinically and biologically heterogeneous syndromes that are unlikely to be captured by a single blood biomarker. Although plasma neurofilament light (NfL) chain reflects neuroaxonal injury, multiplex profiling may identify complementary inflammatory, vascular and proteostatic processes relevant to clinical phenotype and progression. We characterised plasma CNS-related and inflammatory proteomic signatures across FTLD-related clinical syndromes and tested their associations with neuroaxonal injury, clinical severity, regional brain atrophy and survival. Plasma samples from 177 participants with behavioural variant frontotemporal dementia (bvFTD), primary progressive aphasia (PPA), progressive supranuclear palsy (PSP) or corticobasal syndrome (CBS) and 67 healthy controls were analysed using the high-sensitivity nucleic acid-linked immuno-sandwich assay (NULISA) platform. We quantified 130 CNS-related and 250 inflammation-related proteins. Differential expression and multivariate classification analyses were complemented by hierarchical clustering, penalised regression and survival analyses. Canonical correlation analysis tested multivariate associations between disease-associated proteins and regional MRI brain volumes. Sixty-one CNS-related and 61 inflammation-related proteins differed between participants with FTLD-related syndromes and controls after false-discovery-rate correction. NfL was the strongest individual discriminator between all patients and controls, with an area under the receiver operating characteristic curve of 0.91. In exploratory full-cohort models, the multivariate CNS signature achieved an area under the curve of 0.93, compared with 0.82 for the multivariate inflammation signature. However, inflammatory proteins provided stronger discrimination for selected syndrome comparisons, including PPA versus PSP. Among inflammation-related proteins, SPP1 and CD276 showed the strongest positive associations with NfL. Higher CD276 and SPP1 were associated with increased mortality, whereas higher amyloid-{beta}42 was associated with lower mortality. A multivariate protein signature weighted towards tumour necrosis factor and immune signalling was associated with lower regional brain volumes across predominantly frontotemporal and limbic regions (canonical correlation 0.43; permutation p=0.018). Plasma multiplex proteomics identified complementary dimensions of neuroaxonal injury and immune dysregulation across frontotemporal lobar degeneration-related syndromes. CNS-related proteins, particularly NfL, predominantly captured case-control neurodegeneration, whereas inflammatory signatures provided additional information about syndrome heterogeneity, structural brain changes and survival. SPP1, CD276 and broader immune-related signatures are candidates for independent biomarker validation and mechanistic investigation.

Spatial transcriptomics approaches provide crucial insights into gene expression distribution within intact tissue architecture, but they encounter limitations in detecting morphologically complex cell types, assessing their spatial associations with pathology, and accurately annotating cell types using RNA data alone. Therefore, we developed a robust post-processing workflow integrating Xenium spatial technology with iterative indirect immunofluorescence imaging (4i) on formalin-fixed paraffin-embedded (FFPE) human brain tissue. The post-Xenium 4i protocol presented here allows multi-omic tissue mapping, enabling deeper investigation of pathological microenvironments defined by the spatial distribution of neuropathological hallmarks and enrichment of specific cell populations. We applied this workflow on calcarine cortex tissue sections where cerebral amyloid angiopathy (CAA) burden is present, in addition to amyloid plaques and tau pathology, generating a 15-plex image that captures the complexity of the pathological microenvironment.

Current state of the art computational pathology foundation models attain satisfying accuracy most general pathology tasks including detection of pancreatic ductal adenocarcinoma (PDAC) but fail to provide sufficient explainability, operating as "black boxes". Some interpretability effort has been made, with models providing attention maps that project back on the slide the models high attended areas that pushed for the given prediction result, without further naming or reasoning explanation which can limit pathologists trust toward the prediction. We developed C3PRO, a cell-centric framework attempting to provide deeper understanding and explainability, by using embeddings extracted using a pathology foundation model, from cell centered tiles containing both the cell and its close neighborhood, which were then clustered into recurrent morphological phenotypes. A spatial graph was additionally constructed on top of the back-projected vocabulary to account for recurrent motifs. The evaluation of the model performance involved a cohort of 809 patients (825 diagnostic biopsies) using nested cross validation. C3PRO was benchmarked against four different multiple instance learning frameworks and achieved on-par performance for diagnostic (0.958 AUC), survival (0.617) and resectability (0.696) prediction, while providing additional insights compared to the strongest baseline models. Each clustered phenotype was probed and extracted as a representative tile mosaic and annotated by two expert pathologists, constituting a phenotype vocabulary available for statistical analysis. These phenotypes were scored by pathologists with physical brightfield microscopy and replicated the independent statistical significance impact on overall survival prediction from diagnostic biopsies.

