NeuroImage最新文献_第7页

DeepMultiConnectome: Deep multi-task prediction of structural connectomes directly from diffusion MRI tractography DeepMultiConnectome：直接从扩散MRI束状图对结构连接体进行深度多任务预测。

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-29 DOI: 10.1016/j.neuroimage.2026.121765

Marcus J. Vroemen , Yuqian Chen , Yui Lo , Tengfei Xue , Weidong Cai , Fan Zhang , Josien P.W. Pluim , Lauren J. O'Donnell

Diffusion MRI (dMRI) tractography enables in vivo mapping of brain structural connections, but traditional connectome generation is time-consuming and requires gray matter parcellation, posing challenges for large-scale studies. We introduce DeepMultiConnectome, a deep-learning model that predicts structural connectomes directly from tractography, bypassing the need for gray matter parcellation while supporting multiple parcellation schemes. Using a point-cloud-based neural network with multi-task learning, the model classifies streamlines according to their connected regions across two parcellation schemes, sharing a learned representation. By classifying individual streamlines, our method’s output serves as a flexible prerequisite for constructing a wide range of differently weighted connectomes. We train and validate DeepMultiConnectome on tractography from the Human Connectome Project Young Adult dataset (N = 1000), labeled with an 84 and 164 region gray matter parcellation scheme. DeepMultiConnectome predicts multiple structural connectomes from a 3-million-streamline tractogram in ∼40 seconds. DeepMultiConnectome is evaluated by comparing predicted connectomes with traditional connectomes generated using the conventional method of labeling streamlines using a gray matter parcellation. The predicted connectomes show high agreement with traditionally generated connectomes across two parcellation schemes and multiple weighting strategies, and largely preserve network properties. Pearson correlations were r = 0.992 and 0.986 for streamline-count-weighted connectomes, r = 0.995 and 0.992 for SIFT2-weighted connectomes, and r = 0.775 and 0.727 for mean-FA-weighted connectomes. Test-retest analysis and downstream predictions of age and cognitive function demonstrate performance and reproducibility comparable to traditionally generated connectomes. Overall, DeepMultiConnectome provides a fast and scalable model for generating subject-specific connectomes across multiple parcellation and weighting schemes.

弥散MRI （Diffusion MRI, dMRI）神经束成像技术能够在体内绘制大脑结构连接，但传统的连接组生成耗时且需要对灰质进行分割，这给大规模研究带来了挑战。我们介绍了DeepMultiConnectome，这是一种深度学习模型，可以直接从神经束图预测结构连接体，绕过对灰质分割的需要，同时支持多种分割方案。该模型使用具有多任务学习功能的基于点云的神经网络，根据流线在两种分割方案中的连接区域对流线进行分类，共享学习表征。通过对单个流线进行分类，我们的方法的输出作为构建大范围不同加权连接体的灵活先决条件。我们在Human Connectome Project Young Adult数据集（n=1000）的神经束图上训练并验证了DeepMultiConnectome，该数据集使用84和164区域的灰质分割方案进行标记。DeepMultiConnectome在约40秒内从300万流线图中预测出多个结构连接体。DeepMultiConnectome通过比较预测的连接体与使用传统方法使用灰质包裹标记流线生成的传统连接体来评估。通过两种分割方案和多种加权策略，预测的连接体与传统生成的连接体高度一致，并且在很大程度上保留了网络特性。流线计数加权连接体的Pearson相关性r = 0.992和0.986，sift2加权连接体的r = 0.995和0.992，平均fa加权连接体的r = 0.775和0.727。测试-再测试分析和年龄和认知功能的下游预测证明了与传统生成的连接体相当的性能和可重复性。总的来说，DeepMultiConnectome提供了一个快速和可扩展的模型，用于跨多个分割和加权方案生成特定主题的连接体。

{"title":"DeepMultiConnectome: Deep multi-task prediction of structural connectomes directly from diffusion MRI tractography","authors":"Marcus J. Vroemen , Yuqian Chen , Yui Lo , Tengfei Xue , Weidong Cai , Fan Zhang , Josien P.W. Pluim , Lauren J. O'Donnell","doi":"10.1016/j.neuroimage.2026.121765","DOIUrl":"10.1016/j.neuroimage.2026.121765","url":null,"abstract":"<div><div>Diffusion MRI (dMRI) tractography enables in vivo mapping of brain structural connections, but traditional connectome generation is time-consuming and requires gray matter parcellation, posing challenges for large-scale studies. We introduce DeepMultiConnectome, a deep-learning model that predicts structural connectomes directly from tractography, bypassing the need for gray matter parcellation while supporting multiple parcellation schemes. Using a point-cloud-based neural network with multi-task learning, the model classifies streamlines according to their connected regions across two parcellation schemes, sharing a learned representation. By classifying individual streamlines, our method’s output serves as a flexible prerequisite for constructing a wide range of differently weighted connectomes. We train and validate DeepMultiConnectome on tractography from the Human Connectome Project Young Adult dataset (<em>N</em> = 1000), labeled with an 84 and 164 region gray matter parcellation scheme. DeepMultiConnectome predicts multiple structural connectomes from a 3-million-streamline tractogram in ∼40 seconds. DeepMultiConnectome is evaluated by comparing predicted connectomes with traditional connectomes generated using the conventional method of labeling streamlines using a gray matter parcellation. The predicted connectomes show high agreement with traditionally generated connectomes across two parcellation schemes and multiple weighting strategies, and largely preserve network properties. Pearson correlations were <em>r</em> = 0.992 and 0.986 for streamline-count-weighted connectomes, <em>r</em> = 0.995 and 0.992 for SIFT2-weighted connectomes, and <em>r</em> = 0.775 and 0.727 for mean-FA-weighted connectomes. Test-retest analysis and downstream predictions of age and cognitive function demonstrate performance and reproducibility comparable to traditionally generated connectomes. Overall, DeepMultiConnectome provides a fast and scalable model for generating subject-specific connectomes across multiple parcellation and weighting schemes.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"328 ","pages":"Article 121765"},"PeriodicalIF":4.5,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146097191","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Unmasking the bias: Can diffusion-weighted imaging reliably assess glymphatic function in awake and anesthetized brain? 揭露偏见：弥散加权成像能可靠地评估清醒和麻醉大脑的淋巴功能吗？

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-28 DOI: 10.1016/j.neuroimage.2026.121761

Ryszard Stefan Gomolka , Antonio Ladrón-de-Guevara , Søren Grubb , Lydiane Hirschler , Isabelle Strom , Pia Weikop , Matthias van Osch , Maiken Nedergaard , Yuki Mori

