Category: Statistics

J Comput Neurosci. 2026 Jan 15. doi: 10.1007/s10827-025-00918-1. Online ahead of print.

ABSTRACT

Understanding how neurons encode multiple simultaneous stimuli is a fundamental question in neuroscience. We have previously introduced a novel theory of stochastic encoding patterns wherein a neuron’s spiking activity dynamically switches among its constituent single-stimulus activity patterns when presented with multiple stimuli (Groh et al., 2024). Here, we present an enhanced, comprehensive statistical testing framework for such “multiplexing”. As before, our approach evaluates whether dual-stimulus responses can be accounted for as mixtures of Poissons related to single-stimulus benchmarks. Our enhanced framework improves upon previous methods in two key ways. First, it introduces a stronger set of foils for multiplexing, including an “overreaching” category that captures overdispersed activity patterns unrelated to the single-stimulus benchmarks, reducing false detection of multiplexing. Second, it detects continuous mixtures, potentially indicating faster fluctuations – i.e. at sub-trial timescales – that would have been overlooked before. We utilize a Bayesian inference framework, considering the hypothesis with the highest posterior probability as the winner, and employ the predictive recursion marginal likelihood method for non-parametric estimation of the latent mixing distributions. Reanalysis of previous findings confirms the general observation of fluctuating activity and indicates that fluctuations may well occur on faster timescales than previously suggested. We further confirm that multiplexing is more prevalent for (a) combinations of face stimuli than for faces and non-face objects in the inferotemporal face patch system; and (b) distinct vs fused objects in the primary visual cortex.

PMID:41537936 | DOI:10.1007/s10827-025-00918-1

Eur Radiol. 2026 Jan 15. doi: 10.1007/s00330-025-12261-1. Online ahead of print.

ABSTRACT

OBJECTIVES: This study aims to develop a deep learning model to assist physicians in accurately classifying negative, equivocal, and positive β-amyloid (Aβ) deposition stages in Alzheimer’s disease (AD).

MATERIALS AND METHODS: 1327 subjects from two cohorts underwent [¹⁸F]Florbetapir PET and were grouped by Aβ deposition. A cascaded attention-guided vision transformer (CA-ViT) framework was proposed to extract biologically significant regional information for fine-grained classification. To evaluate clinical utility, we assessed the diagnostic performance of physicians with and without the assistance of our proposed method.

RESULTS: The CA-ViT model demonstrated outstanding cross-center performance, achieving accuracies of 82.8% [79.1%, 86.5%] (96% confidence interval, CI) and 78.0% [75.1%, 80.9%] in three-class classification tasks in the two cohorts, respectively. Our proposed model-assisted physicians exhibited significant improvements in diagnostic accuracy (from 56% to 66% and from 50% to 77%).

CONCLUSION: The CA-ViT model effectively decodes fine-grained pathological information from [¹⁸F]Florbetapir PET imaging, enabling accurate stratification of Aβ deposition to assist physicians in early monitoring of AD.

KEY POINTS: Question Deep learning has the potential to assist physicians in accurately classifying β-amyloid deposition stages in early Alzheimer’s disease. Findings The proposed diagnostic model is a promising computer-aided tool for early assessment of amyloid deposition and demonstrates improved physician performance. Clinical relevance Equivocal amyloid deposition often complicates early Alzheimer’s disease diagnosis and may delay optimal interventions. Our model, validated on PET scans from multiple centers, enhances the identification of these equivocal cases and improves diagnostic accuracy among less-experienced physicians.

PMID:41537783 | DOI:10.1007/s00330-025-12261-1

Eur Radiol. 2026 Jan 15. doi: 10.1007/s00330-025-12253-1. Online ahead of print.

ABSTRACT

OBJECTIVES: Besides clinical examination, cranial CT plays a critical role in diagnostics in neurosurgery. In trauma cases or perioperatively, having low-barrier access to CT-like imaging would be highly beneficial. Therefore, this feasibility study examines at an early stage if and how well synthetic cranial CT imaging can be generated from biplanar radiographs of adult neurosurgical patients using deep learning.

MATERIALS AND METHODS: Two 2D to 3D generative adversarial networks (GANs) were trained for the generation of synthetic cranial CTs using radiographs taken in two planes as input. Model 1 uses digitally reconstructed radiographs (DRRs) as input, while model 2 was trained using real X-rays. In total, model 1 was trained and validated using 235 images from three separate centers. Model 2 was trained and tested using 1323 images from a single center.

