Science North

Gamma synchrony is ubiquitous in visual cortex, but whether it contributes to perceptual grouping remains contentious based on observations that gamma frequency is not consistent across stimulus features and that gamma synchrony depends on distances between image elements. These stimulus dependencies have been argued to challenge the idea that the visual system groups image elements by synchronizing the neural assemblies that encode them. Here, we argue instead that these dependencies may shape synchrony in perceptually meaningful ways. Indeed, according to the theory of weakly coupled oscillators (TWCO), synchrony-based grouping mechanisms require stimulus dependence. Synchronization among coupled oscillators depends on frequency dissimilarity and coupling strength, which in early visual cortex relate to local feature dissimilarity and physical distance, respectively. We manipulated these factors in a texture segregation experiment wherein human observers identified the orientation of a figure defined by reduced contrast heterogeneity compared to the background. Human performance followed TWCO predictions both qualitatively and quantitatively, as formalized in a computational model. Moreover, we found that when enriched with a Hebbian learning rule, our model also predicted human learning effects: Increases in model gamma synchrony due to perceptual learning predicted improvements in texture segregation across sessions. Taken together, our data suggest that the stimulus-dependence of gamma synchrony captures local image statistics and is linked to the stimulus-dependence of texture segregation, and that the effect of visual experience on gamma synchrony provides a viable perceptual learning mechanism for training-induced improvements in texture segregation. Our results suggest that gamma synchrony with its inherent stimulus dependencies can provide a plausible mechanistic basis for perceptual grouping and visual scene segmentation.

representations that capture latent regulatory and functional features. Compared with task-specific models, FMs offer enhanced generalization, cross-species transferability, and scalability, making them particularly attractive for plant systems characterized by genomic complexity and limited functional annotations. Providing a systematic overview of this paradigm, Xu et al. present a mini-review synthesizing recent advances in foundation models for plant molecular biology. The review traces the evolution from general DNA language models to plant-specific tools and highlights key challenges unique to plant systems, including polyploidy, repetitive genomes, and sparse experimental annotations. By outlining future directions such as multimodal integration and computational efficiency, this work establishes a conceptual framework for understanding how FMs are redefining computational plant biology and guiding next-generation model development.Building on this FM paradigm, several contributions demonstrate how representation learning can be applied to concrete biological problems. Zhang et al. applied a DNABERT-2-based framework combined with gradient boosting to identify DNA N6-methyladenine modifications in rice, illustrating how foundation models can enhance epigenetic marker detection while mitigating data sparsity. This work exemplifies a broader shift toward pretraining-based strategies in plant genomics, with implications for cross-species prediction and regulatory annotation. Taken together, the contributions in this Research Topic highlight the transformative role of machine learning and foundation models in plant functional genomics. By advancing representation learning, model architecture, interpretability, and multi-omics integration, these studies move the field beyond traditional sequence-based annotation toward predictive, mechanism-aware, and application-oriented frameworks. Continued synergy between computational innovation and experimental validation will be essential for translating these advances into resilient, high-yield crops capable of meeting future agricultural and environmental challenges.

This study evaluates the impact of an educational intervention aimed at improving the management of benign paroxysmal positional vertigo (BPPV) in a clinical setting. Interrupted Time Series (ITS) analysis was used to assess changes in the use of key diagnostic maneuvers—Dix–Hallpike and Supine Roll—as well as the treatment maneuver, Canalith Repositioning Maneuvers (CRM), and the use of CT scans. The intervention aimed to promote evidence-based practices and minimize unnecessary imaging. ITS models accounted for confounders such as history of BPPV and dizziness. The results showed significant improvements in the use of the Supine Roll and CRM maneuvers, with both immediate increases and sustained upward trends following the intervention. For the Dix–Hallpike maneuver, an initial increase was observed after the intervention; however, this effect was no longer statistically significant after adjusting for confounders, suggesting that the observed change may have been influenced by shifts in the patient population rather than the intervention itself. A positive long-term trend persisted. No significant reduction in CT scan usage was found. These findings demonstrate the intervention’s effectiveness in enhancing adherence to recommended diagnostic and treatment practices for BPPV. However, they also suggest that reducing unnecessary imaging may require additional strategies beyond provider education.

We know vanishingly little about how long-lived apes experience senescence in the wild, particularly with respect to their foraging behaviors. Chimpanzees use tools during foraging, and given the cognitive and physical challenges presented by tool use, tool-use behaviors are potentially at a heightened risk of senescence, though this has never been investigated in wild individuals. Accordingly, we sampled data from a longitudinal video archive that contained footage of wild chimpanzees using stone hammers and anvils to crack hard-shelled nuts (nut cracking) at an ‘outdoor laboratory’ over a 17-year period (with focal chimpanzees aging from approximately 39–44 to 56–61 years across this period). Over time, elderly chimpanzees began attending experimental nut-cracking sites less frequently than younger individuals. Several elderly chimpanzees exhibited reductions in efficiency across multiple stages of nut cracking, including taking longer to both select stone tools prior to use and use tools to crack open nuts and consume the associated pieces of kernel. Two chimpanzees began using less streamlined behavioral sequences to crack nuts, including a greater number of actions (such as more numerous hammer strikes). Notably, we report interindividual variability in the extent to which elderly chimpanzees’ tool-use behaviors changed during our sample period – ranging from small to profound reductions in engagement and efficiency – as well as differences in the specific aspects of nut cracking that changed for each individual. We discuss the possible causes of these changes – and recommendations for future research – with reference to literature surrounding the senescence of captive and wild primates.

