Neuropathology and Applied Neurobiology

Amyotrophic Lateral Sclerosis (ALS) is a phenotypically diverse neurodegenerative disorder characterized by the degeneration of motor neurons, ultimately resulting in loss of motor function. Currently, patients are classified based on clinical factors which does not fully capture the heterogeneity. Using the Olink Explore HT, we pro-filed ~5400 plasma proteins in ALS patients (n = 235) from the ALSrisc Study at Karolinska Institutet, Sweden, and applied consensus clustering, which revealed three patient subgroups significantly differing in survival outcome. Weighted gene correlation network analysis (WGCNA) and gene ontology (GO) analysis identified protein modules related to biological functions including cellular stress response and Rho-GTPase activity for the shorter surviving patient group. Cell-type enrichment analysis indicating a broad cellular origin of the circulating proteome, with contributions from nervous system cells, vascular cells, immune cells, and muscle cells. Our work highlights the potential of plasma proteomics for clinically meaningful patient stratification in ALS.

Myasthenia gravis (MG) is an autoimmune disease caused by pathogenic autoantibodies against proteins at the neuromuscular junction (NMJ). The diagnosis and clinical management of MG patients largely relies on the detection of antigen-specific autoantibodies targeting acetylcholine receptor (AChR) or muscle-specific kinase (MuSK). Yet a subset of patients remains seronegative for known MG autoantibodies, highlighting a critical need for alternative approaches to identify pathogenic NMJ antibodies. We established a new human in vitro model of the NMJ based on primary human muscle cells that recapitulates key features of the NMJ: differentiation to myotubes, expression of key NMJ proteins and formation of postsynaptic AChR clusters in response to agrin stimulation. The model allows new insights into myogenesis and genetic muscle diseases, and the new muscle cell-based assay (CBA) detected autoantibodies in sera from patients with AChR- and MuSK-positive MG with 96.43% sensitivity and 100% specificity, while healthy control sera showed no reactivity. Incubation with patient sera significantly reduced AChR clustering compared to controls, demonstrating functional pathogenic effects. Thus, we established a physiologically relevant human NMJ model that enables detection and functional characterization of neuromuscular autoantibodies. This novel approach addresses a key limitation of current antigen-specific diagnostics and provides a method for improved detection and characterization of MG antibodies, independent of antigen specificity. One Sentence SummaryWe established a postsynaptic human in vitro neuromuscular junction model to assess binding and pathogenicity of MG autoantibodies. Key messagesO_ST_ABSWhat is already known on this topic?C_ST_ABSCurrent diagnosis of myasthenia gravis (MG) relies largely on the detection of antigen-specific autoantibodies against AChR and MuSK, leaving a clinically relevant subset of patients seronegative. What are the new findings?We established a physiologically relevant human in vitro neuromuscular junction model based on primary human muscle cells and developed a novel muscle cell-based assay (CBA) for the detection of neuromuscular autoantibodies. How might this impact on clinical practice or future developments?The CBA detected autoantibodies in patients with AChR- or MuSK-positive MG with high sensitivity and specificity and demonstrated their functional pathogenic effects on AChR clustering. This antigen-independent approach may improve the detection and functional characterization of MG autoantibodies, particularly in patients who are seronegative in current diagnostic assays. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=130 SRC="FIGDIR/small/743478v1_ufig1.gif" ALT="Figure 1000"> View larger version (38K): org.highwire.dtl.DTLVardef@18ed154org.highwire.dtl.DTLVardef@151036corg.highwire.dtl.DTLVardef@1b7ab34org.highwire.dtl.DTLVardef@1490fe9_HPS_FORMAT_FIGEXP M_FIG C_FIG