RESULTS: The performance of the model using DDRs as input reached a peak-signal-to-noise ratio (PSNR) of 15.61 and a structural similarity index measure (SSIM) of 0.782 during external validation. The second model, using real X-rays as input, attained a PSNR of 14.69 and an SSIM of 0.717 upon internal validation.

CONCLUSIONS: At the present stage, the synthetic cranial tomography scans generated as part of this study show promise but do not seamlessly correspond to ground-truth CTs. However, this proof-of-concept study is the first to derive such artificial cranial images using deep learning and can serve as a starting point for further investigation.

KEY POINTS: Question Cranial computed tomography involves radiation, logistical challenges, and access is limited in rural areas. Generating synthetic CT images with deep learning could address these challenges. Findings Two deep-learning models were trained to produce CT images from radiographs. Reconstruction from DRRs is promising, but using real X-rays remains more challenging. Clinical relevance As a proof-of-concept, the models’ exact clinical relevance remains to be defined. The proposed approach may broaden access to tomographic neuroimaging, reduce radiation, and enhance intraoperative and maybe even diagnostic support, potentially improving outcomes in neurosurgery and neuro-critical care.

PMID:41537782 | DOI:10.1007/s00330-025-12253-1

Eur Radiol. 2026 Jan 15. doi: 10.1007/s00330-025-12246-0. Online ahead of print.

ABSTRACT

OBJECTIVES: Pretreatment MRI was employed to develop and validate a combined model integrating clinical features with deep learning for rectal cancer.

MATERIALS AND METHODS: We retrospectively collected 458 patients from three hospitals and followed them up for at least 3 years. Clinical, pathological and imaging data were collected. Multi-instance learning (MIL) was used to integrate prediction across multiple slices to improve the performance of the model. To improve predictive performance, a nomogram combining deep learning features and clinicopathologic parameters was constructed. Model performance was assessed using Harrell’s C-index and time-dependent ROC curves.

RESULTS: The training set included 268 patients, 115 patients in the validation set and 75 patients in the external test set. For OS, the MIL model achieved a C-index of 0.757 in the training cohort, 0.754 in the validation cohort, and 0.741 in the test cohort, compared to 0.666, 0.772, and 0.756 for the clinical model, respectively. The combined model, which integrates MIL features with clinical features, further improved predictive performance, with C-index values for OS at 0.819, 0.822 and 0.759 and for DFS at 0.768, 0.750 and 0.721 across the training, validation and external test cohorts.

CONCLUSIONS: By leveraging the complementary strengths of clinical and deep learning approaches, the combined model enhances predictive robustness, enabling more accurate and personalized pretreatment risk assessment in rectal cancer.

KEY POINTS: Question Rectal cancer management requires more precise prognostic models to optimize treatment strategies and improve clinical decision-making for individual patients. Findings The combined model leverages synergistic effects between clinical and deep learning features, achieving enhanced prognostic performance and enabling more personalized pretreatment risk stratification. Critical relevance This study demonstrates that MIL extracts deep learning features complementary to clinical knowledge. The combined model leverages this synergy, providing clinicians with a more powerful tool for personalized prognostic assessment.

PMID:41537781 | DOI:10.1007/s00330-025-12246-0

Eur Radiol. 2026 Jan 15. doi: 10.1007/s00330-025-12180-1. Online ahead of print.

ABSTRACT

BACKGROUND: The CT-derived density of coronary artery calcification is increasingly associated with the risk of ischemic heart disease. Whether this principle also applies to intracranial artery calcifications (IAC) and cerebrovascular disease risk is unknown, primarily due to the lack of population-based estimates of IAC density and its determinants. We investigated these facets in this cohort study.

MATERIALS AND METHODS: In 2464 community-living individuals who underwent non-contrast CT, we measured IAC density and assessed its correlation with IAC volume using Spearman’s ρ. We described its distribution in intracranial carotid artery calcification (ICAC), with specific estimates for its subtypes, and vertebrobasilar artery calcification (VBAC). We investigated associations between risk factors and IAC density using multivariable ordinal regression models.

RESULTS: The prevalence of IAC was 82.8%, with a median density of 232 (IQR 189-287) HU. IAC density correlated moderately with volume (ρ 0.67, 95% CI [0.65-0.70]). ICAC was predominantly composed of higher density, with 80.1% of affected participants having components of ICAC above 400 HU, whereas only 32.0% of participants with VBAC had components above 400 HU. Intimal subtype ICACs showed a predominance for lower densities when compared to medial subtype ICACs. The main determinants of IAC density were hypertension, use of lipid-lowering medication, and smoking, with adjusted odds ratios of 1.59 [1.28-1.90], 1.55 [1.26-1.91], and 1.33 [1.10-1.61], respectively.