This study investigated whether magnetic resonance (MR) diffusivity parameters derived from diffusion-weighted imaging (DWI) can serve as biomarkers of glymphatic function in awake and anesthetized mice. Spectral apparent diffusion coefficient (ADC) analysis, obtained using an inverse Laplace transform, revealed that both Isoflurane (ISO) and Ketamine/Xylazine (K/X) anesthesia reduced the magnitude and range of diffusivities associated with interstitial fluid space (≤1 µm²/ms) compared with the awake state. Perfusion-related diffusivities (20–80 µm²/ms) increased under ISO but decreased under K/X. Monoexponential ADC and biexponential intravoxel incoherent motion (IVIM) modeling showed that ISO dose-dependently elevated, while K/X reduced, both slow and fast diffusivities across the brain. Dynamic contrast-enhanced MRI (DCE-MRI) indicated intermediate glymphatic influx in awake mice relative to both anesthetic conditions, without regional correlation to DWI-derived parameters. Perfusion micro–computed tomography (µCT) further demonstrated that ADC and IVIM metrics correlated regionally with mean transit time, suggesting a confounding by cerebral blood flow (CBF). Two-photon microscopy confirmed anesthesia-induced changes in cortical microvessel diameters consistent with perfusion alterations. Collectively, these findings indicate that MR diffusivity measures are strongly influenced by state-dependent physiological changes, and in particular perfusion. This represents an important limitation that warrants caution when using DWI to compare extracellular space, interstitial fluid flow, or glymphatic activity across different physiological or anesthetic states. Therefore, although technically challenging, we suggest that DWI studies aimed at assessing glymphatic or interstitial dynamics should be performed in awake conditions to minimize variability and anesthesia-related perfusion confounds across studies.

本研究探讨了由扩散加权成像（DWI）获得的磁共振（MR）扩散系数参数是否可以作为清醒和麻醉小鼠淋巴功能的生物标志物。利用拉普拉斯逆变换获得的光谱表观扩散系数（ADC）分析显示，与清醒状态相比，异氟醚（ISO）和氯胺酮/噻嗪（K/X）麻醉均降低了与间质液空间相关的扩散系数的大小和范围（≤1 μ m²/ms）。灌注相关扩散系数（20-80µm²/ms）在ISO下增加，但在K/X下下降。单指数ADC和双指数体内非相干运动（IVIM）模型显示，ISO剂量依赖性升高，而K/X降低，慢速和快速脑扩散率。动态对比增强MRI （DCE-MRI）显示，相对于两种麻醉状态，清醒小鼠的淋巴内流处于中等水平，与dwi衍生参数没有区域相关性。灌注微计算机断层扫描（µCT）进一步表明，ADC和IVIM指标与平均传输时间有区域相关性，提示受脑血流量（CBF）的影响。双光子显微镜证实麻醉引起的皮质微血管直径变化与灌注改变一致。总的来说，这些发现表明MR扩散率测量受到状态依赖性生理变化，特别是灌注的强烈影响。这是一个重要的局限性，在使用DWI比较不同生理或麻醉状态下的细胞外间隙、间质液流动或淋巴活性时需要谨慎。因此，尽管在技术上具有挑战性，但我们建议在清醒状态下进行旨在评估淋巴或间质动力学的DWI研究，以尽量减少研究间的变异性和麻醉相关灌注混淆。

{"title":"Unmasking the bias: Can diffusion-weighted imaging reliably assess glymphatic function in awake and anesthetized brain?","authors":"Ryszard Stefan Gomolka , Antonio Ladrón-de-Guevara , Søren Grubb , Lydiane Hirschler , Isabelle Strom , Pia Weikop , Matthias van Osch , Maiken Nedergaard , Yuki Mori","doi":"10.1016/j.neuroimage.2026.121761","DOIUrl":"10.1016/j.neuroimage.2026.121761","url":null,"abstract":"<div><div>This study investigated whether magnetic resonance (MR) diffusivity parameters derived from diffusion-weighted imaging (DWI) can serve as biomarkers of glymphatic function in awake and anesthetized mice. Spectral apparent diffusion coefficient (ADC) analysis, obtained using an inverse Laplace transform, revealed that both Isoflurane (ISO) and Ketamine/Xylazine (K/X) anesthesia reduced the magnitude and range of diffusivities associated with interstitial fluid space (≤1 µm²/ms) compared with the awake state. Perfusion-related diffusivities (20–80 µm²/ms) increased under ISO but decreased under K/X. Monoexponential ADC and biexponential intravoxel incoherent motion (IVIM) modeling showed that ISO dose-dependently elevated, while K/X reduced, both slow and fast diffusivities across the brain. Dynamic contrast-enhanced MRI (DCE-MRI) indicated intermediate glymphatic influx in awake mice relative to both anesthetic conditions, without regional correlation to DWI-derived parameters. Perfusion micro–computed tomography (µCT) further demonstrated that ADC and IVIM metrics correlated regionally with mean transit time, suggesting a confounding by cerebral blood flow (CBF). Two-photon microscopy confirmed anesthesia-induced changes in cortical microvessel diameters consistent with perfusion alterations. Collectively, these findings indicate that MR diffusivity measures are strongly influenced by state-dependent physiological changes, and in particular perfusion. This represents an important limitation that warrants caution when using DWI to compare extracellular space, interstitial fluid flow, or glymphatic activity across different physiological or anesthetic states. Therefore, although technically challenging, we suggest that DWI studies aimed at assessing glymphatic or interstitial dynamics should be performed in awake conditions to minimize variability and anesthesia-related perfusion confounds across studies.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"328 ","pages":"Article 121761"},"PeriodicalIF":4.5,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146093446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Detection of perivascular spaces at the gray-white matter interface using heavily T2-weighted MRI at 7T 7T重t2加权MRI检测灰质界面血管周围间隙

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-27 DOI: 10.1016/j.neuroimage.2026.121757

Gael Saib , Zeynep H. Demir , Paul A. Taylor , S. Lalith Talagala , Alan P. Koretsky

Background

There is increasing interest in high-contrast cerebrospinal fluid (CSF) MRI for imaging perivascular spaces (PVSs). Dilated PVSs, associated with aging, dementia, and various other conditions, are readily detected within the white matter (WM), basal ganglia, and midbrain. While 7T MRI enables detection of smaller PVSs, cortical PVS burden has received limited attention despite its potential value for understanding neurological conditions.

Purpose

To investigate the detectability of cortical PVS segments in healthy participants using heavily T₂-weighted MRI at 7T.

