Neuroimage-Clinical最新文献_第5页

A multimodal Neuroimaging-Based risk score for mild cognitive impairment 轻度认知障碍的多模态神经影像学风险评分。

IF 3.4 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2024.103719

Elaheh Zendehrouh , Mohammad S.E. Sendi , Anees Abrol , Ishaan Batta , Reihaneh Hassanzadeh , Vince D. Calhoun

Introduction

Alzheimer’s disease (AD), the most prevalent age-related dementia, leads to significant cognitive decline. While genetic risk factors and neuroimaging biomarkers have been extensively studied, establishing a neuroimaging-based metric to assess AD risk has received less attention. This study introduces the Brain-wide Risk Score (BRS), a novel approach using multimodal neuroimaging data to assess the risk of mild cognitive impairment (MCI), a precursor to AD.

Methods

Participants from the OASIS-3 cohort (N = 1,389) were categorized into control (CN) and MCI groups. Structural MRI (sMRI) data provided gray matter (GM) segmentation maps, while resting-state functional MRI (fMRI) data yielded functional network connectivity (FNC) matrices via spatially constrained independent component analysis. Similar imaging features were computed from the UK Biobank (N = 37,780). The BRS was calculated by comparing each participant’s neuroimaging features to the difference between average features of CN and MCI groups. Both GM and FNC features were used. The BRS effectively differentiated CN from MCI individuals within OASIS-3 and in an independent dataset from the ADNI cohort (N = 729), demonstrating its ability to identify MCI risk.

Results

Unimodal analysis revealed that sMRI provided greater differentiation than fMRI, consistent with prior research. Using the multimodal BRS, we identified two distinct groups: one with high MCI risk (negative GM and FNC BRS) and another with low MCI risk (positive GM and FNC BRS). Additionally, 46 UK Biobank participants diagnosed with AD showed FNC and GM patterns similar to the high-risk groups.

Conclusion

Validation using the ADNI dataset confirmed our results, highlighting the potential of FNC and sMRI-based BRS in early Alzheimer’s detection.

简介：阿尔茨海默病（AD）是最常见的与年龄相关的痴呆症，可导致显著的认知能力下降。虽然遗传风险因素和神经影像学生物标志物已被广泛研究，但建立基于神经影像学的指标来评估AD风险却很少受到关注。本研究引入了全脑风险评分（BRS），这是一种使用多模态神经成像数据来评估轻度认知障碍（MCI）风险的新方法，轻度认知障碍是AD的前兆。方法：将OASIS-3队列（N = 1389）的参与者分为对照（CN）组和MCI组。结构MRI （sMRI）数据提供灰质（GM）分割图，而静息状态功能MRI （fMRI）数据通过空间约束的独立成分分析产生功能网络连接（FNC）矩阵。从UK Biobank （N = 37,780）计算了类似的成像特征。BRS是通过比较每个参与者的神经影像学特征与CN组和MCI组平均特征之间的差异来计算的。同时使用GM和FNC特征。BRS在绿洲-3和来自ADNI队列（N = 729）的独立数据集中有效地将CN与MCI个体区分开来，证明了其识别MCI风险的能力。结果：单峰分析显示sMRI比fMRI提供更大的分化，与先前的研究一致。使用多模式BRS，我们确定了两个不同的组：一组具有高MCI风险（阴性GM和FNC BRS），另一组具有低MCI风险（阳性GM和FNC BRS）。此外，46名被诊断为AD的英国生物银行参与者显示出与高危人群相似的FNC和GM模式。结论：ADNI数据集的验证证实了我们的结果，突出了FNC和基于smri的BRS在早期阿尔茨海默病检测中的潜力。

{"title":"A multimodal Neuroimaging-Based risk score for mild cognitive impairment","authors":"Elaheh Zendehrouh , Mohammad S.E. Sendi , Anees Abrol , Ishaan Batta , Reihaneh Hassanzadeh , Vince D. Calhoun","doi":"10.1016/j.nicl.2024.103719","DOIUrl":"10.1016/j.nicl.2024.103719","url":null,"abstract":"<div><h3>Introduction</h3><div>Alzheimer’s disease (AD), the most prevalent age-related dementia, leads to significant cognitive decline. While genetic risk factors and neuroimaging biomarkers have been extensively studied, establishing a neuroimaging-based metric to assess AD risk has received less attention. This study introduces the Brain-wide Risk Score (BRS), a novel approach using multimodal neuroimaging data to assess the risk of mild cognitive impairment (MCI), a precursor to AD.</div></div><div><h3>Methods</h3><div>Participants from the OASIS-3 cohort (<em>N</em> = 1,389) were categorized into control (CN) and MCI groups. Structural MRI (sMRI) data provided gray matter (GM) segmentation maps, while resting-state functional MRI (fMRI) data yielded functional network connectivity (FNC) matrices via spatially constrained independent component analysis. Similar imaging features were computed from the UK Biobank (<em>N</em> = 37,780). The BRS was calculated by comparing each participant’s neuroimaging features to the difference between average features of CN and MCI groups. Both GM and FNC features were used. The BRS effectively differentiated CN from MCI individuals within OASIS-3 and in an independent dataset from the ADNI cohort (<em>N</em> = 729), demonstrating its ability to identify MCI risk.</div></div><div><h3>Results</h3><div>Unimodal analysis revealed that sMRI provided greater differentiation than fMRI, consistent with prior research. Using the multimodal BRS, we identified two distinct groups: one with high MCI risk (negative GM and FNC BRS) and another with low MCI risk (positive GM and FNC BRS). Additionally, 46 UK Biobank participants diagnosed with AD showed FNC and GM patterns similar to the high-risk groups.</div></div><div><h3>Conclusion</h3><div>Validation using the ADNI dataset confirmed our results, highlighting the potential of FNC and sMRI-based BRS in early Alzheimer’s detection.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"45 ","pages":"Article 103719"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11664180/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142787700","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Assessing workflow impact and clinical utility of AI-assisted brain aneurysm detection: A multi-reader study 评估人工智能辅助脑动脉瘤检测的工作流程影响和临床应用：一项多读者研究