CONCLUSION: IAC density differs significantly between the anterior and posterior cerebropetal arteries. While IAC density correlated only moderately with its volume, the associations between cardiovascular risk factors and IAC density were mostly similar to those observed with IAC volume.

KEY POINTS: Question Drivers of the CT density of intracranial artery calcifications are unknown and may reveal novel risk targets for population-based prevention strategies. Findings Calcifications of the anterior cerebral circulation are denser than those of the posterior circulation. Hypertension, diabetes, and smoking are key drivers of calcification density, resembling most drivers of its volume. Clinical relevance Calcification density may serve in distinguishing subtypes of intracranial calcifications, improving detection of subtype-specific effects. Further research is warranted to determine the role of intracranial arteriosclerosis density in prevention strategies for cerebrovascular diseases.

PMID:41537780 | DOI:10.1007/s00330-025-12180-1

Nat Methods. 2026 Jan 15. doi: 10.1038/s41592-025-02923-9. Online ahead of print.

ABSTRACT

Microbial community analysis typically involves determining which microbial features are associated with properties such as environmental or health phenotypes. This task is impeded by data characteristics, including sparsity (technical or biological) and compositionality. Here we introduce MaAsLin 3 (microbiome multivariable associations with linear models) to simultaneously identify both abundance and prevalence relationships in microbiome studies with modern, potentially complex designs. MaAsLin 3 can newly account for compositionality either experimentally (for example, quantitative PCR or spike-ins) or computationally, and it expands the range of testable biological hypotheses and covariate types. On a variety of synthetic and real datasets, MaAsLin 3 outperformed state-of-the-art differential abundance methods, and when applied to the Inflammatory Bowel Disease Multi-omics Database, MaAsLin 3 corroborated previously reported associations, identifying 77% with feature prevalence rather than abundance. In summary, MaAsLin 3 enables researchers to identify microbiome associations more accurately and specifically, especially in complex datasets.

PMID:41540124 | DOI:10.1038/s41592-025-02923-9

Nat Methods. 2026 Jan 15. doi: 10.1038/s41592-025-02948-0. Online ahead of print.

ABSTRACT

Spatially mapping the transcriptome and proteome in the same tissue section can profoundly advance our understanding of cellular heterogeneity and function. Here we present Deterministic Barcoding in Tissue sequencing plus (DBiTplus), an integrative multimodal spatial omics approach combining sequencing-based spatial transcriptomics and multiplexed protein imaging on the same section, enabling both single-cell-resolution cell typing and transcriptome-wide interrogation of biological pathways. DBiTplus utilizes spatial barcoding and RNase H-mediated cDNA retrieval, preserving tissue architecture for multiplexed protein imaging. We developed computational pipelines to integrate these modalities, allowing imaging-guided deconvolution to generate single-cell-resolved spatial transcriptome atlases. We demonstrate DBiTplus across diverse samples including frozen mouse embryos, and formalin-fixed paraffin-embedded human lymph nodes and lymphoma tissues, highlighting its compatibility with challenging clinical specimens. DBiTplus uncovered mechanisms of lymphomagenesis, progression and transformation in human lymphomas. Thus, DBiTplus is a unified workflow for spatially resolved single-cell atlasing and unbiased exploration of biological mechanisms in a cell-by-cell manner at transcriptome scale.

PMID:41540123 | DOI:10.1038/s41592-025-02948-0

Neural Netw. 2026 Jan 7;198:108520. doi: 10.1016/j.neunet.2025.108520. Online ahead of print.

ABSTRACT

The paper presents novel Universum-enhanced classifiers: the Universum Generalized Eigenvalue Proximal Support Vector Machine (U-GEPSVM) and the Improved U-GEPSVM (IU-GEPSVM) for EEG signal classification. Using the computational efficiency of generalized eigenvalue decomposition and the generalization benefits of Universum learning, the proposed models address critical challenges in EEG analysis: non-stationarity, low signal-to-noise ratio, and limited labeled data. U-GEPSVM extends the GEPSVM framework by incorporating Universum constraints through a ratio-based objective function, while IU-GEPSVM enhances stability through a weighted difference-based formulation that provides independent control over class separation and Universum alignment. The models are evaluated on the Bonn University EEG dataset across two binary classification tasks: (O vs S)-healthy (eyes closed) vs seizure, and (Z vs S)-healthy (eyes open) vs seizure. IU-GEPSVM achieves peak accuracies of 85% (O vs S) and 80% (Z vs S), with mean accuracies of 81.29% and 77.57% respectively, outperforming baseline methods. Rigorous statistical validation confirms these improvements: Friedman tests reveal significant overall differences, pairwise Wilcoxon signed-rank tests with Bonferroni correction establish IU-GEPSVM’s superiority over all baselines, and win-tie-loss analysis demonstrates practical significance. Overall, integrating interictal Universum data yields an efficient and reliable solution for neurological diagnosis.