This research investigates Computer-Assisted Audit Techniques (CAATTs) within the Internal Audit Unit at the City of Ekurhuleni Metropolitan Municipality (EMM). The study explores the extent to which CAATTs are utilized, the challenges faced in their implementation, and their impact on the efficiency and effectiveness of internal audits. Data was collected through a mixed-methods approach, including surveys and interviews with internal auditors. The findings reveal that while CAATTs, primarily TeamMate, are increasingly being used to automate audit processes, their adoption is hindered by limited access, insufficient training, and high costs. Despite these challenges, CAATTs have been shown to improve audit accuracy and efficiency, contributing to more timely and reliable audit outcomes. The study highlights the need for enhanced IT infrastructure, greater executive support, and comprehensive training to optimize the use of CAATTs. Based on these findings, the research provides several recommendations, including expanding the use of diverse CAATTs, investing in IT support, and fostering a culture of innovation to maximize the benefits of these technologies in internal auditing. This study contributes to the growing body of knowledge on CAATTs and offers practical insights for municipal internal audit units seeking to enhance their audit capabilities through technology.

Abstract Gamma synchrony is ubiquitous in visual cortex, but whether it contributes to perceptual grouping remains contentious based on observations that gamma frequency is not consistent across stimulus features and that gamma synchrony depends on distances between image elements. These stimulus dependencies have been argued to render synchrony among neural assemblies encoding components of the same object difficult. Alternatively, these dependencies may shape synchrony in meaningful ways. Using the theory of weakly coupled oscillators (TWCO), we demonstrate that stimulus dependence is crucial for gamma’s role in perception. Synchronization among coupled oscillators depends on frequency dissimilarity and coupling strength, which in early visual cortex relate to local feature dissimilarity and physical distance, respectively. We manipulated these factors in a texture segregation experiment wherein human observers identified the orientation of a figure defined by reduced contrast heterogeneity compared to the background. Human performance followed TWCO predictions both qualitatively and quantitatively, as formalized in a computational model. Moreover, we found that when enriched with a Hebbian learning rule, our model also predicted human learning effects. Increases in gamma synchrony due to perceptual learning predicted improvements in behavioral performance across sessions. This suggests that the stimulus-dependence of gamma synchrony is adaptable to the statistics of visual experiences, providing a viable neural grouping mechanism that can improve with visual experience. Together our results highlight the functional role of gamma synchrony in visual scene segmentation and provide a mechanistic explanation for its stimulus-dependent variability.

The shortage of nurses remains a persistent challenge in the U.S. healthcare system, largely driven by a scarcity of clinical training opportunities for nursing students. This underscores the urgent need for innovative approaches in nursing education. While serious games have shown effectiveness, they still face limitations, particularly in providing diverse experiences and overcoming resource constraints. Leveraging advances in artificial intelligence (AI), Nurse Town is a novel digital simulation game designed to enhance critical nursing competencies, including clinical decision-making, effective communication, and empathy. Using Large Language Models (LLMs), we developed virtual patient avatars capable of simulating medical scenarios and engaging in natural, context-aware conversations. The game also features an LLM-based automated performance assessment component. Compared to traditional training methods, Nurse Town significantly reduces the demand for educational resources while ensuring consistent training quality. We present a minimum viable product (MVP) featuring one virtual patient with hypertension. Future development plans include expanding scenario diversity, enhancing avatar realism, and integrating Virtual Reality (VR).

Purpose To develop a highly generalizable weakly supervised model to automatically detect and localize image-level intracranial hemorrhage (ICH) by using study-level labels. Materials and Methods In this retrospective study, the proposed model was pretrained on the image-level Radiological Society of North America dataset and fine-tuned on a local dataset by using attention-based bidirectional long short-term memory networks. This local training dataset included 10 699 noncontrast head CT scans in 7469 patients, with ICH study-level labels extracted from radiology reports. Model performance was compared with that of two senior neuroradiologists on 100 random test scans using the McNemar test, and its generalizability was evaluated on an external independent dataset. Results The model achieved a positive predictive value (PPV) of 85.7% (95% CI: 84.0, 87.4) and an area under the receiver operating characteristic curve of 0.96 (95% CI: 0.96, 0.97) on the held-out local test set (n = 7243, 3721 female) and 89.3% (95% CI: 87.8, 90.7) and 0.96 (95% CI: 0.96, 0.97), respectively, on the external test set (n = 491, 178 female). For 100 randomly selected samples, the model achieved performance on par with two neuroradiologists, but with a significantly faster (P < .05) diagnostic time of 5.04 seconds per scan (vs 86 seconds and 22.2 seconds for the two neuroradiologists, respectively). The model's attention weights and heatmaps visually aligned with neuroradiologists' interpretations. Conclusion The proposed model demonstrated high generalizability and high PPVs, offering a valuable tool for expedited ICH detection and prioritization while reducing false-positive interruptions in radiologists' workflows. Keywords: Computer-Aided Diagnosis (CAD), Brain/Brain Stem, Hemorrhage, Convolutional Neural Network (CNN), Transfer Learning Supplemental material is available for this article. © RSNA, 2024 See also the commentary by Akinci D'Antonoli and Rudie in this issue.

Ion channels are biological transistors that control ionic flux across cell membranes to regulate electrical transmission and signal transduction. They are found in all biological membranes and their conductive state kinetics are frequently disrupted in human diseases. Organelle ion channels are among the most resistant to functional and pharmacological interrogation. Traditional channel protein reconstitution methods rely upon exogenous expression and/or purification from endogenous cellular sources which are frequently contaminated by resident ionophores. Here we describe a fully synthetic method to assay functional properties of polycystin channels that natively traffic to primary cilia and endoplasmic reticulum organelles. Using this method, we characterize their oligomeric assembly, membrane integration, orientation and conductance while comparing these results to their endogenous channel properties. Outcomes define a novel synthetic approach that can be applied broadly to investigate channels resistant to biophysical analysis and pharmacological characterization.