IF 3.4 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103835

Tommaso Di Noto , Sofyan Jankowski , Francesco Puccinelli , Guillaume Marie , Sebastien Tourbier , Yasser Alemán-Gómez , Oscar Esteban , Ricardo Corredor-Jerez , Guillaume Saliou , Patric Hagmann , Meritxell Bach Cuadra , Jonas Richiardi

Despite the plethora of AI-based algorithms developed for anomaly detection in radiology, subsequent integration into clinical setting is rarely evaluated. In this work, we assess the applicability and utility of an AI-based model for brain aneurysm detection comparing the performance of two readers with different levels of experience (2 and 13 years). We aim to answer the following questions: 1) Do the readers improve their performance when assisted by the AI algorithm? 2) How much does the AI algorithm impact routine clinical workflow? We reuse and enlarge our open-access, Time-Of-Flight Magnetic Resonance Angiography dataset (N = 460). We use 360 subjects for training/validating our algorithm and 100 as unseen test set for the reading session. Even though our model reaches state-of-the-art results on the test set (sensitivity = 74 %, false positive rate = 1.6), we show that neither the junior nor the senior reader significantly increase their sensitivity (p = 0.59, p = 1, respectively). In addition, we find that reading time for both readers is significantly higher in the “AI-assisted” setting than in the “Unassisted” (+15 s, on average;

p = 3 \times 10^{- 4}

junior,

p = 3 \times 10^{- 5}

senior). The confidence reported by the readers is unchanged across the two settings, indicating that the AI assistance does not influence the certainty of the diagnosis. Our findings highlight the importance of clinical validation of AI algorithms in a clinical setting involving radiologists. This study should serve as a reminder to the community to always examine the real-word effectiveness and workflow impact of proposed algorithms.

尽管基于人工智能的算法开发了大量用于放射学异常检测，但随后与临床环境的整合很少得到评估。在这项工作中，我们评估了基于人工智能的脑动脉瘤检测模型的适用性和实用性，比较了两名不同经验水平（2年和13年）的读者的表现。我们的目标是回答以下问题：1)在人工智能算法的帮助下，读者是否提高了他们的表现？2)人工智能算法对常规临床工作流程的影响有多大？我们重用并扩大了开放获取的飞行时间磁共振血管造影数据集（N = 460）。我们使用360个主题来训练/验证我们的算法，100个主题作为阅读部分的未见测试集。尽管我们的模型在测试集上达到了最先进的结果（灵敏度= 74%，假阳性率= 1.6），但我们发现初级和高级读者都没有显著提高他们的灵敏度（p = 0.59, p = 1）。此外，我们发现，在“人工智能辅助”设置下，两种读者的阅读时间都明显高于“无辅助”设置(平均+15秒；P =3×10-4 junior, P =3×10-5 senior)。在两种设置中，读者报告的置信度没有变化，这表明人工智能辅助不会影响诊断的确定性。我们的研究结果强调了人工智能算法在涉及放射科医生的临床环境中临床验证的重要性。这项研究应该作为一个提醒，社区总是检查实际的有效性和工作流程的影响所提出的算法。

{"title":"Assessing workflow impact and clinical utility of AI-assisted brain aneurysm detection: A multi-reader study","authors":"Tommaso Di Noto , Sofyan Jankowski , Francesco Puccinelli , Guillaume Marie , Sebastien Tourbier , Yasser Alemán-Gómez , Oscar Esteban , Ricardo Corredor-Jerez , Guillaume Saliou , Patric Hagmann , Meritxell Bach Cuadra , Jonas Richiardi","doi":"10.1016/j.nicl.2025.103835","DOIUrl":"10.1016/j.nicl.2025.103835","url":null,"abstract":"<div><div>Despite the plethora of AI-based algorithms developed for anomaly detection in radiology, subsequent integration into clinical setting is rarely evaluated. In this work, we assess the applicability and utility of an AI-based model for brain aneurysm detection comparing the performance of two readers with different levels of experience (2 and 13 years). We aim to answer the following questions: 1) Do the readers improve their performance when assisted by the AI algorithm? 2) How much does the AI algorithm impact routine clinical workflow? We reuse and enlarge our open-access, Time-Of-Flight Magnetic Resonance Angiography dataset (N = 460). We use 360 subjects for training/validating our algorithm and 100 as unseen test set for the reading session. Even though our model reaches state-of-the-art results on the test set (sensitivity = 74 %, false positive rate = 1.6), we show that neither the junior nor the senior reader significantly increase their sensitivity (p = 0.59, p = 1, respectively). In addition, we find that reading time for both readers is significantly higher in the “AI-assisted” setting than in the “Unassisted” (+15 s, on average; <span><math><mrow><mi>p</mi><mo>=</mo><mn>3</mn><mo>×</mo><msup><mrow><mn>10</mn></mrow><mrow><mo>-</mo><mn>4</mn></mrow></msup></mrow></math></span> junior, <span><math><mrow><mi>p</mi><mo>=</mo><mn>3</mn><mo>×</mo><msup><mrow><mn>10</mn></mrow><mrow><mo>-</mo><mn>5</mn></mrow></msup></mrow></math></span> senior). The confidence reported by the readers is unchanged across the two settings, indicating that the AI assistance does not influence the certainty of the diagnosis. Our findings highlight the importance of clinical validation of AI algorithms in a clinical setting involving radiologists. This study should serve as a reminder to the community to always examine the real-word effectiveness and workflow impact of proposed algorithms.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"47 ","pages":"Article 103835"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144549662","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

Modulations of thalamo-cortical coupling during voluntary movement in patients with essential tremor 特发性震颤患者随意运动时丘脑-皮质偶联的调节

IF 3.4 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103848

Alexandra Steina , Sarah Sure , Markus Butz , Jan Vesper , Alfons Schnitzler , Jan Hirschmann

The ventral intermediate nucleus of the thalamus (VIM) is the main thalamic hub for cerebellar inputs and the primary deep brain stimulation target in essential tremor (ET). As such, it presumably plays a critical role in motor control. However, this structure is rarely studied in humans, and existing studies mostly focus on tremor. Here, we studied neural oscillations in the VIM and their coupling to cortical oscillations during voluntary movement.