Materials and methods

A T₂-weighted 3D-TSE sequence was optimized at 7T to detect CSF with high resolution and contrast-to-noise ratio (CNR) while minimizing signal from surrounding tissues. A semi-automated pipeline was developed to extract PVSs and quantify their density in the whole brain, including the cortex.

Results

Seventeen healthy volunteers (40

\pm

14 years) were scanned at 7T. Optimized TSE achieved a CSF-to-tissue CNR of ∼180:1, enabling detection of small PVSs throughout the brain and leukocortical segments. About 20% of WM PVSs contained a leukocortical segment. WM PVSs with a leukocortical segment represented 70% of the total PVS volume. PVS density in the cortex was ∼0.7% (∼6-fold lower than WM), with highest in the insula and lowest in the auditory cortex.

Conclusion

High-resolution CSF imaging using optimized 3D-TSE MRI at 7T allows detection and quantification of leukocortical PVS segments at the gray-white matter interface in healthy individuals. This study lays the groundwork for exploring regional PVS changes related to the cortex and their potential use in diagnosis or prognosis of neurological diseases.

高对比脑脊液（CSF） MRI对血管周围间隙（pvs）成像的兴趣越来越大。与衰老、痴呆和各种其他疾病相关的扩张性PVSs在白质（WM）、基底神经节和中脑内很容易被检测到。虽然7T MRI可以检测较小的PVS，但皮质PVS负荷受到的关注有限，尽管它对了解神经系统疾病有潜在价值。目的探讨7T重t2加权MRI对健康受试者皮质PVS节段的可探测性。材料和方法在7T时优化sa t2加权3D-TSE序列，以高分辨率和高噪比（CNR）检测脑脊液，同时最大限度地减少周围组织的信号。开发了一种半自动管道来提取pvs并量化其在整个大脑（包括皮层）中的密度。结果17例健康志愿者（40±14岁）7T扫描。优化后的TSE实现了CSF-to-tissue的CNR为180:1，能够检测整个大脑和白皮质段的小PVSs。约20%的WM PVSs含有白质皮层段。伴白脑皮层段的WM PVS占总PVS体积的70%。皮层的PVS密度为~ 0.7%（比WM低~ 6倍），其中岛叶最高，听觉皮层最低。结论采用优化的3D-TSE MRI在7T时进行高分辨率脑脊液成像，可以检测和定量健康个体灰质-白质界面白质皮层PVS段。本研究为探索与皮层相关的PVS区域变化及其在神经系统疾病诊断和预后中的潜在应用奠定了基础。

{"title":"Detection of perivascular spaces at the gray-white matter interface using heavily T2-weighted MRI at 7T","authors":"Gael Saib , Zeynep H. Demir , Paul A. Taylor , S. Lalith Talagala , Alan P. Koretsky","doi":"10.1016/j.neuroimage.2026.121757","DOIUrl":"10.1016/j.neuroimage.2026.121757","url":null,"abstract":"<div><h3>Background</h3><div>There is increasing interest in high-contrast cerebrospinal fluid (CSF) MRI for imaging perivascular spaces (PVSs). Dilated PVSs, associated with aging, dementia, and various other conditions, are readily detected within the white matter (WM), basal ganglia, and midbrain. While 7T MRI enables detection of smaller PVSs, cortical PVS burden has received limited attention despite its potential value for understanding neurological conditions.</div></div><div><h3>Purpose</h3><div>To investigate the detectability of cortical PVS segments in healthy participants using heavily T<sub>2</sub>-weighted MRI at 7T.</div></div><div><h3>Materials and methods</h3><div>A T<sub>2</sub>-weighted 3D-TSE sequence was optimized at 7T to detect CSF with high resolution and contrast-to-noise ratio (CNR) while minimizing signal from surrounding tissues. A semi-automated pipeline was developed to extract PVSs and quantify their density in the whole brain, including the cortex.</div></div><div><h3>Results</h3><div>Seventeen healthy volunteers (40<span><math><mo>±</mo></math></span>14 years) were scanned at 7T. Optimized TSE achieved a CSF-to-tissue CNR of ∼180:1, enabling detection of small PVSs throughout the brain and leukocortical segments. About 20% of WM PVSs contained a leukocortical segment. WM PVSs with a leukocortical segment represented 70% of the total PVS volume. PVS density in the cortex was ∼0.7% (∼6-fold lower than WM), with highest in the insula and lowest in the auditory cortex.</div></div><div><h3>Conclusion</h3><div>High-resolution CSF imaging using optimized 3D-TSE MRI at 7T allows detection and quantification of leukocortical PVS segments at the gray-white matter interface in healthy individuals. This study lays the groundwork for exploring regional PVS changes related to the cortex and their potential use in diagnosis or prognosis of neurological diseases.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121757"},"PeriodicalIF":4.5,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146079732","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Converse or reverse? Machine-learning modeling for disease progression: A study based on Alzheimer’s disease continuum cohort 反向还是反向？疾病进展的机器学习建模：基于阿尔茨海默病连续队列的研究。

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-25 DOI: 10.1016/j.neuroimage.2026.121754

Yujing Huang (黄玉晶) , Hao Zhang (张灏) , Buqing Ma (马步青) , Zhe Yu (俞哲) , Shenyi Dai (戴珅懿) , Lu Cheng (程璐) , Li Su (苏里) , Alzheimer’s Disease Neuroimaging Initiative (ADNI), Gaoyi Yang (杨高怡) , Qingguo Ma (马庆国)

Introduction

Longitudinal trajectories from healthy aging to Mild Cognitive Impairment and Alzheimer’s Disease involve complex mechanisms.

Methods

We evaluated five machine learning approaches (Random Forest, Support Vector Machines, Radial Basis Function Networks, Backpropagation Networks, Convolutional Neural Network) to assess the importance of potential predictive markers across the health-to-dementia continuum. Using the ADNI cohort across four phases (ADNI1, ADNIGO, ADNI2, ADNI3), we analyzed participants with distinct trajectories: stable, convertible, and reverse progression.

Results

Random Forest outperformed other models across key effectiveness metrics and achieved a macro-averaged sensitivity of 70.8 % and specificity of 96.8 % across all participant groups. Random Forest identified visuospatial and memory-related cognitive dysfunction as key predictive clinical features and several amyloid-related neuroimaging biomarkers — including temporal variations of amyloid uptake within inferior lateral ventricles, para-hippocampus—for classifying participant groups. Additionally, plasma APOE4 and long neurofilament light chain levels emerged as promising predictors for tracking progression.

Conclusion

These findings highlight the potential of machine learning in classifying disease trajectories.