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disorder traditionally characterized by motor neuron degeneration, but emerging evidence indicates sensory system involvement. Despite reports of sensory abnormalities in some patients, the molecular and functional alterations in dorsal root ganglion (DRG) neurons remain insufficiently characterized. We investigated DRG pathology at disease onset in 12-week-old SOD1G93A mice using an integrated transcriptomic, morphological, and electrophysiological approach. RNA sequencing of lumbar DRG identified 35 differentially expressed genes, predominantly upregulated, enriched in oxidative stress-related and phagosome pathways. Comparative analysis with motor neuron transcriptomes revealed distinct gene expression profiles, indicating sensory neuron-specific molecular responses. Immunohistochemistry demonstrated reduced soma diameter in both A- and C-fiber DRG neurons. Nav channel colocalization increased for Nav1.7 in A fibers and Nav1.8 in both fiber types, whereas Nav1.6 was unchanged. Whole-cell patch-clamp recordings showed depolarized resting membrane potential, increased spike amplitude, and enhanced repetitive firing in A-fiber neurons, consistent with hyperexcitability, while C fibers showed no significant functional changes. These findings demonstrate early molecular, structural, and functional alterations in primary sensory neurons in ALS, supporting pathology beyond motor neurons and identifying sensory neuron excitability as a potential therapeutic target.

Gross brain findings help guide differential diagnosis, but their diagnostic value across major neurodegenerative diseases remains incompletely characterized. We evaluated whether gross descriptions in autopsy reports could classify cases by final neuropathologic diagnosis. We analyzed 5,613 autopsy cases from the Mayo Clinic Brain Bank collected between 1998 and 2023, including seven diagnostic categories: Alzheimer disease (AD), Lewy body disease (LBD), combined AD and LBD (AD-LBD), frontotemporal lobar degeneration, corticobasal degeneration, progressive supranuclear palsy (PSP), and multiple system atrophy (MSA). A fine-tuned large language model (LLM) converted narrative gross descriptions into semi-quantitative scores for 39 features. Extraction accuracy was 0.95 in 200 manually annotated feature-level test examples. A CatBoost classifier used the extracted scores as input, whereas a second fine-tuned LLM used standardized textual descriptions of the same features. Both classifiers also included age at death, sex, and brain weight. Performance was evaluated in the same held-out test set of 562 cases. CatBoost achieved an accuracy of 0.73, a Cohen's kappa of 0.65, and a macro-average area under the receiver operating characteristic curve of 0.92. The text-based LLM achieved an accuracy of 0.75 and a kappa of 0.68. Macro-average sensitivity was 0.66 for both models. PSP sensitivity was 0.92 with CatBoost and 0.93 with text-based LLM. Corresponding sensitivities were 0.87 and 0.92 for MSA, but only 0.21 and 0.03 for AD-LBD. Feature attribution analysis identified subthalamic nucleus atrophy and putaminal abnormalities as contributors to PSP and MSA predictions, respectively. Narrative gross descriptions contained diagnostic information, particularly for disorders with distinctive macroscopic patterns, whereas combined AD-LBD remained difficult to distinguish.

Background and Objectives: Facioscapulohumeral muscular dystrophy (FSHD) is characterised by asynchronous muscle degeneration and marked pathological heterogeneity. While conventional MRI can detect fatty replacement and oedema-like abnormalities, it provides limited insight into other relevant biological processes such as extracellular matrix remodelling and tissue perfusion. This study investigated whether multispectral optoacoustic tomography (MSOT), an emerging imaging technique that non-invasively interrogates tissue composition through endogenous optical absorbers, can provide complementary molecular information on muscle involvement in FSHD. Methods: MSOT was performed on 568 muscles from 15 genetically confirmed FSHD patients and 9 healthy volunteers using a standardized protocol. The analysis included single-wavelength signal intensity and spectrally unmixed signals related to lipid, collagen, deoxygenated haemoglobin, oxygenated haemoglobin, total haemoglobin, and tissue oxygen saturation. Patients also underwent conventional muscle MRI together with clinical strength and disease severity assessments. Feasibility, reproducibility, discrimination between patients and controls, and associations with MRI and clinical measures were evaluated. Results: MSOT acquisition was well tolerated and highly reproducible (intra- and inter-rater ICC = 0.88). FSHD muscles displayed a distinct optoacoustic profile characterised by reduced signal intensity at wavelengths below 900 nm, increased lipid signal and reduced oxygenated and total haemoglobin signals. Lipid signal discriminated FSHD from control muscles (AUC 0.830, p = 0.008) and increased with MRI-defined fatty replacement (p < 0.0001). Short tau inversion recovery (STIR)-positive muscles on MRI exhibited a distinct molecular phenotype, with higher collagen-associated and lipid signals together with increased deoxygenated and total haemoglobin signals, consistent with active tissue remodelling. Although FSHD muscles classified as normal by conventional imaging showed group-level optoacoustic profiles comparable to controls, 13 of 141 (9%) demonstrated at least one abnormal MSOT-derived parameter. Reduced haemoglobin-related signals correlated with greater global disease severity, whereas increased lipid signal was associated with lower muscle strength. Discussion: MSOT is a feasible and reproducible imaging technique in FSHD that provides complementary molecular information beyond conventional structural imaging. Its ability to detect signatures related to fatty replacement, STIR positivity, functional impairment, and isolated abnormalities in radiologically normal muscles supports further longitudinal studies of MSOT-derived readouts as potential biomarkers of disease activity, progression, and therapeutic response.

