Progressive Supranuclear Palsy (PSP) is the most prevalent primary 4R-tauopathy, characterized by the pathogenic accumulation of misfolded tau protein within neurons and glial cells. Historically, clinical diagnosis relied upon the identification of Richardson’s Syndrome, however, the recognition of diverse clinical phenotypes that overlap with Parkinson’s disease, corticobasal syndrome, and frontotemporal dementia has complicated the diagnostic landscape and hindered the success of developing therapeutic interventions. As the field transitions toward a precision medicine paradigm, there is a growing need for validated biomarkers that can provide molecular specificity, facilitate early diagnosis, and accurately track disease progression. This paper reviews the recent advancements in neuroimaging and fluid-based biomarkers, assessing their potential to delineate PSP from similar neurodegenerative conditions and unlock the 4R-tau therapeutic pipeline. In the domain of neuroimaging, while structural magnetic resonance imaging (MRI) and the Magnetic Resonance Parkinsonism Index (MRPI) continue to provide measures of subcortical atrophy, the emergence of second-generation tau-selective positron emission tomography (PET) radioligands represents a transformative shift. New tau PET tracers offer the ability to visualize tau pathology in vivo, providing a more direct assessment of the underlying proteinopathy than traditional volumetric measures. These advancements are complemented by significant progress in fluid biomarkers. Plasma phosphorylated tau at residue 217 (p-tau217) has gained prominence as a robust marker for Alzheimer’s disease, and its primary utility in PSP research currently serves as a critical negative signature to exclude amyloid-associated co-pathology. In contrast, novel assays targeting microtubule-binding region tau fragments show burgeoning potential for the specific identification of 4R-tau isoforms. Furthermore, neurofilament light chain (NfL) has been firmly established as a sensitive, albeit non-specific, indicator of neuroaxonal injury and clinical severity. Additional advancements with digital health approaches and electrophysiological assessments add to the opportunities for improved objective measures. This review concludes that the shift from clinical-only diagnostic criteria to a biomarker-enabled molecular framework is the necessary catalyst for developing effective disease-modifying therapies for PSP and related 4R-tauopathies. The synthesis of these multimodal biomarkers into a unified framework will be essential to improve participant stratification, enable the use of adaptive trial models, and provide supportive evidence of target engagement for future clinical trials.
NeuroCon-AutismNet: a privacy-preserving multimodal framework toward autism screening via diffusion-regularized EEG biomarkers and empathy-aware multilingual dialogue
IntroductionAutism Spectrum Disorder (ASD) screening requires multimodal biomarkers to capture the heterogeneous neurological and behavioral phenotypes. Current screening approaches remain siloed across EEG analysis and conversational assessment, limiting integrated diagnostic architecture. Privacy-preserving machine learning frameworks for mental health screening are underdeveloped, particularly for multilingual deployment contexts. This paper presents NeuroCon-AutismNet, a candidate multimodal architecture integrating diffusion-regularized EEG synthesis, multilingual conversational screening, and formal differential privacy as architectural proof-of-concept. No diagnostic discrimination capability is claimed; all validation is scoped to synthetic evaluation.MethodsNeuroCon-AutismNet comprises four modules: (1) Temporal Diffusion Biomarker Generator (TDBG), a latent diffusion model over VAE-encoded 19-channel EEG; (2) Multilingual Affective Dialogue Screening Network (MADSN), a fine-tuned GPT-2-small module deployed in English, Spanish, and Hindi; (3) Neuro-Linguistic Fusion Transformer (NLFT), enforcing positional alignment as a design prior rather than learned cross-modal association; and (4) Adaptive Mixture-of-Experts Layer (AMEL-X) for entropy-regularized multimodal fusion. Formal (ε, δ)-differential privacy (ε = 1.0, δ = 1e-5) is verified via DP-SGD RDP composition (σ = 1.2, q = 0.0914, T = 550 steps, verified ε = 0.97). Privacy verification establishes architectural readiness for future real-data deployment; no real patient records are present in the training set.Results and DiscussionWithin closed synthetic evaluation, held-out diagnostic AUC is 0.503 (95% CI: 0.487–0.519, DeLong p = 0.67), statistically indistinguishable from chance and the central limitation of this study. Two partial external benchmarks are provided. Spectral comparison against three independently published real ASD EEG studies yields Pearson r = 0.87 across five frequency bands; delta and alpha directions are reproduced, but theta and gamma reproduce poorly with large amplitude errors (delta MAE 14.79%, alpha MAE 11.57%). Expert evaluation of MADSN outputs by 50 annotators under single-blind protocol yields 90% empathy satisfaction and Cohen’s κ = 0.82, reflecting text quality rather than clinical screening validity. The null diagnostic AUC and synthetic-only evaluation prevent any current screening or clinical-utility claims. Real-data EEG validation, clinician-caregiver interaction studies for MADSN, and DP-protected training on real patient records are prerequisites for future clinical deployment.