We investigated thalamo-cortical coupling, combining recordings of thalamic local field potentials and magnetoencephalography, in 10 ET patients with externalized deep brain stimulation electrodes. During the recording, patients repeatedly pressed a button in response to a visual cue. In a whole-brain analysis of VIM-cortex coherence, we contrasted activity during pre-movement baseline and button pressing.

Button pressing was associated with a bilateral decrease of thalamic alpha (8–12 Hz) and beta (13–21 Hz) power and a contralateral gamma (35–90 Hz) power increase. Alpha/low-beta (8–20 Hz) coherence decreased during movement. This effect localized to the supplementary motor area and premotor cortex. A high-beta (21–35 Hz) coherence increase occurred in the same region but was more focal than the suppression. Pre-movement levels of thalamo-cortex low-beta coherence correlated with reaction time.

Our results demonstrate that voluntary movement is associated with modulations of behaviourally relevant thalamic coupling, primarily to premotor areas. We observed a clear distinction between low- and high-beta frequencies and our results suggest that the concept of “antikinetic” beta oscillations, originating from research on Parkinson’s disease, is transferable to ET.

丘脑腹侧中间核（VIM）是小脑输入的主要丘脑中枢，也是特发性震颤（ET）的主要深部脑刺激靶点。因此，它可能在运动控制中起着关键作用。然而，这种结构很少在人类中进行研究，现有的研究主要集中在震颤上。在这里，我们研究了在自主运动过程中VIM的神经振荡及其与皮质振荡的耦合。我们研究了10例ET患者的丘脑-皮质耦合，结合丘脑局部场电位和脑磁图的记录。在录音过程中，患者根据视觉提示反复按下按钮。在全脑分析中，我们对比了运动前基线和按下按钮时的活动。按下按钮与双侧丘脑α (8-12 Hz)和β (13-21 Hz)功率下降和对侧γ (35-90 Hz)功率增加有关。α /低- β (8-20 Hz)相干性在运动过程中下降。这种作用局限于辅助运动区和运动前皮层。高β (21-35 Hz)相干性增加发生在同一区域，但比抑制更集中。运动前丘脑-皮层低相干性水平与反应时间相关。我们的研究结果表明，自主运动与行为相关的丘脑耦合调节有关，主要是运动前区域。我们观察到低频率和高频率之间的明显区别，我们的结果表明，起源于帕金森氏病研究的“反动力学”β振荡概念可转移到ET。

{"title":"Modulations of thalamo-cortical coupling during voluntary movement in patients with essential tremor","authors":"Alexandra Steina , Sarah Sure , Markus Butz , Jan Vesper , Alfons Schnitzler , Jan Hirschmann","doi":"10.1016/j.nicl.2025.103848","DOIUrl":"10.1016/j.nicl.2025.103848","url":null,"abstract":"<div><div>The ventral intermediate nucleus of the thalamus (VIM) is the main thalamic hub for cerebellar inputs and the primary deep brain stimulation target in essential tremor (ET). As such, it presumably plays a critical role in motor control. However, this structure is rarely studied in humans, and existing studies mostly focus on tremor. Here, we studied neural oscillations in the VIM and their coupling to cortical oscillations during voluntary movement.</div><div>We investigated thalamo-cortical coupling, combining recordings of thalamic local field potentials and magnetoencephalography, in 10 ET patients with externalized deep brain stimulation electrodes. During the recording, patients repeatedly pressed a button in response to a visual cue. In a whole-brain analysis of VIM-cortex coherence, we contrasted activity during pre-movement baseline and button pressing.</div><div>Button pressing was associated with a bilateral decrease of thalamic alpha (8–12 Hz) and beta (13–21 Hz) power and a contralateral gamma (35–90 Hz) power increase. Alpha/low-beta (8–20 Hz) coherence decreased during movement. This effect localized to the supplementary motor area and premotor cortex. A high-beta (21–35 Hz) coherence increase occurred in the same region but was more focal than the suppression. Pre-movement levels of thalamo-cortex low-beta coherence correlated with reaction time.</div><div>Our results demonstrate that voluntary movement is associated with modulations of behaviourally relevant thalamic coupling, primarily to premotor areas. We observed a clear distinction between low- and high-beta frequencies and our results suggest that the concept of “antikinetic” beta oscillations, originating from research on Parkinson’s disease, is transferable to ET.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"48 ","pages":"Article 103848"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144662572","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

Broader functionality of language areas at the left middle frontal gyrus in patients with Broca’s area tumors 布洛卡区肿瘤患者左额叶中回语言区更广泛的功能

IF 3.6 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103860

Riho Nakajima , Akitoshi Ogawa , Masashi Kinoshita , Takahiro Osada , Hirokazu Okita , Seiki Konishi , Mitsutoshi Nakada