Amyotrophic lateral sclerosis (ALS) is a heterogeneous neurodegenerative disorder lacking established biomarkers of underlying disease biology. Here, we identify lipidomic signatures of circulating extracellular vesicles (EVs) that distinguish ALS from neurologically healthy controls and related neuromuscular diseases. Using mass spectrometry-based profiling and multivariate modelling, we separate ALS (n=77) from healthy controls (n=89) with an area under the receiver operating characteristic curve (AUROC) of 0.90 (95% CI 0.85-0.94), rising to 0.92 (0.89-0.96) when low-confidence predictions are withheld, and retaining performance in an independent cohort (n=119; AUROC 0.78, 0.70-0.86). In Tofersen-treated SOD1-ALS patients, biomarker trajectories change in parallel with clinical measures while individuals with spinal muscular atrophy (n=50) exhibit a lipid fingerprint distinct from that observed in ALS, supporting disease specificity within a shared systemic framework of abnormal systemic lipogenesis. These findings establish EV-associated lipid signatures as a systemic, blood-based biomarker platform for ALS diagnosis, stratification, and treatment monitoring, with structurally distinct lipid species indicating widespread membrane remodelling.

Frontotemporal lobar degeneration (FTLD), predominantly comprised of tau and TDP-43 proteinopathies, is a common cause of early-onset dementia. Biomarkers that reliably classify FTLD subtypes are limited, thereby hindering in-vivo diagnosis and clinical trial stratification. We conducted large-scale, multi-platform cerebrospinal fluid proteomics in 566 individuals, including 274 FTLD-spectrum (48 autopsy-confirmed FTLD-TDP, 90 autopsy-confirmed FTLD-tau), 171 Alzheimers disease, and 95 controls. In neuropathological subtype contrasts, elevated matrix metalloproteinase MMP10 classified Picks disease (PiD) versus other subtypes with near-perfect accuracy across proteomic platforms. We further validated CSF MMP10 as a potential PiD diagnostic biomarker via a targeted immunoassay in an independent, autopsy-confirmed FTLD cohort. Post-synaptic proteins HOMER1 and DLG4 exhibited elevations across FTLD-TDP subtypes (A, B, and C) versus FTLD-tau. Network analysis identified RNA metabolism and pro-inflammatory signatures in FTLD-TDP, and cytoskeletal and heat shock signatures in progressive supranuclear palsy. These findings reveal candidate diagnostic biomarkers and proteomic pathways underlying FTLD clinico-pathological heterogeneity.

Accurate classification of lung adenocarcinoma growth patterns from hematoxylin and eosin (H\&E)-stained whole-slide images is essential for treatment planning and prognosis. We develop and evaluate an exponentially weighted ensemble deep learning approach that combines five architectures (EfficientNet-B3, DeiT3, Swin Transformer, Vision Transformer, ConvNeXt) with weights determined by a softmax transformation of per-model Quadratic Weighted Kappa (QWK) scores. Using 25,545 image patches at 40x magnification from the Dartmouth-Hitchcock dataset, we evaluated performance with stratified 5-fold cross-validation. Individual model performance ranged from Cohen's kappa kappa = 0.5563 (ConvNeXt) to kappa = 0.9561 (EfficientNet-B3). The ensemble achieved 96.52% accuracy and kappa = 0.9648, exceeding the best individual model by delta kappa = +0.0087 with statistical significance (McNemar's test, chi^2 = 50.84, p < 10^-12; bootstrap 95% CI on delta kappa [+0.0056, +0.0117]. Pairwise disagreement rates of 16.6--43.2% between architectures indicate that the models capture different aspects of tissue morphology, and the framework automatically down-weights unreliable components, as illustrated by ConvNeXt receiving only 5.7% ensemble weight despite a single-fold training failure. The approach offers a clinically grounded method for ensemble aggregation in digital pathology.