Prospective validation of imaging and serum diagnostic biomarkers of steatohepatitis and fibrosis in MASLD: the LITMUS Imaging Study
Nature Medicine, Published online: 24 July 2026; doi:10.1038/s41591-026-04496-2
In the LITMUS study, serum biomarkers outperformed imaging for identifying patients at risk of MASH, whereas elastography was superior for staging advanced fibrosis and cirrhosis.
Early detection of Alzheimer’s disease with circular RNA from blood
Nature Medicine, Published online: 23 July 2026; doi:10.1038/s41591-026-04563-8
Early diagnosis of Alzheimer’s disease is important to ensure timely and accurate treatment. We show that a blood-based circular RNA signature can accurately detect Alzheimer’s disease as well as, or better than, current biomarkers and even predict who will develop symptoms before they appear. This less invasive approach could offer earlier diagnosis and improved monitoring of Alzheimer’s disease.
<![CDATA[ASCP updates spotlight ketamine safety guardrails, at‑home agitation therapy, psilocybin phase 3 momentum, and biomarker-driven precision psychiatry.]]>
Sex-related molecular phenotypes in anxiety-depressive disorders: a machine learning analysis of routine blood biomarkers
BackgroundAnxiety disorders and depressive disorders are the most prevalent mental disorders worldwide. Their diagnosis has long relied on clinical symptom assessment, and objective blood−based biomarkers remain lacking. Sex is a critical risk factor for these disorders; however, sex−specific divergence in blood biochemical profiles has yet to be systematically characterized.MethodsThis retrospective study enrolled 778 patients diagnosed with anxiety−depressive state at China−Japan Friendship Hospital. Demographic data, complete blood count parameters, and blood biochemical parameters were collected. Following missing value processing and multiple imputation, Mann–Whitney U tests were applied to identify sex−differentially expressed biomarkers. A random forest classifier was constructed to evaluate the discriminative capacity of combined multi−marker panels, with model performance comprehensively assessed through receiver operating characteristic curve analysis, SHAP−based explainability analysis, and multi−classifier probability projection. Age−stratified analyses were performed with a threshold of 50 years to explore the potential modifying effect of age on sex differences.ResultsSeveral biomarkers exhibiting significant differences between males and females were identified (FDR < 0.05), among which creatinine, hemoglobin, hematocrit, red blood cell count, and uric acid demonstrated the largest effect sizes. The random forest model achieved an area under the receiver operating characteristic curve of 0.902 on the independent test set. Multi−classifier probability projection following hyperparameter tuning yielded a Silhouette coefficient of 0.464 in the two−dimensional space, with permutational multivariate analysis of variance confirming highly significant centroid differences between groups (p < 0.001). Age−stratified analysis using hemoglobin as an example revealed that levels in males were significantly higher than those in females across both age strata, with the magnitude of the sex difference attenuated in the ≥50−year group compared with the <50−year group.ConclusionsRobust sex−related signals are embedded in routine blood biochemical markers. Although complete separation is difficult to achieve under unsupervised dimensionality reduction, these signals can be efficiently integrated through ensemble learning algorithms. This study provides a molecular phenotypic basis related to sex in patients with anxiety−depressive state and underscores the importance of fully considering sex as a variable in clinical laboratory testing.
Integration of imaging and clinical biomarkers for cerebral infarction diagnosis via NeuroFusionNet
BackgroundCerebral infarction remains a leading cause of mortality and long-term disability worldwide, demanding rapid and accurate diagnostic strategies. However, current assessments primarily rely on imaging interpretation, often neglecting valuable clinical and laboratory information that could enhance diagnostic precision.MethodsWe developed NeuroFusionNet, a multi-modal deep learning framework that integrates imaging features with clinical biomarkers for binary classification of cerebral infarction and healthy controls. The model combines a ResNet-based visual encoder with a multilayer perceptron branch for clinical indicators, achieving end-to-end feature fusion and joint optimization.ResultsNeuroFusionNet achieved superior diagnostic performance with an accuracy of 0.9655, precision of 0.9584, recall of 0.9584, and F1-score of 0.9584, significantly outperforming baseline models including ResNet, MobileNet, and GhostNet. The integration of imaging and clinical biomarkers effectively enhanced model sensitivity and robustness, demonstrating strong potential for real-world clinical application.ConclusionOur findings highlight the clinical value of integrating imaging and laboratory data for precision diagnosis of cerebral infarction. NeuroFusionNet provides a scalable and interpretable framework that may support early detection and personalized management of cerebrovascular diseases in routine clinical practice.