The frontal language area (FLA; left posterior inferior frontal gyrus [pIFG] or Broca’s area), critical for language processing can reorganize in response to lesion progression. While reorganization in the contralateral hemisphere is well known, how reorganization occurs within the ipsilateral hemisphere, especially in the perilesional region, remains unclear. Direct electrical stimulation (DES) during awake surgery enables identification of causal relationships between brain regions and language functions with high spatial resolution. In this study, we investigated cortical reorganization within the ipsilateral hemisphere of the FLA. Seventy-two patients with left hemisphere gliomas were studied. Patients were divided into FLA and non-FLA groups based on whether lesions included the pIFG (n = 10 and n = 62, respectively). All patients underwent DES during a picture-naming task, as recommended by awake surgery guidelines. A subset also underwent resting-state functional MRI (rsfMRI) before surgery to calculate betweenness centrality, an index of network importance of brain areas. DES revealed that the pIFG exhibited positive (impaired) responses to the picture-naming task in both groups. Notably, the frequency of positive responses in the middle frontal gyrus (MFG) was significantly higher in the FLA group than in the non-FLA group. RsfMRI-based network analyses revealed that two areas in the MFG, one in the anterior part and the other in the posterior part, showed higher centrality than surrounding frontal areas in both groups, especially the posterior one. These results suggest that language areas can be observed in the perilesional MFG regions following tumor progression, and raise the possibility that network hubs contribute to maintaining cognitive functions after brain lesions.

额叶语言区；左侧额下后回（pIFG）或布洛卡区)，对语言处理至关重要，可以在病变进展中重组。虽然对侧半球的重组是众所周知的，但同侧半球的重组是如何发生的，特别是在病灶周围区域，尚不清楚。清醒手术期间的直接电刺激（DES）能够以高空间分辨率识别大脑区域与语言功能之间的因果关系。在这项研究中，我们研究了FLA同侧半球的皮质重组。本文对72例左脑胶质瘤患者进行了研究。根据病变是否包含pIFG将患者分为FLA组和非FLA组（n = 10和n = 62）。根据清醒手术指南的建议，所有患者在图片命名任务期间都进行了DES。一个子集在手术前接受静息状态功能MRI （rsfMRI）来计算中间性中心性，这是脑区域网络重要性的指标。DES显示，两组的pIFG对图片命名任务都表现出积极（受损）的反应。值得注意的是，与非FLA组相比，FLA组中额叶回（MFG）阳性反应的频率明显更高。基于rsfmri的网络分析显示，两组MFG的两个区域，一个在前部，另一个在后部，都比周围额叶区域表现出更高的中心性，尤其是后部。这些结果表明，在肿瘤进展后，语言区可以在病灶周围的MFG区域观察到，并提出了网络枢纽有助于维持脑病变后认知功能的可能性。

{"title":"Broader functionality of language areas at the left middle frontal gyrus in patients with Broca’s area tumors","authors":"Riho Nakajima , Akitoshi Ogawa , Masashi Kinoshita , Takahiro Osada , Hirokazu Okita , Seiki Konishi , Mitsutoshi Nakada","doi":"10.1016/j.nicl.2025.103860","DOIUrl":"10.1016/j.nicl.2025.103860","url":null,"abstract":"<div><div>The frontal language area (FLA; left posterior inferior frontal gyrus [pIFG] or Broca’s area), critical for language processing can reorganize in response to lesion progression. While reorganization in the contralateral hemisphere is well known, how reorganization occurs within the ipsilateral hemisphere, especially in the perilesional region, remains unclear. Direct electrical stimulation (DES) during awake surgery enables identification of causal relationships between brain regions and language functions with high spatial resolution. In this study, we investigated cortical reorganization within the ipsilateral hemisphere of the FLA. Seventy-two patients with left hemisphere gliomas were studied. Patients were divided into FLA and non-FLA groups based on whether lesions included the pIFG (n = 10 and n = 62, respectively). All patients underwent DES during a picture-naming task, as recommended by awake surgery guidelines. A subset also underwent resting-state functional MRI (rsfMRI) before surgery to calculate betweenness centrality, an index of network importance of brain areas. DES revealed that the pIFG exhibited positive (impaired) responses to the picture-naming task in both groups. Notably, the frequency of positive responses in the middle frontal gyrus (MFG) was significantly higher in the FLA group than in the non-FLA group. RsfMRI-based network analyses revealed that two areas in the MFG, one in the anterior part and the other in the posterior part, showed higher centrality than surrounding frontal areas in both groups, especially the posterior one. These results suggest that language areas can be observed in the perilesional MFG regions following tumor progression, and raise the possibility that network hubs contribute to maintaining cognitive functions after brain lesions.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"48 ","pages":"Article 103860"},"PeriodicalIF":3.6,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144827430","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

Simultaneous MRI and laser Doppler Flowmetry: Assessing cerebral Macro- and microcirculation in neurointensive care 同时MRI和激光多普勒血流仪：评估神经重症监护患者的大脑宏观和微循环

IF 3.4 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103821

Sofie Tapper , Stina Mauritzon , Marcelo P. Martins , Fredrik Ginstman , Anders Tisell , Peter Zsigmond , Karin Wårdell

Objective

Subarachnoid hemorrhage (SAH) patients are monitored in the neurointensive care unit (NICU) to avoid additional brain injuries, yet methods for monitoring cerebral blood flow (CBF) are limited. The aim was to investigate the feasibility of simultaneous MRI, using arterial spin labeling (ASL) and 2D-flow MRI, and probe-based laser Doppler flowmetry (LDF) in NICU patients.

Method

Four SAH patients, three of whom received optical probes during routine surgery, were included. Compatibility of simultaneous LDF and MRI was evaluated before performing MRI 3–4 times over several days per patient. The methods were compared using mean CBF in grey matter obtained from the absolute ASL CBF-maps, the total inflow calculated from 2D-flow MRI, and local average LDF perfusion. Additionally, regional mean CBF from ASL and each arterial flowrate was compared for each hemisphere.