从健康衰老到轻度认知障碍和阿尔茨海默病的纵向轨迹涉及复杂的机制。方法：我们评估了五种机器学习方法（随机森林、支持向量机、径向基函数网络、反向传播网络、卷积神经网络），以评估健康到痴呆连续体中潜在预测标记的重要性。使用四个阶段（ADNI1, adnio, ADNI2, ADNI3）的ADNI队列，我们分析了具有不同轨迹的参与者：稳定，可转换和反向进展。结果：随机森林在关键有效性指标上优于其他模型，在所有参与者组中实现了70.8%的宏观平均灵敏度和96.8%的特异性。随机森林将视觉空间和记忆相关的认知功能障碍确定为关键的预测性临床特征和几种淀粉样蛋白相关的神经成像生物标志物，包括下侧脑室、海马体旁淀粉样蛋白摄取的时间变化，用于对参与者群体进行分类。此外，血浆APOE4和长神经丝轻链水平成为跟踪进展的有希望的预测指标。结论：这些发现突出了机器学习在分类疾病轨迹方面的潜力。

{"title":"Converse or reverse? Machine-learning modeling for disease progression: A study based on Alzheimer’s disease continuum cohort","authors":"Yujing Huang (黄玉晶) , Hao Zhang (张灏) , Buqing Ma (马步青) , Zhe Yu (俞哲) , Shenyi Dai (戴珅懿) , Lu Cheng (程璐) , Li Su (苏里) , Alzheimer’s Disease Neuroimaging Initiative (ADNI), Gaoyi Yang (杨高怡) , Qingguo Ma (马庆国)","doi":"10.1016/j.neuroimage.2026.121754","DOIUrl":"10.1016/j.neuroimage.2026.121754","url":null,"abstract":"<div><h3>Introduction</h3><div>Longitudinal trajectories from healthy aging to Mild Cognitive Impairment and Alzheimer’s Disease involve complex mechanisms.</div></div><div><h3>Methods</h3><div>We evaluated five machine learning approaches (Random Forest, Support Vector Machines, Radial Basis Function Networks, Backpropagation Networks, Convolutional Neural Network) to assess the importance of potential predictive markers across the health-to-dementia continuum. Using the ADNI cohort across four phases (ADNI1, ADNIGO, ADNI2, ADNI3), we analyzed participants with distinct trajectories: stable, convertible, and reverse progression.</div></div><div><h3>Results</h3><div>Random Forest outperformed other models across key effectiveness metrics and achieved a macro-averaged sensitivity of 70.8 % and specificity of 96.8 % across all participant groups. Random Forest identified visuospatial and memory-related cognitive dysfunction as key predictive clinical features and several amyloid-related neuroimaging biomarkers — including temporal variations of amyloid uptake within inferior lateral ventricles, para-hippocampus—for classifying participant groups. Additionally, plasma APOE4 and long neurofilament light chain levels emerged as promising predictors for tracking progression.</div></div><div><h3>Conclusion</h3><div>These findings highlight the potential of machine learning in classifying disease trajectories.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121754"},"PeriodicalIF":4.5,"publicationDate":"2026-01-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146065352","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Personalized repetitive transcranial magnetic stimulation reduces frontal eeg complexity in patients with obsessive–compulsive disorder 个性化重复经颅磁刺激降低强迫症患者额叶脑电图复杂性。

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-23 DOI: 10.1016/j.neuroimage.2026.121751

Yihao Sheng , Woxin Pan , Haibing Huang , Shenghao Geng , Fangyan Jia , Liyv Lu , Ke Chen , Chunyan Zhu , Dandan Li

Background

Repetitive transcranial magnetic stimulation (rTMS) shows therapeutic potential for obsessive-compulsive disorder (OCD). Brain entropy has recently emerged as a candidate biomarker in neuropsychiatry, yet its modulation by rTMS in OCD remains unclear. Given EEG's superior temporal resolution for capturing rapid fluctuations in neural complexity, it was used to evaluate the effects of fMRI-neuronavigated rTMS on frontal entropy and its potential as an objective treatment marker.

Methods

Resting-state EEG was recorded from 44 OCD patients and 24 healthy controls (HCs) to compute frontal entropy- and complexity-based measures, including approximate entropy (ApEn), sample entropy (SampEn), and Lempel–Ziv complexity (LZC). Patients were randomized to an active (n = 22) or sham (n = 22) rTMS group, with the active group receiving individualized 1 Hz stimulation over the right pre-supplementary motor area for 14 consecutive days. EEG was repeated post-intervention.

Results

At baseline, OCD patients exhibited higher frontal complexity than healthy controls across all three measures. Linear mixed-effects models consistently revealed significant main effects of time and stimulation, as well as their interaction. Bayesian and FDR-corrected analyses confirmed significant reductions in all three measures following active stimulation. Post-treatment, frontal complexity remained elevated in the sham group relative to healthy controls, whereas no such difference was observed in the active stimulation group.

Conclusion

OCD is characterized by increased frontal neural complexity as indexed by multiple entropy- and complexity-based EEG measures. Individualized rTMS modulated these abnormalities, supporting frontal EEG complexity as a promising objective biomarker of neuromodulatory effects.

背景：重复经颅磁刺激（rTMS）显示出治疗强迫症（OCD）的潜力。脑熵最近成为神经精神病学的候选生物标志物，但rTMS对强迫症的调节作用尚不清楚。鉴于脑电图在捕捉神经复杂性的快速波动方面具有优越的时间分辨率，我们利用它来评估fmri神经导航rTMS对额叶熵的影响及其作为客观治疗指标的潜力。方法：记录44例强迫症患者和24例健康对照（hc）的静息状态脑电图，计算基于额叶熵和复杂度的测量，包括近似熵（ApEn）、样本熵（SampEn）和Lempel-Ziv复杂度（LZC）。患者被随机分为活跃（n = 22）或假（n = 22）rTMS组，活跃组在右侧辅助前运动区域连续14天接受个体化1hz刺激。干预后复查脑电图。结果：在基线时，强迫症患者在所有三项测量中表现出比健康对照组更高的额叶复杂性。线性混合效应模型一致地揭示了时间和刺激的显著主效应及其相互作用。贝叶斯和fdr校正分析证实，在积极增产后，这三种措施的产量都显著降低。治疗后，假手术组的额叶复杂性仍然高于健康对照组，而在积极刺激组中没有观察到这种差异。结论：基于多重熵和复杂性的脑电图测量表明，强迫症的特征是额神经复杂性增加。个体化rTMS调节了这些异常，支持额叶脑电图复杂性作为神经调节作用的一个有希望的客观生物标志物。