Optimizing Treatment Strategies in the Bipolar Disorder Spectrum With Classical AI Approaches: Systematic Review of Performance, Bias, and Clinical Applicability
Background: Bipolar disorder (BD) is a complex and heterogeneous psychiatric condition, characterized by fluctuating clinical courses that affect approximately 1%‐2% of the global population in their lifetime. Despite pharmacological advances, treatment response varies significantly among patients, making the identification of individualized treatment strategies a major challenge. Artificial Intelligence (AI), through its classical approaches, has emerged as a powerful tool in precision psychiatry to identify subtle patterns in complex data and inform personalized clinical decisions. Objective: The present systematic review aimed to examine the current evidence on classical AI-supported treatment optimization in the BD spectrum. Methods: The review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines. Four databases (PubMed, Web of Science, Scopus, and Embase) were searched for original studies published after 2015 on the application of classical AI in the treatment of BD in adult patients. Publication bias was evaluated by visual inspection of a funnel plot. The methodological quality, risk of bias, and clinical applicability of the predictive models were assessed using the Prediction Model Risk Of Bias Assessment Tool for prediction models using regression or AI methods (PROBAST+AI; PROBAST+AI Working Group) tool. Results: A total of 35 studies were included and classified into 5 outcome-based categories, including acute symptomatic response, long-term maintenance response, relapse and readmission risk, safety and dose optimization, and brain aging and phenotyping. Acute symptomatic response models performed modestly (pooled area under the curve [AUC] 0.68), while imaging improved accuracy (74%‐77%). Long-term maintenance response models showed moderate-to-high performance (pooled AUC 0.80), with biomarker- and cellular-based models reaching 96%‐99% accuracy. Relapse and readmission prediction achieved a pooled AUC of 0.71, with digital phenotyping and rule-based methods performing best (AUC 0.85‐0.88). Safety and dose optimization models achieved 85%‐97% accuracy. Brain aging and phenotyping studies highlighted accelerated brain aging in BD, partially mitigated by lithium, and revealed novel data-driven subgroups. However, 3 studies were considered at high risk of bias due to small sample sizes associated with disproportionately high-performance estimates. An additional study was identified as potentially biased because it lay markedly distant from the funnel plot’s confidence line. Finally, the PROBAST+AI assessment revealed a high risk of bias in most studies, primarily due to data analysis limitations, small sample sizes, and lack of external validation. Conclusions: The adoption of classical AI tools in BD serves as a driver for therapeutic optimization, although current AI tools in BD should still be considered exploratory rather than ready for clinical use. Effective implementation in real-world clinical scenarios requires more robust, transparent, and externally validated models to ensure reliability and generalizability.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/df358c951a2c44fbd20077fdda306baf" />
Blood Protein Panel May Help Distinguish Major Dementia Types
Diagnosing dementia is rarely straightforward, particularly early in the disease course. Alzheimer’s disease, dementia with Lewy bodies, and frontotemporal dementia can overlap clinically, while mixed pathology is common in older patients. This creates a major barrier for precision medicine: treatment selection, trial enrolment, prognosis, and patient counselling increasingly require a more biologically grounded diagnosis.
Blood biomarkers have already begun to change Alzheimer’s disease diagnostics, especially for detecting amyloid and tau pathology. But the field still lacks robust plasma tools for distinguishing Alzheimer’s disease from dementia with Lewy bodies (DLB) or frontotemporal dementia (FTD).
Proteomics across dementia cohorts
A new study, published in Nature Aging, used proximity extension assay proteomics to profile plasma proteins across international dementia cohorts. In the discovery phase, researchers analyzed more than 1,300 plasma samples, including controls and individuals across preclinical, mild cognitive impairment, and dementia stages of Alzheimer’s disease, dementia with Lewy bodies, and frontotemporal dementia.