Familial recurrence of amyotrophic lateral sclerosis (ALS) or frontotemporal dementia (FTD) is the clinical cornerstone for assuming Mendelian, monogenic inheritance which informs genetic testing, counselling and gene-discovery strategies in standard practice. Polygenic inheritance can lead to familial clustering of ALS, while pathogenic variants with reduced penetrance may present sporadically in small, modern-day pedigrees. As both rare genetic variants and polygenic disease contribute to the occurrence of ALS, the accuracy of a positive family history to predict the underlying genetic architecture remains unknown. Using growing knowledge of the genetic architecture of ALS, we designed the APECS simulator: a large-scale pedigree simulator using historical demographic parameters, unifying frameworks for monogenic and polygenic disease models. We simulated 1,000,000 three-generation pedigrees and modelled ALS, FTD, and other dementias under both inheritance models. For index patients with ALS, we examined the performance of proposed familial ALS (fALS) definitions to predict the underlying genetic architecture. Simulation results were benchmarked against real-world population estimates, over a wide range of demographic and genetic parameters. We assessed the phenocopy rate for monogenic index patients, defined as relatives with polygenic ALS or a pathogenic allele from a different ancestor. Finally, we illustrated the value of distant relatedness among apparently sporadic, but monogenic ALS patients. Simulations generated 55.3 million individuals, yielding 5,431 ALS-affected index patients (16.4% monogenic). The proposed criterion for 'definite fALS' (2 or more relatives with ALS among relatives within two degrees) had a 31.1% sensitivity, while falsely classifying 10.0% of 'fALS' as monogenic. The sensitivity of 'definite fALS' was improved to 43.3% when also considering relatives with FTD. Considering any type of dementia raised the false positive rate to 35.9%. In total, 5.3% of monogenic index patients had a phenocopy with ALS in relatives up to the third degree. Of seemingly sporadic monogenic index patients, 83.1% had distant relatives with ALS/FTD up to the fifth degree. A family history provides limited accuracy for inferring genetic architecture in ALS. Considering unspecified dementias in the family history reduces this accuracy, while adding definite FTD improves accuracy. Relying on traditional fALS criteria risks both missing cases for genetic testing and treatments, and falsely alarming families with polygenic ALS. Given the limited value of a family history and risk of phenocopies, routine genetic testing regardless of family history and systematic approaches for detecting distant relatedness will help to find known and novel genetic causes in ALS.

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.