{"title":"Personalized repetitive transcranial magnetic stimulation reduces frontal eeg complexity in patients with obsessive–compulsive disorder","authors":"Yihao Sheng , Woxin Pan , Haibing Huang , Shenghao Geng , Fangyan Jia , Liyv Lu , Ke Chen , Chunyan Zhu , Dandan Li","doi":"10.1016/j.neuroimage.2026.121751","DOIUrl":"10.1016/j.neuroimage.2026.121751","url":null,"abstract":"<div><h3>Background</h3><div>Repetitive transcranial magnetic stimulation (rTMS) shows therapeutic potential for obsessive-compulsive disorder (OCD). Brain entropy has recently emerged as a candidate biomarker in neuropsychiatry, yet its modulation by rTMS in OCD remains unclear. Given EEG's superior temporal resolution for capturing rapid fluctuations in neural complexity, it was used to evaluate the effects of fMRI-neuronavigated rTMS on frontal entropy and its potential as an objective treatment marker.</div></div><div><h3>Methods</h3><div>Resting-state EEG was recorded from 44 OCD patients and 24 healthy controls (HCs) to compute frontal entropy- and complexity-based measures, including approximate entropy (ApEn), sample entropy (SampEn), and Lempel–Ziv complexity (LZC). Patients were randomized to an active (n = 22) or sham (n = 22) rTMS group, with the active group receiving individualized 1 Hz stimulation over the right pre-supplementary motor area for 14 consecutive days. EEG was repeated post-intervention.</div></div><div><h3>Results</h3><div>At baseline, OCD patients exhibited higher frontal complexity than healthy controls across all three measures. Linear mixed-effects models consistently revealed significant main effects of time and stimulation, as well as their interaction. Bayesian and FDR-corrected analyses confirmed significant reductions in all three measures following active stimulation. Post-treatment, frontal complexity remained elevated in the sham group relative to healthy controls, whereas no such difference was observed in the active stimulation group.</div></div><div><h3>Conclusion</h3><div>OCD is characterized by increased frontal neural complexity as indexed by multiple entropy- and complexity-based EEG measures. Individualized rTMS modulated these abnormalities, supporting frontal EEG complexity as a promising objective biomarker of neuromodulatory effects.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121751"},"PeriodicalIF":4.5,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146046675","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Longitudinal P2X7R and myelin PET reveal distinct neuroinflammatory and white matter signatures compared with TSPO PET and DTI-MRI in the TgF344-AD rat model 在TgF344-AD大鼠模型中，与TSPO PET和DTI-MRI相比，纵向P2X7R和髓磷脂PET显示出不同的神经炎症和白质特征。

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-23 DOI: 10.1016/j.neuroimage.2026.121753

Oscar Moreno , Izaro Fernández , Sandra Plaza-García , Daniel Padro , Zuriñe Baz , Pedro Ramos-Cabrer , Abraham Martín , Jordi Llop

Emerging evidence links glial activation, particularly microglia, to Alzheimer’s disease (AD) progression. While TSPO PET (positron emission tomography) imaging detects neuroinflammation, its limitations drive interest in alternative targets like the P2X7 receptor. Myelin loss, potentially tied to chronic inflammation, is increasingly recognized as a key hallmark in AD, though the timing and relationship between neuroinflammation and demyelination remain poorly understood.

We conducted a longitudinal PET study from 4 to 22 months of age in TgF344-AD rats and wild-type controls to assess neuroinflammation with [¹⁸F]JNJ-64413739 (P2X7R) and [¹⁸F]DPA-714 (TSPO) only at 22 months, alongside myelin content using [¹⁸F]Florbetaben. Diffusion tensor imaging (DTI) was used to study variations on myelin structure in old AD and WT rats. In vitro studies, including autoradiography, immunofluorescence and staining were used to support the in vivo results.

[¹⁸F]JNJ-64413739 PET showed increased P2X7 receptor expression in AD and control animals over time, while [¹⁸F]DPA-714 PET showed significant differences between groups at 22 months. [¹⁸F]Florbetaben PET showed different uptake in white matter rich areas between groups with observed demyelination in AD rats at 20 months in the brain stem, supported by diffusional MRI findings.

In our study, P2X7R overexpression was attributed to aging rather than genotype effects, and no link was found to the observed demyelination in AD rats. Conversely, increased TSPO neuroinflammation in TgF344-AD rats correlated with myelin loss and the reported cognitive decline in this model. Our results support the use of the TgF344-AD model to study early AD pathology, focusing on neuroinflammation and white matter integrity.

新出现的证据表明，神经胶质细胞的激活，特别是小胶质细胞，与阿尔茨海默病（AD）的进展有关。虽然TSPO PET（正电子发射断层扫描）成像检测神经炎症，但其局限性促使人们对P2X7受体等替代靶标感兴趣。髓磷脂丢失，可能与慢性炎症有关，越来越被认为是阿尔茨海默病的一个关键标志，尽管神经炎症和脱髓鞘之间的时间和关系仍然知之甚少。我们在TgF344-AD大鼠和野生型对照中进行了4至22月龄的纵向PET研究，仅在22月龄时使用[18F]JNJ-64413739 （P2X7R）和[18F]DPA-714 （TSPO）评估神经炎症，同时使用[18F]Florbetaben评估髓磷脂含量。采用弥散张量成像（DTI）研究老年AD和WT大鼠髓鞘结构的变化。体外研究，包括放射自显影，免疫荧光和染色被用来支持体内的结果。[18F]JNJ-64413739 PET显示，随着时间的推移，AD和对照动物的P2X7受体表达增加，而[18F]DPA-714 PET在22个月时组间差异显著。[18F]弥散性MRI结果支持了20月龄AD大鼠脑干脱髓鞘组间富白质区Florbetaben PET摄取的差异。在我们的研究中，P2X7R过表达归因于衰老而不是基因型效应，并且没有发现与AD大鼠观察到的脱髓鞘有关。相反，TgF344-AD大鼠的TSPO神经炎症增加与髓磷脂丢失和该模型中报道的认知能力下降相关。我们的研究结果支持使用TgF344-AD模型来研究早期AD病理，重点关注神经炎症和白质完整性。