The analysis identified more than 200 dysregulated proteins across disease groups. For Alzheimer’s disease, glial fibrillary acidic protein, or GFAP, showed the strongest increase across the disease continuum. Neurofilament light chain also rose with clinical stage, while several other proteins declined as Alzheimer’s disease progressed. However, the authors emphasize that these proteomic markers did not outperform established plasma markers of amyloid and tau pathology for detecting Alzheimer’s disease.
The more clinically novel findings came from the non-Alzheimer’s dementias. In dementia with Lewy bodies, integrin alpha-V and integrin alpha-M were consistently reduced, including in analyses stratified by amyloid status and in autopsy-confirmed Lewy body disease. The same integrin-related signal was also seen in Parkinson’s disease data from the PPMI cohort, supporting a broader link to Lewy body pathology.
For frontotemporal dementia, neurofilament light chain remained one of the strongest markers, consistent with its role as a general marker of neuroaxonal injury. Lower GFAP helped distinguish frontotemporal dementia from Alzheimer’s disease, while proteins such as OSM appeared more relevant in earlier frontotemporal degeneration.
A 21-protein dementia panel
The researchers then refined these signals into a custom 21-protein plasma panel and tested it in an independent multicenter cohort. The panel showed its strongest value in differential diagnosis, helping separate dementia with Lewy bodies and frontotemporal dementia from both controls and Alzheimer’s disease dementia. Its performance was moderate to good across these comparisons, suggesting that plasma proteomics may be most useful as an added layer of biological stratification when clinical symptoms overlap.
These values are not sufficient to replace specialist clinical assessment, CSF testing, imaging, or established Alzheimer’s blood biomarkers. But they suggest that plasma proteomics could provide clinically useful support where diagnostic uncertainty remains high.
Toward biomarker-based dementia stratification
The immediate relevance is not simply another biomarker list. The study addresses a practical gap in dementia medicine: identifying scalable blood-based tools that help separate biologically different diseases with overlapping symptoms. That could improve referral pathways, enrich clinical trials with the right patient populations, and support future disease-modifying therapies beyond Alzheimer’s disease.
Important limitations remain. Many DLB and FTD diagnoses were clinical rather than autopsy-confirmed, prodromal groups were relatively small, and biomarker performance may depend on assay platform and cohort calibration. Mixed pathology, especially coexisting Alzheimer’s and Lewy body disease, remains a major challenge.
Even so, the study provides a strong proof of concept. Plasma proteomics may help move dementia diagnostics from broad syndromic categories toward molecular stratification, an essential step if precision neurology is to match the progress already seen in Alzheimer’s biomarker development.
The post Blood Protein Panel May Help Distinguish Major Dementia Types appeared first on Inside Precision Medicine.
Altered tryptophan metabolism as a contributor to cognitive impairment in chronic kidney disease: a narrative review
Approximately 40% of patients with chronic kidney disease (CKD) experience cognitive impairment (CI), which is strongly associated with increased mortality. CI is driven by multiple factors, including vascular injury, accumulation of uremic toxins, disruption of the blood–brain barrier, and chronic inflammation. Recent evidence suggests that kidney disease and neurocognitive decline are mechanistically linked through dysregulated tryptophan metabolism. Tryptophan is metabolised through three main pathways: the kynurenine, indole, and serotonin pathways, each producing bioactive metabolites with distinct neurophysiological effects. The hallmarks of CKD include chronic inflammation, gut microbial dysbiosis, and impaired renal clearance, all of which alter tryptophan metabolism. Inflammation drives tryptophan metabolism towards the kynurenine pathway, increasing the formation of neurotoxic compounds that promote oxidative stress, excitotoxicity, and neuronal injury. However, reduced availability of tryptophan for serotonin synthesis impairs serotonergic signalling and neurotransmission, as well as melatonin biosynthesis, thereby contributing to circadian rhythm disturbances and impaired glymphatic clearance. Concurrently, gut dysbiosis and reduced renal clearance promote the accumulation of indole-derived uremic toxins, leading to endothelial dysfunction, neuroinflammation, and disruption of the blood–brain barrier. This review highlights the current evidence of dysregulated tryptophan metabolism in CKD and its impact on the pathogenesis of neurocognitive complications. The review also discusses potential biomarkers and therapeutic strategies, including kynurenine pathway inhibitors, gut microbiota modulation, uremic toxin adsorption, melatonin supplementation and personalised medicine to mitigate cognitive impairment in CKD.