BackgroundSensitive biomarkers that objectively stage Huntington disease (HD) are needed to improve participant stratification and facilitate the enrichment of clinical trials with biologically and clinically homogeneous populations. The HDClarity study, an international longitudinal biofluid collection initiative for HD, provides a unique resource for large-scale proteomic profiling of matched CSF and serum samples spanning the disease spectrum. Here, we leveraged baseline proteomic data from HDClarity to characterize protein signatures associated with HD stage and clinical severity, compare measurements across analytical platforms and biofluid compartments, and identify candidate multi-protein panels for disease staging. MethodsBaseline proteomic data generated using Olink Explore ([~]3,000 proteins) and SomaScan v4.1 ([~]7,000 proteins) were analyzed in matched CSF and serum samples from 315 HD gene-expansion carriers and 92 non-HD controls. A total of 2,119 proteins overlapped between Olink and SomaScan, enabling assessment of cross-platform concordance, while CSF-serum relationships were evaluated using all available protein measurements within each assay. Covariate-adjusted linear regression models were used to assess disease stage-associated differences in protein abundance, while partial correlation analyses evaluated relationships between protein abundance, clinical severity in HD gene-expansion carriers, and estimated years to disease onset in premanifest participants. A nested machine-learning pipeline incorporating univariate feature ranking, penalized regression-based feature selection, and repeated cross- validation was used to derive compact multi-protein classifiers for HD staging. ResultsCross-platform and CSF-serum correlations were highly protein-dependent, with some analytes showing strong concordance and others exhibiting weak or inverse relationships. These findings highlight substantial heterogeneity in biomarker behaviour across analytical platforms and biofluids. Adjusted models identified both known HD-associated markers (NEFL, GFAP, CHI3L1) and less well-characterized proteins in CSF and serum whose baseline abundance differed across HD-Integrated Staging System (HD-ISS) and clinical stages. Partial correlation analyses revealed additional candidate biomarkers associated with clinical severity and estimated time to disease onset. Machine-learning models derived compact CSF and serum protein panels that accurately classified participants across HD-ISS stages 0 and 1, as well as the transition from premanifest to early manifest disease. ConclusionsThis study provides the first large-scale orthogonal comparison of matched CSF and serum proteomes in HDClarity, establishing robust baseline proteomic signatures across the HD continuum. Our findings demonstrate the importance of considering both analytical platform and biofluid when interpreting protein biomarkers and identify compact protein panels with potential utility for objective disease staging, patient stratification, and clinical trial enrichment in HD. Trial RegistrationNot applicable. One Sentence SummaryCaron et al. analyzed matched baseline CSF and serum proteomic data from the HDClarity study generated using two orthogonal proteomic platforms, identifying reproducible multi-protein panels capable of staging and stratifying Huntington disease.

Chronic Wasting Disease (CWD) is a fatal transmissible spongiform encephalopathy (TSE) that is confined to cervids (deer, moose, elk, and reindeer) but shares key properties with human neurodegenerative conditions such as Alzheimers, Parkinsons, Huntingtons disease and frontal-temporal dementia. CWD and other TSEs are caused by the misfolded prion protein (PrP). Although the identification of diagnostic and prognostic biomarkers at all stages of disease progression is becoming exceedingly critical as CWD continues to increase in prevalence, accurate antemortem testing techniques are extremely limited. This study made use of cervidized transgenic mice (mice carrying the cervid PrP) that recapitulate CWD in various disease stages and investigated the utility of neurological biomarkers and neurobehavioral manifestations for CWD detection. Neurofilament light chain (NFL), glial fibrillary acidic protein (GFAP), and total Tau (t-Tau) were assessed under the hypothesis that combined biomarker signatures might more reliably reflect CWD-related neurodegeneration and disease progression. Analyses at 90, 132, 174, and 230 days post-CWD inoculation show distinct biomarker elevation, with all three biomarkers significantly elevated in the CWD animals by 132 days post-inoculation. To our knowledge, this is the first demonstration that these three plasma biomarkers are useful not only for detecting CWD, but also for identifying it at early antemortem stages of disease. Novel phenotypes were also revealed by comprehensive phenotypic profiling, including rigid tail elevation, increased grip strength, and impaired coordination, to lend further support to plasma biomarker data indicating neurologic impairment associated with brain pathology. Ultimately, the goal is to improve antemortem, non-invasive CWD detection methods to enable earlier detection and assist with disease management.