Background: To determine the crude and age-adjusted point prevalence of motor neuron disease (MND) and its phenotypes (classical amyotrophic lateral sclerosis [ALS], primary lateral sclerosis [PLS] and progressive muscular atrophy [PMA]) using a comprehensive multi-source case ascertainment strategy. Methods: Cross-sectional population-based study with a prevalence date of 1 January 2025. Cases were identified through five complementary regional data sources and confirmed by clinical review against the revised El Escorial and Gold Coast criteria. Cases were classified by phenotype as ALS, PLS or PMA. The reference population was the population aged 15 years or older in Valencian Community (n=4,433,999). Crude prevalence was calculated per 100,000 with 95% Poisson confidence intervals; age-adjusted rates were obtained by direct standardisation to the 2013 European Standard Population. Results: A total of 374 confirmed MND cases were identified. The overall crude prevalence of MND was 8.43 per 100,000 (95% CI: 7.60-9.29) and the age-adjusted prevalence was 7.85 per 100,000 (95% CI: 7.05-8.65). By phenotype, ALS accounted for 325 cases (86.9%; crude prevalence 7.33 per 100,000), PLS for 26 (7.0%; 0.59 per 100,000) and PMA for 23 (6.1%; 0.52 per 100,000). Prevalence was consistent across provinces. Mean age was 65.9 years, the male-to-female ratio was 1.37:1, and 83.4% were receiving riluzole. Conclusions: This is the first comprehensive, multi-source population-based estimate of MND prevalence and its subtypes in the Valencian Community. The estimate is higher than the previous single-unit figure and consistent with recent European data, underscoring the value of multi-source ascertainment.

Importance: Blood biomarkers for Alzheimer disease(AD), Parkinson disease(PD), and amyotrophic lateral sclerosis(ALS) are typically developed as disease-specific single-analyte tests. Whether one shared plasma panel can support disease-selective classification remains unclear. Objective: To test whether the Olink Target 48 Neurodegeneration plasma panel, analyzed with disease-specific Elastic Net models, can discriminate AD, PD, and ALS in a heterogeneous neurological cohort. Design, Setting, and Participants: Single-center diagnostic cohort study at Keio University Hospital, Tokyo, Japan. After quality control and one consent withdrawal, 302 participants were analyzed(healthy controls n = 62; AD continuum n = 93; PD n = 39; ALS n = 27; other neurological disorders n = 81) and allocated by diagnosis-stratified sampling into a discovery training set(n = 211) and a locked hold-out test set(n = 91). Exposures: Plasma normalized protein expression from 42 assays retained after protein quality control. Plasma DOPA decarboxylase was excluded a priori from all classifiers because blood levels primarily reflect dopaminergic treatment exposure. Main Outcomes and Measures: Hold-out receiver operating characteristic area under the curve(AUC). Sensitivity and specificity at a training-set Youden threshold locked for Test. Results: Hold-out AUCs were 0.974(95% CI, 0.942-0.995) for AD, 0.926(0.820-1.000) for PD, and 0.995(0.980-1.000) for ALS. At locked Youden cutoffs, sensitivity was 0.929 (95% CI, 0.765-0.991) and specificity was 0.921 (95% CI, 0.824-0.974) for AD, 0.833 (95% CI, 0.516-0.979) and 0.975 (95% CI, 0.912-0.997) for PD, and 1.000 (95% CI, 0.631-1.000) and 0.940 (95% CI, 0.865-0.980) for ALS. Conclusions and Relevance: A single plasma neurodegeneration panel supported disease-selective classification of AD, PD, and ALS in internal validation. Multicenter external validation is needed before clinical application.