PMID:41538899 | DOI:10.1016/j.neunet.2025.108520

J Radiol Prot. 2026 Jan 15. doi: 10.1088/1361-6498/ae38ee. Online ahead of print.

ABSTRACT

Thirteen years after the Fukushima nuclear accident, Fukushima Prefecture still faces major challenges in recovery, especially concerning the negative image. Perception of spatial stigma refers to the residents’ concerns about the negative image of their region and its residents as perceived by the public. The present study aims to clarify the perception of spatial stigma and its associated factors among residents.
A questionnaire survey was conducted among local residents from December 2023 to January 2024 in Tomioka, Okuma, and Futaba towns. Statistical analysis was done using the chi-square test and logistic regression.
67.8% of participants expressed a strong perception of spatial stigma. Perception of spatial stigma was independently correlated with living in the FDNPP location, high radiation health risk perception, anxiety about treated water release, uncertainty of returning, and poor mental health.
Actively addressing the stigma of Fukushima through targeted countermeasures is crucial for alleviating residents’ perception of spatial stigma. These efforts are vital for fostering recovery and achieving comprehensive revitalization of Fukushima Prefecture.

PMID:41538886 | DOI:10.1088/1361-6498/ae38ee

JMIR Form Res. 2026 Jan 15;10:e70809. doi: 10.2196/70809.

ABSTRACT

BACKGROUND: Heart failure (HF) and chronic obstructive pulmonary disease (COPD) are responsible for a significant amount of the economic and chronic disease burden that impacts the Ontario health system. Telehomecare, a home self-management program launched by the Ontario Telemedicine Network (OTN), was created to improve access to quality care and limit health care use. However, few data are available on patient-, caregiver-, and provider-reported outcomes of telehomecare.

OBJECTIVE: This study aims to evaluate the impact of the OTN telehomecare program on the health-related quality of life (HRQoL), disease-management skills, and satisfaction of patients with HF and those with COPD; informal caregiver strain index; and nurse satisfaction with telehomecare.

METHODS: We used a prospective longitudinal cohort design, including patients with HF and those with COPD enrolled in Ontario’s telehomecare program, informal caregivers of patients in the program, and nurses providing services in telehomecare. Patients and informal caregivers were administered telephone surveys at baseline, month 3, month 6, and month 12 follow-up from July 2016 to December 2019. The outcomes for the longitudinal surveys were patient-perceived HRQoL, disease self-management skills, perception of telehomecare (ease of use and usefulness), satisfaction with telehomecare, and informal caregiver-perceived strain. Cross-sectional surveys were conducted with nurses to assess nurse perception and satisfaction with telehomecare. Participant data were analyzed using general linear mixed models in SAS Statistical Software (version 9.4; SAS Institute Inc).

RESULTS: Overall, a total of 194 patients (HF, n=117; COPD, n=77), 62 caregivers, and 24 nurses participated, with an overall response rate of 51% (280/551). The average age of patients with HF and those with COPD was 71 (SD 11.3) years and 70 (SD 11.1) years, respectively, and 52% (100/194) were men. A significant improvement in overall HRQoL was observed among patients with HF at month 12 (-18.37, P<.001). Minimal clinically important differences were observed across all HRQoL domains for people with HF, indicating clinically meaningful improvement over the study period. No statistically significant improvement in HRQoL was observed among patients with COPD; however, minimal clinically important differences were observed in the physical functioning dimension. Patients reported being confident in self-managing their diseases throughout the study, but as patients aged, their perception of and satisfaction with telehomecare was shown to decrease (P=.002 and P=.002, respectively). Caregivers reported relatively low strain scores (mean 10.3, SD 5.9) throughout the program, and nurses reported moderate levels of satisfaction (mean 6.7, SD 1.5) with telehomecare at follow-up.

CONCLUSIONS: In this population, telehomecare demonstrated an ability to improve the HRQoL of patients with HF and those with COPD. However, the long-term sustainability of HRQoL improvements in patients following telehomecare requires further investigation. Furthermore, telehomecare was shown to decrease informal caregiver-perceived strain, and nurses described moderate levels of satisfaction and perceived quality of care with telehomecare.

PMID:41538796 | DOI:10.2196/70809