{"title":"Longitudinal P2X7R and myelin PET reveal distinct neuroinflammatory and white matter signatures compared with TSPO PET and DTI-MRI in the TgF344-AD rat model","authors":"Oscar Moreno , Izaro Fernández , Sandra Plaza-García , Daniel Padro , Zuriñe Baz , Pedro Ramos-Cabrer , Abraham Martín , Jordi Llop","doi":"10.1016/j.neuroimage.2026.121753","DOIUrl":"10.1016/j.neuroimage.2026.121753","url":null,"abstract":"<div><div>Emerging evidence links glial activation, particularly microglia, to Alzheimer’s disease (AD) progression. While TSPO PET (positron emission tomography) imaging detects neuroinflammation, its limitations drive interest in alternative targets like the P2X7 receptor. Myelin loss, potentially tied to chronic inflammation, is increasingly recognized as a key hallmark in AD, though the timing and relationship between neuroinflammation and demyelination remain poorly understood.</div><div>We conducted a longitudinal PET study from 4 to 22 months of age in TgF344-AD rats and wild-type controls to assess neuroinflammation with [<sup>18</sup>F]JNJ-64413739 (P2X7R) and [<sup>18</sup>F]DPA-714 (TSPO) only at 22 months, alongside myelin content using [<sup>18</sup>F]Florbetaben. Diffusion tensor imaging (DTI) was used to study variations on myelin structure in old AD and WT rats. <em>In vitro</em> studies, including autoradiography, immunofluorescence and staining were used to support the <em>in vivo</em> results.</div><div>[<sup>18</sup>F]JNJ-64413739 PET showed increased P2X7 receptor expression in AD and control animals over time, while [<sup>18</sup>F]DPA-714 PET showed significant differences between groups at 22 months. [<sup>18</sup>F]Florbetaben PET showed different uptake in white matter rich areas between groups with observed demyelination in AD rats at 20 months in the brain stem, supported by diffusional MRI findings.</div><div>In our study, P2X7R overexpression was attributed to aging rather than genotype effects, and no link was found to the observed demyelination in AD rats. Conversely, increased TSPO neuroinflammation in TgF344-AD rats correlated with myelin loss and the reported cognitive decline in this model. Our results support the use of the TgF344-AD model to study early AD pathology, focusing on neuroinflammation and white matter integrity.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121753"},"PeriodicalIF":4.5,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146046383","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Short-term dynamic changes in interbrain synchrony during first social interaction between strangers 陌生人初次社交互动中脑间同步的短期动态变化。

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-22 DOI: 10.1016/j.neuroimage.2026.121752

Dovrat Atias , Andrey Markus , Simone Shamay-Tsoory

Emerging evidence highlights the role of interbrain synchrony in fostering the formation of relationships. Yet a key question remains: how does interbrain synchrony dynamically evolve throughout a first meeting between strangers, and to what extent do interbrain networks exhibit dynamic changes during the interaction?

In this study, we investigate short-term changes in behavioral and interbrain synchrony during initial face-to-face encounters between strangers. To assess experience-dependent changes in interbrain synchrony, we tracked these dynamics with functional near-infrared spectroscopy (fNIRS) over a five-minute interaction period, and examined whether these changes predict changes in movement synchronization. A total of 106 previously unacquainted participants were randomly paired into dyads and instructed to engage in a five-minute acquaintance conversation. Results indicate that movement synchronization increased throughout the interaction and predicted dyadic connectedness. The neuroimaging findings revealed heightened interbrain synchrony in the dorsolateral prefrontal cortex (dlPFC), inferior frontal gyrus (IFG), and premotor cortex (PMC). Critically, we observed a gradual increase in interbrain synchrony in the l.dlPFC- l.IFG and the r.PMC-r.IFG. Finally, we showed that interbrain synchrony in the r.PMC-r.IFG predicted movement synchronization only at the end of the meeting. These findings position interbrain synchrony as a rapidly plastic, experience-dependent coupling mechanism that emerges during first-time natural conversation, rather than a static correlate of shared state. It shows that minute-by-minute increases in behavioral and neural alignment play a foundational role in the formation of social bonds, potentially through motor pathways.

新出现的证据强调了脑间同步在促进关系形成中的作用。然而，一个关键的问题仍然存在：在陌生人的第一次会面中，脑间同步是如何动态演变的？在互动过程中，脑间网络在多大程度上表现出动态变化？在这项研究中，我们调查了陌生人初次面对面接触时行为和脑间同步的短期变化。为了评估脑间同步的经验依赖变化，我们使用功能性近红外光谱（fNIRS）在5分钟的相互作用期间跟踪这些动态，并检查这些变化是否预测运动同步的变化。共有106名之前不认识的参与者被随机分成两组，并被要求进行五分钟的熟人交谈。结果表明，运动同步在整个交互过程中增加，并预测了二元连接。神经影像学结果显示，背外侧前额叶皮层（dlPFC）、额下回（IFG）和运动前皮层（PMC）的脑间同步增强。关键的是，我们观察到ldlpfc - lifg和rpmc - rifg的脑间同步逐渐增加。最后，我们发现了rpmc -r的脑间同步。IFG仅在会议结束时预测运动同步。这些发现表明，脑间同步是一种快速可塑的、经验依赖的耦合机制，它出现在第一次自然对话中，而不是共享状态的静态相关。它表明，行为和神经一致性的分分秒秒增加，在社会纽带的形成中起着基础作用，可能是通过运动途径形成的。

{"title":"Short-term dynamic changes in interbrain synchrony during first social interaction between strangers","authors":"Dovrat Atias , Andrey Markus , Simone Shamay-Tsoory","doi":"10.1016/j.neuroimage.2026.121752","DOIUrl":"10.1016/j.neuroimage.2026.121752","url":null,"abstract":"<div><div>Emerging evidence highlights the role of interbrain synchrony in fostering the formation of relationships. Yet a key question remains: how does interbrain synchrony dynamically evolve throughout a first meeting between strangers, and to what extent do interbrain networks exhibit dynamic changes during the interaction?</div><div>In this study, we investigate short-term changes in behavioral and interbrain synchrony during initial face-to-face encounters between strangers. To assess experience-dependent changes in interbrain synchrony, we tracked these dynamics with functional near-infrared spectroscopy (fNIRS) over a five-minute interaction period, and examined whether these changes predict changes in movement synchronization. A total of 106 previously unacquainted participants were randomly paired into dyads and instructed to engage in a five-minute acquaintance conversation. Results indicate that movement synchronization increased throughout the interaction and predicted dyadic connectedness. The neuroimaging findings revealed heightened interbrain synchrony in the dorsolateral prefrontal cortex (dlPFC), inferior frontal gyrus (IFG), and premotor cortex (PMC). Critically, we observed a gradual increase in interbrain synchrony in the l.dlPFC- l.IFG and the r.PMC-r.IFG. Finally, we showed that interbrain synchrony in the r.PMC-r.IFG predicted movement synchronization only at the end of the meeting. These findings position interbrain synchrony as a rapidly plastic, experience-dependent coupling mechanism that emerges during first-time natural conversation, rather than a static correlate of shared state. It shows that minute-by-minute increases in behavioral and neural alignment play a foundational role in the formation of social bonds, potentially through motor pathways.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121752"},"PeriodicalIF":4.5,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146044039","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Temporal evolution of neural codes: The added value of a geometric approach to linear coefficients 神经编码的时间演化：线性系数几何方法的附加价值。