Background and Objectives Huntington's disease is a rare neurodegenerative disease in which optimizing clinical trial design is essential. Although baseline characteristics are known to predict disease progression, the statistical efficiency gains achievable by formally integrating prognostic predictions into trial analyses have not been systematically quantified. We aimed to benchmark prognostic modeling approaches, quantify the trial efficiency gains from prognostic covariate adjustment, and deliver a ready-to-deploy framework for immediate implementation. Methods: We developed a two-step modeling pipeline using longitudinal Enroll-HD natural history data (NCT01574053) to train nine machine learning algorithms to predict disease progression across five UHDRS endpoints. For this, we used 16 predictive baseline covariates representing stage of disease (baseline score levels), disease specific (CAP, CAG length), demographics and co-medication use that could be quantified in one clinical visit. For Total Function Capacity (TFC), frequently used as a primary endpoint in clinical trials, we also devised a minimal prognostic model based on 6 covariates. Model performance was tested using the placebo arm of the GENERATION HD1 phase 3 trial (NCT03761849). Results: Models were trained on 16,398 Enroll-HD participants and tested on 260 participants from the GENERATION HD1 placebo arm. Models that take into account nonlinear relationships between the covariates showed no performance advantage over linear-based models on the test set. For TFC, a minimal six-covariate linear prognostic model performed on par with the full 16-feature model. Adjusting for the prognostic score increased the signal-to-noise ratio across different trial visits by 8-18%, raising statistical power from 80% to up to 90% at a fixed sample size, or, equivalently, reducing required sample size by 15-28% compared to unadjusted analyses (no covariate adjustment). Compared to routinely used covariate adjustments (including baseline score, age, CAG repeat length, and CAP score), adjusting for a prognostic score consistently delivered meaningful additional efficiency gains. Discussion: Prognostic covariate adjustment using a simple six-covariate linear score derived from a single baseline visit yields meaningful gains in statistical power for HD clinical trials, while remaining fully embedded within standard, regulator-accepted analysis frameworks. This approach offers a practical and immediately deployable tool to improve trial design.

Charcot Marie Tooth type 1A (CMT1A) is the most common inherited peripheral neuropathy, caused by duplication of the peripheral myelin protein 22 (PMP22) gene and characterized by dysmyelination and defects in myelin maturation. Hypercellularity is a prominent feature of the disease, although its origin, persistence, and underlying mechanisms remain controversial. Here, we performed a longitudinal analysis of sciatic nerves from wild-type (WT) and CMT1A rats between 2 and 12 months of age. We demonstrated that CMT1A nerves exhibit persistent cellular proliferation extending into adulthood. This phenotype is confirmed by bulk RNA-seq data showing sustained enrichment of proliferation-associated gene signatures. Cell type signature analysis identifies overrepresentation of immature, neural crest related, and epithelial mesenchymal transitional (EMT) transcriptional programs in CMT1A nerves. Among the most upregulated genes, RUNX2 showed strong and prolonged expression at both transcript and protein levels. Consistently with the key role of RUNX2 in the activation of the injury response, CMT1A nerves also showed a marked enrichment in SOX10+JUN+ cells, that only partially overlap with SOX10+RUNX2+ cells, indicating a maladaptive activation of the injury-like Schwann cell (SC) program. Interestingly, we detected robust upregulation of SMOC2, an extracellular matrix protein selectively expressed by endoneurial fibroblasts, pointing to activation of non-canonical cell types and altered extracellular matrix remodeling. Together, these data demonstrate that CMT1A peripheral nerves are characterized by chronic hypercellularity driven by persistent proliferation, EMT, and dysregulated injury transcriptional programs, collectively pointing at progressive and multicellular nerve remodeling as a relevant pathomechanism in the disease.

Maximizing the morphological and molecular fidelity of preserved mammalian brain tissue is essential for basic neuroscience and brain banking, where tissue quality determines the reliability of downstream analyses. Aldehyde fixation and cryogenic storage are the most powerful preservation techniques available. However, how to best combine them is incompletely characterized. Here, we tested twelve different approaches to preserve the murine brain, including standard aldehyde perfusion fixation, variants of aldehyde-stabilized cryopreservation (ASC) using ethylene glycol (EG), aldehyde fixation followed by unprotected freezing, and several cryopreservation protocols without preceding aldehyde fixation, such as interleaved equilibration with vitrification solution. Preservation quality was assessed using light microscopy, transmission electron microscopy, and patch-level feature analysis with the DINOv2 vision foundation model. We found that ASC with sodium dodecyl sulfate (SDS)-mediated blood-brain barrier permeabilization preserved nuclear morphology, myelin periodicity, and neuropil texture comparable to standard aldehyde perfusion fixation, providing the first independent replication of ASC. Omitting SDS caused severe parenchymal dehydration, due to a mismatch between water and cryoprotectant transport. Aldehyde fixation followed by unprotected slow freezing confined ice damage primarily to perivascular zones, while fast freezing led to intranuclear clefts and cavities. Among non-fixation protocols, interleaved equilibration with vitrification solution preserved myelinated axon profiles and chromatin patterns, but induced perivascular edema. High concentration glycerol perfusion caused significant osmotic dehydration. Straight freezing of unfixed tissue without cryoprotectant was followed by membrane disruption, but resulted in the visualization of more structure than expected, likely as a result of structural restitution upon thawing. Taken together, our results provide a framework for matching preservation strategies to the needs of different types of brain banking and neuroscience research.