Result

Twelve MRI measurements were successfully performed, eight of which included simultaneous LDF. With careful routines and the MR scanner uniquely located in the NICU, neither patient safety nor data quality was compromised, demonstrating the feasibility of concurrent measurements. All methods showed longitudinal dynamic changes, following the same increasing or decreasing trends. In three patients, ASL and 2D-flow data were closely related, with dynamic changes within 10 %.

Conclusion

This novel approach offers longitudinal, simultaneous estimates of macro- and microcirculatory components locally, regionally, and globally in the human brain. This concept has potential to provide insights into the interplay of different aspects of CBF in NICU patients and thereby aid in prevention of secondary brain injuries.

目的在神经重症监护病房（NICU）对蛛网膜下腔出血（SAH）患者进行监测，以避免额外的脑损伤，但监测脑血流（CBF）的方法有限。目的是探讨在NICU患者中同时使用动脉自旋标记（ASL）和2d血流MRI以及基于探针的激光多普勒血流仪（LDF）的可行性。方法选取4例SAH患者，其中3例在常规手术中接受了光学探针。在每位患者数天内进行3-4次MRI检查前，评估LDF与MRI同时检查的兼容性。通过ASL脑血流绝对图获得的灰质平均脑血流、2D-flow MRI计算的总脑流和局部平均LDF灌注对两种方法进行比较。此外，比较ASL的区域平均CBF和每个半球的动脉流速。结果12次MRI测量成功，其中8次同时进行LDF测量。通过仔细的例行程序和独特的位于NICU的MR扫描仪，患者的安全性和数据质量都没有受到损害，证明了并发测量的可行性。所有方法均表现出纵向动态变化，具有相同的增减趋势。3例患者ASL与2D-flow数据密切相关，动态变化在10%以内。这种新方法提供了对人类大脑局部、区域和全局宏观和微循环成分的纵向、同步估计。这一概念有可能为新生儿重症监护病房患者脑血流不同方面的相互作用提供见解，从而有助于预防继发性脑损伤。

{"title":"Simultaneous MRI and laser Doppler Flowmetry: Assessing cerebral Macro- and microcirculation in neurointensive care","authors":"Sofie Tapper , Stina Mauritzon , Marcelo P. Martins , Fredrik Ginstman , Anders Tisell , Peter Zsigmond , Karin Wårdell","doi":"10.1016/j.nicl.2025.103821","DOIUrl":"10.1016/j.nicl.2025.103821","url":null,"abstract":"<div><h3>Objective</h3><div>Subarachnoid hemorrhage (SAH) patients are monitored in the neurointensive care unit (NICU) to avoid additional brain injuries, yet methods for monitoring cerebral blood flow (CBF) are limited. The aim was to investigate the feasibility of simultaneous MRI, using arterial spin labeling (ASL) and 2D-flow MRI, and probe-based laser Doppler flowmetry (LDF) in NICU patients.</div></div><div><h3>Method</h3><div>Four SAH patients, three of whom received optical probes during routine surgery, were included. Compatibility of simultaneous LDF and MRI was evaluated before performing MRI 3–4 times over several days per patient. The methods were compared using mean CBF in grey matter obtained from the absolute ASL CBF-maps, the total inflow calculated from 2D-flow MRI, and local average LDF perfusion. Additionally, regional mean CBF from ASL and each arterial flowrate was compared for each hemisphere.</div></div><div><h3>Result</h3><div>Twelve MRI measurements were successfully performed, eight of which included simultaneous LDF. With careful routines and the MR scanner uniquely located in the NICU, neither patient safety nor data quality was compromised, demonstrating the feasibility of concurrent measurements. All methods showed longitudinal dynamic changes, following the same increasing or decreasing trends. In three patients, ASL and 2D-flow data were closely related, with dynamic changes within 10 %.</div></div><div><h3>Conclusion</h3><div>This novel approach offers longitudinal, simultaneous estimates of macro- and microcirculatory components locally, regionally, and globally in the human brain. This concept has potential to provide insights into the interplay of different aspects of CBF in NICU patients and thereby aid in prevention of secondary brain injuries.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"47 ","pages":"Article 103821"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144307676","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

Comparable 40 Hz auditory steady-state responses in children at familial high risk for schizophrenia or bipolar disorder and population-based controls 精神分裂症或双相情感障碍家族性高风险儿童与基于人群的对照中40hz听觉稳态反应的比较

IF 3.4 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103841

Kit Melissa Larsen , Júlia Díaz-i-Calvete , Anna Hester Ver Loren van Themaat , Anne Amalie Elgaard Thorup , Kerstin Jessica Plessen , Merete Nordentoft , Hartwig Roman Siebner

引用次数: 0

Beyond language: empathy and emotion recognition deficits in primary progressive aphasias 超越语言：原发性进行性失语症的移情和情绪识别缺陷。

IF 3.6 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103852

Giulia Giacomucci , Alice Pieri , Valentina Moschini , Chiara Crucitti , Sonia Padiglioni , Carmen Morinelli , Giulia Galdo , Filippo Emiliani , Matilde Nerattini , Silvia Bagnoli , Assunta Ingannato , Sandro Sorbi , Benedetta Nacmias , Valentina Berti , Valentina Bessi

Although primary progressive aphasia (PPA) is considered a language disorder, increasing evidence points to the presence of social cognition impairments in PPA variants. The aims of this study were to explore empathy and emotion recognition deficits in the three PPA variants (sv-PPA, lv-PPA, nfv-PPA) and to identify their neural correlates.