Background: Amyotrophic lateral sclerosis (ALS) is a rare neurodegenerative disease with no known cure. Disease progression in people living with ALS is heterogeneous, hindering personalized treatment development. The current gold standard for measuring disease progression in ALS, the ALS Functional Rating Scale - Revised (ALSFRS-R), is widely used but based on subjective measurements. Blood-based neurofilament light (NfL) has been studied as a diagnostic and prognostic biomarker but less information exists on its utility as a disease progression biomarker. Methods: We present results from blood draws of 300 participants in the FDA-funded Clinic-Based Multi-Site ALS Natural History and Biofluid study of the ALS Natural History Consortium (NHC). Plasma NfL levels were measured and analyzed against different disease progression metrics based on the ALSFRS-R. Results: NfL levels were found to be correlated with the ALSFRS-R average rate of change (r=-0.53, 95% CI -0.62 to -0.42). This association differed at a cutoff value of 61 pg/mL, with stronger correlations below this cutoff (r=-0.51 vs r=-0.18). Survival differed stratifying by this cutoff value, with participants under the cutoff having higher survival probabilities. The predictive value of NfL when predicting time to death was higher compared with the first ALSFRS-R across different event horizons. A model including both was better when predicting events up to 2 years after diagnosis. Conclusions: These results highlight the utility of NfL as a disease progression biomarker in ALS alongside ALSFRS-R based disease progression metrics. The cutoff value can aid in clinical trial stratification, pragmatic trial planning, and clinical care.

Amyotrophic lateral sclerosis (ALS) is a motor neuron disease, leading to progressive muscle weakness and motor impairment. Growing evidence indicates that cerebellar Purkinje cells, which play a central role in motor coordination, are also affected in ALS. However, it is unclear whether the molecular events that initiate neurodegeneration in these ALS-relevant motor-controlling neurons are shared or distinct. Here, we used a MATR3 S85C knock-in (KI) mouse model of early-stage ALS with stage-specific motor phenotypes and selective vulnerability of motor neurons and Purkinje cells to decipher the molecular events underlying neurodegeneration in these two neuronal populations. We found that a profound reduction in detectable MATR3 S85C immunoreactivity (hereafter referred to as MATR3 loss) in both motor neurons and Purkinje cells precedes the onset of motor dysfunction and neuropathology, implicating MATR3 loss as the earliest detectable molecular event. Our bulk cerebellar RNA profiling and motor neuron-specific RNA profiling data at the onset of MATR3 loss revealed distinct molecular signatures. In the cerebellum, Ngfr expression emerged in Purkinje cells before the onset of neuronal loss and remained elevated throughout the disease course. This increase was accompanied by activation of the JNK-mediated cell death pathway. In the motor neurons, elevated Fgf21 and integrated stress response (ISR) gene expression were the first to be observed and persisted throughout disease progression, consistent with previous findings in SOD1 mouse models. Our findings provide mechanistic insights into the initiation of neurodegeneration in ALS-relevant motor-controlling neurons and implicate potential neuron type-specific targets for future therapeutics.

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.

Oxford Nanopore Technologies based methylation profiling enables rapid, accurate CNS tumor classification but is currently mostly performed with MinION flow cells (~$1,000 USD). We benchmarked the low cost Flongle flow cell (~$100, 10 fold reduction) for methylation based tumor classification. Across 12 pediatric CNS tumor samples with highly variable sequencing yields (17 to 330 Mbp), both Sturgeon and nanoDx classifiers achieved perfect diagnostic accuracy when operating above significance thresholds, despite extreme data sparsity. Bootstrap analysis determined empirically defined minimum data thresholds. These results demonstrate that Flongle flow cells enable cost effective deployment of rapid, accurate nanopore based CNS diagnostics while maintaining clinical reliability.