Leukodystrophies (LKD) comprise mostly inherited white matter disorders whose developmental origins and underlying pathomechanisms remain poorly understood. Progress has been limited by restricted access to human primary tissue and by the inability of animal models to fully recapitulate human neurodevelopment. Human induced pluripotent stem cell-derived neural organoids provide an alternative approach, but most culture systems do not capture later, glia-rich developmental stages relevant to these diseases. Recent studies have shown that appropriate patterning cues and long in vitro differentiation can generate neuroglial spheroids comprising multiple neuronal and macroglial populations, including mature oligodendrocytes. Here, we adapt the human oligodendrocyte (hOLS) spheroid platform to establish 3D models of globoid cell leukodystrophy (GLD), 4H leukodystrophy (4H), and Canavan disease (CD). Here we adapt this organoid platform to establish 3D models of globoid cell leukodystrophy (GLD), 4H leukodystrophy (4H), and Canavan disease (CD). Across 150 days of differentiation, hOLS from different genetic backgrounds displayed largely comparable growth kinetics and developmental trajectories, comprising multiple neuroglial cell types with predominantly ventral forebrain and midbrain identities. Comparison of LKD and control hOLS revealed both shared and disease-specific alterations, with prominent transcriptional changes in lysosomal, metabolic, and stress-response pathways. Lineage-resolved pathway analysis further identified signatures across individual LKDs, including mitochondrial and translational stress in GLD glial lineages, disruption of RNA metabolic and cell-cycle pathways in 4H, and prominent astrocyte-centered metabolic dysregulation in CD. Our results establish hOLS as a human 3D in vitro platform for studying the impact of LKD-associated genetic variants on early neural development and for dissecting lineage-specific disease vulnerability.

As we have entered the endemic phase of COVID-19, there is a need to assess the long-term effects of SARS-CoV-2 infection on the trajectory of neurodegenerative disorders. Herein, we investigate whether the SARS-CoV-2 RNA-binding nucleocapsid protein (NCP; N-protein) may influence molecular processes relevant to motor neuron degeneration in amyotrophic lateral sclerosis (ALS). We detected N-protein immunoreactivity in ALS spinal cord motor neurons, including cytoplasmic N-protein-positive puncta that colocalized with cytoplasmic TAR DNA-binding protein 43 (TDP-43). In vitro, optogenetically induced N-protein biomolecular condensates (BMCs) showed reversible, liquid-like behavior and recovered after photobleaching. Using this model, we found that these condensates colocalized with TDP-43 and the RNA-binding proteins G3BP1/2, PABPC1 and TIA1. Condensate formation was associated with TDP-43 cytoplasmic mislocalization, altered TARDBP expression, and altered exon 3 splicing of POLDIP3. The N-terminal portion of N-protein influenced BMC formation, whereas truncated C-terminal proteoforms did not form detectable BMCs or show detectable TDP-43 colocalization in this assay. Extending these observations to a human cell model, we show that optogenetically induced N-protein BMCs colocalize with TDP-43 in human iPSC-derived motor neurons. Collectively, these findings support further investigation into potential molecular links between SARS-CoV-2 exposure and the trajectory of neurodegenerative disease.