Eleven sv-PPA, 34 lv-PPA,11 nfv-PPA patients and 34 healthy controls (HC) were included in this study. Empathy was explored with the Interpersonal Reactivity Index (IRI) (Perspective Taking – PT, Fantasy – FT, Empathic Concern – EC, Personal Distress – PD), rated by caregivers before (T0) and after (T1) the onset of cognitive symptoms. Emotion recognition was evaluated with the Ekman 60Faces (EK-60F) Test and metabolic activity with [18F]FDG-PET.

In all PPA variants, PT score was reduced from T0 to T1 (sv-PPA p = 0.014, lv-PPA p < 0.001, nfv-PPA p = 0.022) and PD score was increased (sv-PPA p = 0.033, lv-PPA p < 0.001, nfv-PPA p = 0.009). Only lv-PPA showed a decrease of FT score (p = 0.024), while EC was spared in all three variants. Sv-PPA patients had the worst performances in the EK-60F Test, followed by lv-PPA and, lastly, by nfv-PPA.

Correlations between EK-60F scores and metabolic activity were found in sv-PPA and lv-PPA, highlighting the involvement of areas participating in the emotion recognition network: cingulate cortex, insula, temporal and orbitofrontal cortices and inferior frontal gyrus.

All PPA variants exhibited impairments in cognitive empathy (PT) and heightened emotional contagion (PD). The most severe deficits in emotion recognition were shown by sv-PPA, while nfv-PPA was the less impaired variant.

虽然原发性进行性失语症（PPA）被认为是一种语言障碍，但越来越多的证据表明，PPA变体中存在社会认知障碍。本研究的目的是探讨三种PPA变体（sv-PPA、lv-PPA、nfv-PPA）的共情和情绪识别缺陷，并确定它们的神经相关性。svv - ppa患者11例，lv-PPA患者34例，nfv-PPA患者11例，健康对照34例。共情采用人际反应指数（IRI）（观点采取- PT，幻想- FT，共情关注- EC，个人困扰- PD），由照顾者在认知症状出现前（T0）和后（T1）评定。用Ekman 60Faces （EK-60F）测试评估情绪识别，用[18F]FDG-PET评估代谢活性。在所有PPA变异中，PT评分从T0降至T1 （sv-PPA p = 0.014, lv-PPA p = 0.014）

{"title":"Beyond language: empathy and emotion recognition deficits in primary progressive aphasias","authors":"Giulia Giacomucci , Alice Pieri , Valentina Moschini , Chiara Crucitti , Sonia Padiglioni , Carmen Morinelli , Giulia Galdo , Filippo Emiliani , Matilde Nerattini , Silvia Bagnoli , Assunta Ingannato , Sandro Sorbi , Benedetta Nacmias , Valentina Berti , Valentina Bessi","doi":"10.1016/j.nicl.2025.103852","DOIUrl":"10.1016/j.nicl.2025.103852","url":null,"abstract":"<div><div>Although primary progressive aphasia (PPA) is considered a language disorder, increasing evidence points to the presence of social cognition impairments in PPA variants. The aims of this study were to explore empathy and emotion recognition deficits in the three PPA variants (sv-PPA, lv-PPA, nfv-PPA) and to identify their neural correlates.</div><div>Eleven sv-PPA, 34 lv-PPA,11 nfv-PPA patients and 34 healthy controls (HC) were included in this study. Empathy was explored with the Interpersonal Reactivity Index (IRI) (Perspective Taking – PT, Fantasy – FT, Empathic Concern – EC, Personal Distress – PD), rated by caregivers before (T0) and after (T1) the onset of cognitive symptoms. Emotion recognition was evaluated with the Ekman 60Faces (EK-60F) Test and metabolic activity with [18F]FDG-PET.</div><div>In all PPA variants, PT score was reduced from T0 to T1 (sv-PPA <em>p</em> = 0.014, lv-PPA <em>p</em> < 0.001, nfv-PPA <em>p</em> = 0.022) and PD score was increased (sv-PPA <em>p</em> = 0.033, lv-PPA <em>p</em> < 0.001, nfv-PPA <em>p</em> = 0.009). Only lv-PPA showed a decrease of FT score (<em>p</em> = 0.024), while EC was spared in all three variants. Sv-PPA patients had the worst performances in the EK-60F Test, followed by lv-PPA and, lastly, by nfv-PPA.</div><div>Correlations between EK-60F scores and metabolic activity were found in sv-PPA and lv-PPA, highlighting the involvement of areas participating in the emotion recognition network: cingulate cortex, insula, temporal and orbitofrontal cortices and inferior frontal gyrus.</div><div>All PPA variants exhibited impairments in cognitive empathy (PT) and heightened emotional contagion (PD). The most severe deficits in emotion recognition were shown by sv-PPA, while nfv-PPA was the less impaired variant.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"48 ","pages":"Article 103852"},"PeriodicalIF":3.6,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144769309","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

Electric field variations across DLPFC targeting methods in TMS therapy for Alzheimer’s disease 经颅磁刺激治疗阿尔茨海默病DLPFC靶向方法的电场变化

IF 3.4 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103847

Nianshuang Wu , Yuxuan Shao , Zhen Wu , Shuxiang Zhu , Penghao Wang , Ziyan Zhu , Cheng Zhang , Changzhe Wu , Xiaolin Huo , Hua Lin , Guanghao Zhang

Background

The dorsolateral prefrontal cortex (DLPFC) is crucial for cognitive control and a primary target for transcranial magnetic stimulation (TMS) in Alzheimer’s disease (AD). However, understanding the distribution of TMS-induced electric field (E-field) across different targeting methods remains limited, as does its relationship to therapeutic outcomes.

Objective

This study assesses differences in TMS-induced E-field using functional versus anatomical targeting methods for DLPFC stimulation.