INTRODUCTION: Understanding concordance across biofluids and platforms is critical for understanding neurodegenerative biomarkers results and translating them into clinical use; yet systematic comparisons remain limited. To address this gap, we performed large-scale proteomic profiling of patient-paired plasma and CSF to characterize cross-modal relationships. METHODS: We profiled paired plasma and CSF from 67 individuals using SomaScan 11K and NULISAseq CNS panels. Disease severity was assessed with the CDR+NACC FTLD-M Global Score. RESULTS: We identified 269 SomaScan and 18 NULISA proteins with significant cross-biofluid correlation. Cross-platform concordance within biofluids was strong. NEFL, NPTX2, TREM2, and CHIT1 demonstrated consistent cross-platform agreement. Associations with disease severity were compartment-specific, with decreased NPTX2 in CSF, increased NEFL and GFAP in plasma, and decreased TREM2 across biofluids. DISCUSSION: Cross-platform consistency supports biomarker robustness, while limited cross-biofluid concordance highlights compartmental biology. Given additional clinical correlations despite varying underlying pathology, these findings may point to shared neurodegenerative disorder pathways.

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.

INTRODUCTION: Biomarkers for monitoring disease activity and treatment response in peripheral neuropathies remain limited. Big tau, a high-molecular-weight isoform of tau, is predominantly expressed in the peripheral nervous system (PNS). We investigated serum levels of big tau, brain-derived tau (BD-tau), and neurofilament light chain (NfL) in peripheral neuropathies, multiple sclerosis (MS), Alzheimer disease (AD), and healthy controls (HC). METHODS: Ultra-sensitive blood-based assays run on an HD-X Single Molecule Array analyser (Quanterix) were used to measure big tau and BD-tau in serum from patients with Guillain-Barr&eacute syndrome (GBS, n=81), Miller Fisher syndrome (MFS, n=20), Charcot-Marie-Tooth disease (CMT, n=102), chronic inflammatory demyelinating polyneuropathy (CIDP, n=43), MS (n=159), AD (n=20), and HC (n=41). NfL was measured in patients with neuropathies using an SR-X Single Molecule Array analyser (Quanterix). RESULTS: Serum big tau levels were higher in GBS than in AD (11.4 vs 2.4 pg/mL, p<0.0001) and MS (11.4 vs 9.0 pg/mL, p=0.01), and similar to CIDP and CMT. Contrarily, serum BD-tau levels in GBS were higher than in CIDP (3.0 vs 2.3 pg/mL, p=0.006) and MS (3.0 vs 1.7 pg/mL, p<0.0001), but similar to CMT, and lower than in AD (3.0 vs 9.8 pg/mL, p<0.0001). Serum NfL levels were higher in GBS than in CIDP (32.5 vs 13.0 pg/mL, p=0.0002), CMT (32.5 vs 12.3 pg/mL, p<0.0001), and HC (32.5 vs 7.6 pg/mL, p<0.0001). Compared with GBS, MFS patients showed higher BD-tau (12.7 vs 3.0 pg/mL, p=0.003), lower big tau (5.4 vs 11.4 pg/mL, p=0.002), and higher NfL levels, although the latter did not reach statistical significance (118.3 vs 32.5 pg/mL, p=0.16). The NfL/big tau ratio was significantly higher in MFS than in GBS, CIDP, and CMT. In GBS, BD-tau correlated with early clinical severity (MRC at 1 week; I-RODS at 4 weeks; maximum GBS-DS and GBS-DS at 4 weeks), whereas neither tau biomarker showed long-term clinical correlations. Higher BD-tau and big tau levels were associated with the need for mechanical ventilation (BD-tau: 8.6 vs 2.9 pg/mL, p=0.019; big tau: 19.7 vs 10.7 pg/mL, p=0.007), while higher BD-tau levels were associated with mortality (10.9 vs 2.9 pg/mL, p=0.003). CONCLUSIONS: Higher big tau levels in peripheral neuropathies than in CNS diseases support its role as a PNS-specific biomarker. In MFS, increased serum BD-tau, reduced big tau, and an elevated NfL/big tau ratio suggest CNS involvement with relative preservation of the PNS.