Spinal muscular atrophy (SMA) is usually considered a lower motor neuron disease; however, growing evidence from both experimental models and clinical studies indicates that SMN deficiency affects the whole CNS, beyond the spinal and brainstem motor system. Here, we investigated the temporal progression and selective vulnerability of cortical projection neurons in the SMN{Delta}7 mouse model at early (P5) and late (P11) symptomatic stages. Combining subtype-specific immunohistochemistry, DiI tracing associated with morphometric reconstruction, and dendritic spine analysis, we identified stage- and subtype-dependent cortical alterations. At P11, corticospinal and infragranular callosal projection neurons were selectively reduced in number and showed marked dendritic remodeling and altered excitatory synaptic organization, whereas corticothalamic neurons and supragranular SATB2-positive populations were comparatively preserved. Notably, cortical abnormalities were already detectable at P5, before overt neuronal loss: indeed, corticospinal neurons exhibited reduced soma size and dendritic complexity, while dendritic spine density remained preserved, but spine morphology was shifted toward immature phenotypes. These early maturation defects persisted during disease progression and preceded the reduction in spine density observed at P11. Together, our findings identify cortical projection neurons as selectively vulnerable to constitutive SMN deficiency and suggest that impaired neuronal and synaptic maturation represents an early feature of cortical pathology preceding subtype-specific neurodegeneration. These results expand the pathophysiological spectrum of SMA beyond lower motor neurons and highlight early brain alterations as a potential component of disease progression.

Dementia is a major global health challenge; however, current diagnostic approaches often rely on invasive or costly procedures. Non-invasive molecular approaches using Raman spectroscopy and sweat-derived extracellular vesicles (EVs) may provide an accessible strategy for characterizing dementia-associated molecular changes. To determine whether Raman spectral profiling of sweat EVs can discriminate individuals with dementia from cognitively healthy controls (HC) and distinguish between Alzheimer disease (AD) and frontotemporal dementia (FTD) including the assessment of a sex-specific spectral signatures. This proof-of-concept cross-sectional study included 88 participants: 26 HC, 37 patients with AD, and 25 patients FTD, comprising female and male participants. Sweat EVs were isolated from sweat patches and characterized using Raman spectroscopy. Differences in Raman spectral signatures between groups and diagnostic performance assessed by receiver operating characteristic analysis. Distinct Raman spectral fingerprints were identified in sweat EVs from dementia patients compared with HC. Participant-level Raman classification discriminated dementia patients from HC with an AUC of 0.76 in the full cohort and 0.83 in a balanced sensitivity analysis. Distinct spectral differences were also observed between AD and FTD, achieving modest classification performance of 0.73 AUC. Sex-stratified analyses further revealed Raman spectral heterogeneity between diagnostic groups, highlighting additional molecular differences that were not observed in the overall cohort. These findings provide proof-of-concept evidence that Raman spectroscopy of sweat EVs can identify dementia-associated molecular signatures potentially related to differences in protein, lipid, and metabolic composition. Furthermore, Raman profiling showed modest ability to differentiate AD from FTD. As a non-invasive approach, our strategy may represent a promising platform for molecular characterization of dementia, warranting validation in larger longitudinal cohorts.

Cardiac amyloidosis is caused by deposits of misfolded proteins in the heart and can lead to heart failure. Histological confirmation typically involves staining tissue sections with Congo red and examining them under polarized light for apple-green birefringence; however, achieving an interpretable, label-free, and observer-independent assessment remains challenging. Here we show that Pearson kurtosis of channel intensities within the amide I window provides an interpretable, fitting-free contrast for cardiac amyloid in unstained human cardiac tissue sections from ten participants imaged by ultrabroadband multiplex coherent anti-Stokes Raman scattering (CARS). In a discovery participant, high-kurtosis regions corresponded with Congo red-positive tissue in two CARS-histology pairs of adjacent sections (descriptive spatial AUCs, 0.79 and 0.73). Across 15 fields from nine additional participants, field-level kurtosis values did not overlap between severe-deposition and Congo red-negative tissue fields. Among the descriptors examined, kurtosis showed the largest standardized participant-level separation. Exploratory multivariate curve resolution yielded a deposit-associated component with aromatic-residue features shared with reagent transthyretin. In one field from the discovery participant, the second harmonic generation (SHG) signal, acquired simultaneously with CARS, showed a broadly inverse spatial pattern relative to the amide I kurtosis map, revealing distinct tissue-architectural contrast. Together, these findings support a potential label-free approach to detecting cardiac amyloid deposits that combines amide I kurtosis derived from broadband CARS with SHG-derived architectural contrast.