Methods

Functional and anatomical targets were identified in 30 (11 M/19F) AD patients and 30 (13 M/17F) age-matched healthy controls (HCs) using T1 and fMRI data. E-field characteristics, including magnitude (E_ROI) and normal component (E_⊥), were calculated via SimNIBS software for comparisons across stimulation targets.

Results

Functional targeting showed greater spatial dispersion compared to anatomical targeting in both groups. Significant E-field differences were observed between the functional target and adjacent anatomical regions when the coil was positioned over the functional target in both groups. Optimal coil orientation exhibited directional specificity: parallel alignment with the LOI E-field produced higher field intensity in the functional target compared to the anatomical target (AD patients: P < 0.001; HCs: P = 0.052), while perpendicular orientation maintained functional stability with reduced anatomical interference (both groups: P < 0.001). And significant variations in E-field ratios were observed across coil orientations.

Conclusion

This study reveals key E-field disparities across DLPFC targeting approaches and establishes coil orientation optimization as a critical strategy to improve TMS precision, offering actionable insights for developing personalized protocols in AD therapy that may enhance treatment efficacy while minimizing adverse effects.

背外侧前额叶皮层（DLPFC）对认知控制至关重要，也是经颅磁刺激（TMS）治疗阿尔茨海默病（AD）的主要靶点。然而，对tms诱导电场（E-field）在不同靶向方法中的分布及其与治疗结果的关系的了解仍然有限。目的利用功能性和解剖性靶向方法，评估经颅磁刺激诱发的脑电场的差异。方法利用T1和fMRI数据对30例（11 M/19F） AD患者和30例（13 M/17F）年龄匹配的健康对照（hc）进行功能和解剖靶点鉴定。电磁场特征，包括震级（EROI）和法向分量（E⊥），是通过SimNIBS软件计算的，以便在刺激目标之间进行比较。结果两组的功能性靶向均表现出更大的空间离散性。在两组中，当线圈放置在功能靶上时，在功能靶和邻近解剖区域之间观察到显著的电场差异。最佳线圈定向表现出方向特异性：与LOI电场平行对准在功能靶点产生比解剖靶点更高的场强度(AD患者：P <；0.001;HCs: P = 0.052)，而垂直取向保持功能稳定，解剖干扰减少(两组：P <；0.001)。在不同的线圈方向上观察到显著的电场比变化。结论本研究揭示了不同DLPFC靶向方法的关键电场差异，并确定了线圈定向优化是提高TMS精度的关键策略，为开发个性化的AD治疗方案提供了可行的见解，这些方案可能会提高治疗效果，同时最大限度地减少不良反应。

{"title":"Electric field variations across DLPFC targeting methods in TMS therapy for Alzheimer’s disease","authors":"Nianshuang Wu , Yuxuan Shao , Zhen Wu , Shuxiang Zhu , Penghao Wang , Ziyan Zhu , Cheng Zhang , Changzhe Wu , Xiaolin Huo , Hua Lin , Guanghao Zhang","doi":"10.1016/j.nicl.2025.103847","DOIUrl":"10.1016/j.nicl.2025.103847","url":null,"abstract":"<div><h3>Background</h3><div>The dorsolateral prefrontal cortex (DLPFC) is crucial for cognitive control and a primary target for transcranial magnetic stimulation (TMS) in Alzheimer’s disease (AD). However, understanding the distribution of TMS-induced electric field (E-field) across different targeting methods remains limited, as does its relationship to therapeutic outcomes.</div></div><div><h3>Objective</h3><div>This study assesses differences in TMS-induced E-field using functional versus anatomical targeting methods for DLPFC stimulation.</div></div><div><h3>Methods</h3><div>Functional and anatomical targets were identified in 30 (11 M/19F) AD patients and 30 (13 M/17F) age-matched healthy controls (HCs) using T1 and fMRI data. E-field characteristics, including magnitude (E<sub>ROI</sub>) and normal component (E<sub>⊥</sub>), were calculated via SimNIBS software for comparisons across stimulation targets.</div></div><div><h3>Results</h3><div>Functional targeting showed greater spatial dispersion compared to anatomical targeting in both groups. Significant E-field differences were observed between the functional target and adjacent anatomical regions when the coil was positioned over the functional target in both groups. Optimal coil orientation exhibited directional specificity: parallel alignment with the LOI E-field produced higher field intensity in the functional target compared to the anatomical target (AD patients: <em>P</em> < 0.001; HCs: <em>P</em> = 0.052), while perpendicular orientation maintained functional stability with reduced anatomical interference (both groups: <em>P</em> < 0.001). And significant variations in E-field ratios were observed across coil orientations.</div></div><div><h3>Conclusion</h3><div>This study reveals key E-field disparities across DLPFC targeting approaches and establishes coil orientation optimization as a critical strategy to improve TMS precision, offering actionable insights for developing personalized protocols in AD therapy that may enhance treatment efficacy while minimizing adverse effects.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"48 ","pages":"Article 103847"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144679686","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

Corrigendum to “Disturbed hierarchy and mediation in reward-related circuits in depression”. [45 (2025) 103739] “抑郁症中奖励相关回路中紊乱的等级和调解”的勘误表。[45(2025) 103739]。

IF 3.6 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103855

Ruikun Yang , Junxia Chen , Suping Yue , Yue Yu , Jiamin Fan , Yuling Luo , Hui He , Mingjun Duan , Sisi Jiang , Dezhong Yao , Cheng Luo

引用次数: 0

Deep convolutional neural networks outperform vanilla machine learning when predicting language outcomes after stroke 深度卷积神经网络在预测中风后的语言结果方面优于普通机器学习

IF 3.6 2区医学 Q2 NEUROIMAGING

Neuroimage-Clinical

Pub Date : 2025-01-01 DOI: 10.1016/j.nicl.2025.103880

Thomas M.H. Hope , Howard Bowman , Alex P. Leff , Cathy J. Price

Background

Current medicine cannot confidently predict patients’ language skills after stroke. In recent years, researchers have sought to bridge this gap with machine learning. These models appear to benefit from access to features describing where and how much brain damage these patients have suffered. Given the very high dimensionality of structural brain imaging data, those brain lesion features are typically post-processed from the images themselves into tabular features. With the introduction of deep Convolutional Neural Networks (CNN), which appear to be much more robust to high dimensional data, it is natural to hope that much of this image post-processing might be unnecessary. But prior attempts to demonstrate this (in the area of post-stroke prognostics) have so far yielded only equivocal results – perhaps because the datasets that those studies could deploy were too small to properly constrain CNNs, which are famously ‘data-hungry’.