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-22 DOI: 10.1016/j.neuroimage.2026.121737

Théo Desbordes , Itsaso Olasagasti , Nicolas Piron , Sophie Schwartz , Nina Kazanina

Multivariate decoding analyses have become a cornerstone method in cognitive neuroscience. When applied to time-resolved brain imaging signals, they provide insights into the temporal dynamics of information processing in the brain. In particular, the temporal generalization (TG) method—where a decoder trained at one time point is tested on others—is commonly used to assess the stability of neural representations over time. However, TG performance can be ambiguous: distinct representational dynamics—such as sparse versus distributed activity, or scaling of activity versus recruitment of new units—can yield similar TG matrices. Moreover, even when generalization is strong, underlying neural representations may still be evolving in ways that TG alone fails to reveal. This ambiguity of performance profiles can mask meaningful changes in the geometry of neural representations. In this study, we use controlled simulations to demonstrate how different dynamic processes can produce indistinguishable TG profiles. To resolve these ambiguities, we propose a complementary approach based on the geometry of the learned linear coefficients. Specifically, we quantify the Rotation Angle θ between decision subspaces (with cosine similarity) and the Feature Density

α

(capturing whether feature contributions are distributed or sparse). Together, these measures complement TG analyses, revealing how neural representations evolve in space and time. Beyond time-resolved decoding, our approach applies broadly to any linear model, offering a geometric perspective on representational dynamics.

多元解码分析已成为认知神经科学的基础方法。当应用于时间分辨率的大脑成像信号时，它们提供了对大脑信息处理的时间动态的见解。特别是，时间泛化（TG）方法-在一个时间点训练的解码器在其他时间点进行测试-通常用于评估神经表征随时间的稳定性。然而，TG性能可能是模糊的：不同的表征动态—例如稀疏与分布式活动，或活动的缩放与新单元的招募—可以产生相似的TG矩阵。此外，即使泛化很强，潜在的神经表征可能仍在以单TG无法揭示的方式进化。性能概况的这种模糊性可以掩盖神经表征几何结构中有意义的变化。在这项研究中，我们使用控制模拟来证明不同的动态过程如何产生难以区分的TG剖面。为了解决这些歧义，我们提出了一种基于学习到的线性系数几何的互补方法。具体来说，我们量化了决策子空间（具有余弦相似度）和特征密度α（捕获特征贡献是分布的还是稀疏的）之间的旋转角θ。总之，这些措施补充了TG分析，揭示了神经表征如何在空间和时间上进化。除了时间分辨解码之外，我们的方法广泛适用于任何线性模型，提供了表征动态的几何视角。

{"title":"Temporal evolution of neural codes: The added value of a geometric approach to linear coefficients","authors":"Théo Desbordes , Itsaso Olasagasti , Nicolas Piron , Sophie Schwartz , Nina Kazanina","doi":"10.1016/j.neuroimage.2026.121737","DOIUrl":"10.1016/j.neuroimage.2026.121737","url":null,"abstract":"<div><div>Multivariate decoding analyses have become a cornerstone method in cognitive neuroscience. When applied to time-resolved brain imaging signals, they provide insights into the temporal dynamics of information processing in the brain. In particular, the temporal generalization (TG) method—where a decoder trained at one time point is tested on others—is commonly used to assess the stability of neural representations over time. However, TG performance can be ambiguous: distinct representational dynamics—such as sparse versus distributed activity, or scaling of activity versus recruitment of new units—can yield similar TG matrices. Moreover, even when generalization is strong, underlying neural representations may still be evolving in ways that TG alone fails to reveal. This ambiguity of performance profiles can mask meaningful changes in the geometry of neural representations. In this study, we use controlled simulations to demonstrate how different dynamic processes can produce indistinguishable TG profiles. To resolve these ambiguities, we propose a complementary approach based on the geometry of the learned linear coefficients. Specifically, we quantify the <strong>Rotation Angle θ</strong> between decision subspaces (with cosine similarity) and the <strong>Feature Density</strong> <span><math><mrow><mi>α</mi></mrow></math></span> (capturing whether feature contributions are distributed or sparse). Together, these measures complement TG analyses, revealing how neural representations evolve in space and time. Beyond time-resolved decoding, our approach applies broadly to any linear model, offering a geometric perspective on representational dynamics.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121737"},"PeriodicalIF":4.5,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146044059","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Temporal dynamics of letter processing revealed by multivariate pattern analysis of EEG data 脑电数据多变量模式分析揭示字母处理的时间动态

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-22 DOI: 10.1016/j.neuroimage.2026.121750

Miguel Domingues , Susana Araújo , Tânia Fernandes , Inês Bramão

Letters are the primitives of reading expertise. Single letter recognition relies on a hierarchy of processing stages, in which early visual features gradually evolve into abstract letter representations, but the temporal organization of these stages remains poorly understood. To address it, we applied multivariate pattern analysis (MVPA) to electroencephalography (EEG) data recorded while adult readers (n = 35) performed a one-back repetition detection task on single letters and pseudoletters. Traditional event-related potential (ERP) analyses revealed differences between letters and pseudoletters in the N1 (140–170 ms), P2 (210–270 ms), and P3 (300–500 ms) components. Multivariate temporal generalization analyses showed that neural patterns distinguishing letters from pseudoletters were highly generalizable from approximately 140 to 600 ms after stimulus onset. A spatiotemporal searchlight analysis indicated that, despite this temporal generalization, the topographic configuration of EEG channels contributing to classification changed along this window, suggesting that neural representations in later processing stages were transformed from earlier perceptual stages. These findings indicate that letter recognition unfolds as a cascade of continuous and interacting processes rather than via discrete stages. Early perceptual letter-specific activity, indexed by the N1 component, remains engaged throughout later, increasingly abstract, orthographic processing stages to jointly support letter identification.