Creutzfeldt-Jakob disease (CJD) is a rapidly progressive, fatal neurodegenerative disorder caused by the misfolding of the normal prion protein (PrPC) into its pathogenic form (PrPSc). Therefore, effective clinical management and infection control require diagnostic assays that combine sub-attomole sensitivity with a turnaround time compatible with routine practice. While real-time quaking-induced conversion (RT-QuIC) offers excellent analytical performance, its protracted reaction time and dependence on plate readers that integrate high-speed shaking with continuous fluorescence detection restrict its widespread implementation. Endpoint QuIC (EP-QuIC) is an established alternative approach that utilizes vigorous shaking in a ThermoMixer(R) C and relies on a single endpoint fluorescence readout. In this study, we introduced minor modifications to assay conditions to enhance the speed, robustness, and reproducibility of EP-QuIC. The refined assay detected prion seeding activity in sporadic CJD brain samples within 6 h--and, in some cases, as early as 2 h. Unlike RT-QuIC, problematic recombinant PrP batches that impaired assay reliability in RT-QuIC--exhibiting delayed kinetics and spontaneous ThT fluorescence in unseeded reactions--did not produce such effects in EP-QuIC. Instead, they yielded strong, specific fluorescence signals that clearly identified sCJD samples, demonstrating improved batch tolerance and diagnostic efficiency of this approach. When applied to CSF, the EP-QuIC assay showed 100% sensitivity and specificity, highlighting its potential as a rapid, reliable diagnostic tool for prion diseases.

Mitochondria are central hubs for cellular metabolism and mitochondrial dysfunction is a hallmark of many chronic diseases. Consequently, changes in mitochondrial DNA copy number (mtDNA-CN), the number of mtDNA genomes per cell or tissue sample, are associated with diseases ranging from cancer and obesity to psoriasis and all-cause mortality. MtDNA-CN especially holds promise as a biomarker for neurodegenerative diseases, but whether and how mtDNA-CN changes with neurodegeneration is controversial. Here, we performed a systematic review and meta-analysis of 76 studies including 156 comparisons of mtDNA-CN in populations with or without a neurodegenerative disease to identify overall trends and potential moderators that explain variation among studies. Overall, mtDNA-CN was not statistically different with neurodegeneration, but heterogeneity among studies was extreme (I2 = 99.5%). The diagnosed disease explained the most variation. For example, Alzheimer's patients showed a 21% decrease in mtDNA-CN, but there was no change in mtDNA-CN with Parkinson's disease. Decreases in mtDNA-CN during neurodegeneration were also more extreme at older ages. Surprisingly, the tissue sampled for mtDNA-CN was not particularly influential, except for certain diseases. Studies published in earlier years also showed more extreme decreases in mtDNA-CN with neurodegeneration. Excessive heterogeneity persisted even after accounting for all moderators and their interactions (I2 = 85.7%). We conclude that the general perception of decreased mtDNA-CN with neurodegeneration is a vast oversimplification that may stem from legacy effects of early studies. However, mtDNA levels offer great promise as biomarkers for neurodegeneration, other diseases, and general health metrics, assuming appropriate complications can be considered.

Spinal muscular atrophy (SMA) is increasingly recognized as a multisystem disorder involving non-neuronal cells, yet the role of Schwann cells (SCs) in late-onset SMA (loSMA) remains unclear. We investigated age-dependent peripheral nerve pathology in a four-copy SMN2 mouse model of loSMA. Sciatic nerves from wild-type and loSMA mice were analyzed at postnatal (P) days 20, 35, 70, and >100 using semi-thin morphometry, immunofluorescence for MBP, Sox10, Sox2, and F4/80, and nerve conduction studies. loSMA nerves showed reduced myelin thickness at all time points and smaller axon diameters at P20 and P35. G- ratios were reduced at P20 but increased from P35 onward, indicating progressively altered axon-myelin relationships. MBP immunofluorescence intensity, compound muscle action potential amplitude, and nerve conduction velocity were reduced in loSMA mice at P>100. The proportion of Sox2+ SCs increased from P35 onward, while Sox10+ cell abundance increased at later stages. F4/80+ macrophages were transiently elevated at P35 and correlated with Sox2+ cell numbers at this stage. These findings demonstrate age-dependent myelin abnormalities, altered SC states, and transient accumulation of macrophages in loSMA peripheral nerves. Whether these changes are SC-autonomous or secondary to chronic axonal dysfunction remains to be determined.

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.