Methods

The study draws on a much larger dataset than has been employed in previous work like this, referring to patients whose language outcomes were assessed once during the chronic phase post-stroke, on or around the same days as they underwent high resolution MRI brain scans. Following the model of our own and others’ past work, we use state of the art ‘vanilla’ machine learning models (boosted ensembles) to predict a variety of language and cognitive outcomes scores. These models employ both demographic variables and features derived from the brain imaging data, which represent where brain damage has occurred. These are our baseline models. Next, we use deep CNNs to predict the same language scores for the same patients, drawing on both the demographic variables, and post-processed brain lesion images: i.e., multi-input models with one input for tabular features and another for 3-dimensional images. We compare the models using 5 × 2-fold cross-validation, with consistent folds.

Results

The CNN models consistently outperform the vanilla machine learning models, in this domain.

Conclusions

Deep CNNs offer state of the art performance when predicting language outcomes after stroke, outperforming vanilla machine learning and obviating the need to post-process lesion images into lesion features.

目前的医学不能自信地预测中风后患者的语言能力。近年来，研究人员试图用机器学习来弥合这一差距。这些模型似乎受益于描述这些患者遭受的脑损伤的位置和程度的特征。考虑到脑结构成像数据的高维度，这些脑损伤特征通常是由图像本身后处理成表格特征。随着深度卷积神经网络（CNN）的引入，它对高维数据的鲁棒性似乎要强得多，人们自然希望大部分图像后处理可能是不必要的。但是之前证明这一点的尝试（在中风后预后领域）到目前为止只产生了模棱两可的结果——也许是因为这些研究可以部署的数据集太小，无法适当地约束cnn，而cnn是出了名的“数据饥渴”。方法：该研究使用的数据集比以前的研究要大得多，研究对象是中风后慢性期的患者，在接受高分辨率核磁共振脑部扫描的当天或同一天左右，对他们的语言结果进行了一次评估。根据我们自己和其他人过去工作的模型，我们使用最先进的“香草”机器学习模型（增强集合）来预测各种语言和认知结果分数。这些模型采用了人口统计学变量和来自脑成像数据的特征，这些数据代表了脑损伤发生的位置。这些是我们的基线模型。接下来，我们使用深度cnn来预测相同患者的相同语言分数，同时利用人口统计学变量和后处理的脑损伤图像：即多输入模型，其中一个输入表格特征，另一个输入三维图像。我们使用5 × 2倍交叉验证比较模型，具有一致的折叠。结果在这一领域，CNN模型始终优于普通机器学习模型。深度cnn在预测中风后的语言结果方面提供了最先进的性能，优于普通机器学习，并且避免了将病变图像后处理为病变特征的需要。

{"title":"Deep convolutional neural networks outperform vanilla machine learning when predicting language outcomes after stroke","authors":"Thomas M.H. Hope , Howard Bowman , Alex P. Leff , Cathy J. Price","doi":"10.1016/j.nicl.2025.103880","DOIUrl":"10.1016/j.nicl.2025.103880","url":null,"abstract":"<div><h3>Background</h3><div>Current medicine cannot confidently predict patients’ language skills after stroke. In recent years, researchers have sought to bridge this gap with machine learning. These models appear to benefit from access to features describing where and how much brain damage these patients have suffered. Given the very high dimensionality of structural brain imaging data, those brain lesion features are typically post-processed from the images themselves into tabular features. With the introduction of deep Convolutional Neural Networks (CNN), which appear to be much more robust to high dimensional data, it is natural to hope that much of this image post-processing might be unnecessary. But prior attempts to demonstrate this (in the area of post-stroke prognostics) have so far yielded only equivocal results – perhaps because the datasets that those studies could deploy were too small to properly constrain CNNs, which are famously ‘data-hungry’.</div></div><div><h3>Methods</h3><div>The study draws on a much larger dataset than has been employed in previous work like this, referring to patients whose language outcomes were assessed once during the chronic phase post-stroke, on or around the same days as they underwent high resolution MRI brain scans. Following the model of our own and others’ past work, we use state of the art ‘vanilla’ machine learning models (boosted ensembles) to predict a variety of language and cognitive outcomes scores. These models employ both demographic variables and features derived from the brain imaging data, which represent where brain damage has occurred. These are our baseline models. Next, we use deep CNNs to predict the same language scores for the same patients, drawing on both the demographic variables, and post-processed brain lesion images: i.e., multi-input models with one input for tabular features and another for 3-dimensional images. We compare the models using 5 × 2-fold cross-validation, with consistent folds.</div></div><div><h3>Results</h3><div>The CNN models consistently outperform the vanilla machine learning models, in this domain.</div></div><div><h3>Conclusions</h3><div>Deep CNNs offer state of the art performance when predicting language outcomes after stroke, outperforming vanilla machine learning and obviating the need to post-process lesion images into lesion features.</div></div>","PeriodicalId":54359,"journal":{"name":"Neuroimage-Clinical","volume":"48 ","pages":"Article 103880"},"PeriodicalIF":3.6,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145219053","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