字母是阅读技巧的基础。单字母识别依赖于处理阶段的层次结构，其中早期的视觉特征逐渐演变为抽象的字母表示，但这些阶段的时间组织仍然知之甚少。为了解决这一问题，我们将多变量模式分析（MVPA）应用于脑电图（EEG）数据，同时成人读者（n = 35）对单个字母和伪字母进行单回重复检测任务。传统的事件相关电位（ERP）分析揭示了字母和伪字母在N1 （140-170 ms）、P2 （210-270 ms）和P3 （300-500 ms）分量上的差异。多元时间泛化分析表明，在刺激开始后约140 - 600 ms内，区分字母和伪字母的神经模式具有高度的泛化性。时空探照灯分析表明，尽管存在这种时间概化，但有助于分类的脑电通道的地形配置沿着这一窗口发生了变化，这表明后期处理阶段的神经表征是从早期感知阶段转变而来的。这些发现表明，字母识别是一个连续的、相互作用的过程，而不是通过离散的阶段展开的。早期的感知字母特定活动，由N1分量索引，在后来越来越抽象的正字法处理阶段保持参与，共同支持字母识别。

{"title":"Temporal dynamics of letter processing revealed by multivariate pattern analysis of EEG data","authors":"Miguel Domingues , Susana Araújo , Tânia Fernandes , Inês Bramão","doi":"10.1016/j.neuroimage.2026.121750","DOIUrl":"10.1016/j.neuroimage.2026.121750","url":null,"abstract":"<div><div>Letters are the primitives of reading expertise. Single letter recognition relies on a hierarchy of processing stages, in which early visual features gradually evolve into abstract letter representations, but the temporal organization of these stages remains poorly understood. To address it, we applied multivariate pattern analysis (MVPA) to electroencephalography (EEG) data recorded while adult readers (<em>n</em> = 35) performed a one-back repetition detection task on single letters and pseudoletters. Traditional event-related potential (ERP) analyses revealed differences between letters and pseudoletters in the N1 (140–170 ms), P2 (210–270 ms), and P3 (300–500 ms) components. Multivariate temporal generalization analyses showed that neural patterns distinguishing letters from pseudoletters were highly generalizable from approximately 140 to 600 ms after stimulus onset. A spatiotemporal searchlight analysis indicated that, despite this temporal generalization, the topographic configuration of EEG channels contributing to classification changed along this window, suggesting that neural representations in later processing stages were transformed from earlier perceptual stages. These findings indicate that letter recognition unfolds as a cascade of continuous and interacting processes rather than via discrete stages. Early perceptual letter-specific activity, indexed by the N1 component, remains engaged throughout later, increasingly abstract, orthographic processing stages to jointly support letter identification.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121750"},"PeriodicalIF":4.5,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146025228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The evaluation of DUNE: a U-Net-based neural network to denoise multi-echo fMRI data 基于u - net的神经网络对多回波fMRI数据去噪的评价

IF 4.5 2区医学 Q1 NEUROIMAGING

NeuroImage

Pub Date : 2026-01-22 DOI: 10.1016/j.neuroimage.2026.121749

Peter Van Schuerbeek , Manon Roose , Alina Monica Ionescu , Maarten Naeyaert , Hubert Raeymaekers

Task based fMRI data suffers from scanner and physiologic noise. Consequently, finding the task based BOLD responses out of the noise is challenging. To improve the power to detect the BOLD responses, multi-echo (ME) fMRI combined with ICA based denoising (MEICA) and single-echo (SE) fMRI at high temporal resolution (<1 s) have been introduced. Both techniques have been found to give better activation maps than a traditional fMRI experiment at low temporal resolution (1.5–3 s).

In this study, we introduced a new U-shaped convolutional neural network DUNE to denoise ME-fMRI data as an alternative to MEICA in 2 ME-fMRI experiments. The resulting activation maps found after denoising with DUNE were compared with those found after denoising with MEICA and similar SE-fMRI experiments at high temporal resolution.

Our results revealed that DUNE was successful in reducing the noise while preserving the BOLD effects of interest comparable to MEICA and SE-fMRI. This latter result showed the potential of using a U-shaped convolutional neural network DUNE to denoise ME-fMRI data.

基于任务的fMRI数据受到扫描噪声和生理噪声的影响。因此，从噪声中找到基于任务的BOLD响应是具有挑战性的。为了提高检测BOLD响应的能力，多回波（ME） fMRI结合基于ICA的去噪（MEICA）和高时间分辨率（<1 s）的单回波（SE） fMRI被引入。在低时间分辨率（1.5-3秒）下，这两种技术都比传统的fMRI实验提供了更好的激活图。在这项研究中，我们引入了一种新的u形卷积神经网络DUNE来对ME-fMRI数据进行降噪，作为2个ME-fMRI实验中MEICA的替代方案。将DUNE去噪后得到的激活图与MEICA去噪和类似SE-fMRI实验在高时间分辨率下得到的激活图进行比较。我们的研究结果表明，DUNE成功地降低了噪声，同时保留了与MEICA和SE-fMRI相当的感兴趣的BOLD效果。后一个结果显示了使用u形卷积神经网络DUNE去噪ME-fMRI数据的潜力。

{"title":"The evaluation of DUNE: a U-Net-based neural network to denoise multi-echo fMRI data","authors":"Peter Van Schuerbeek , Manon Roose , Alina Monica Ionescu , Maarten Naeyaert , Hubert Raeymaekers","doi":"10.1016/j.neuroimage.2026.121749","DOIUrl":"10.1016/j.neuroimage.2026.121749","url":null,"abstract":"<div><div>Task based fMRI data suffers from scanner and physiologic noise. Consequently, finding the task based BOLD responses out of the noise is challenging. To improve the power to detect the BOLD responses, multi-echo (ME) fMRI combined with ICA based denoising (MEICA) and single-echo (SE) fMRI at high temporal resolution (<1 s) have been introduced. Both techniques have been found to give better activation maps than a traditional fMRI experiment at low temporal resolution (1.5–3 s).</div><div>In this study, we introduced a new U-shaped convolutional neural network DUNE to denoise ME-fMRI data as an alternative to MEICA in 2 ME-fMRI experiments. The resulting activation maps found after denoising with DUNE were compared with those found after denoising with MEICA and similar SE-fMRI experiments at high temporal resolution.</div><div>Our results revealed that DUNE was successful in reducing the noise while preserving the BOLD effects of interest comparable to MEICA and SE-fMRI. This latter result showed the potential of using a U-shaped convolutional neural network DUNE to denoise ME-fMRI data.</div></div>","PeriodicalId":19299,"journal":{"name":"NeuroImage","volume":"327 ","pages":"Article 121749"},"PeriodicalIF":4.5,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146025229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0