Marie Louise Liu
Marie Louise Liu has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Hartwig Roman Siebner
Danish Research Centre for Magnetic Resonance, Department of Radiology and Nuclear Medicine, Copenhagen University Hospital - Amager and Hvidovre
2 shared publications - Agnes Mayr
University Clinic of Radiology, Medical University of Innsbruck, Anichstraße 35, 6020 Innsbruck, Austria
1 shared publication Anke Ninija Karabanov
Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital - Amager and Hvidovre
1 shared publicationAnne Francke Christensen
Danish Headache Center and Department of Neurology, Rigshospitalet Glostrup Hospital, Denmark
1 shared publicationAxel Thielscher
Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital - Amager and Hvidovre
1 shared publication- Axel Thielscher
Department of Health Technology, Technical University of Denmark, Kongens Lyngby, Denmark
1 shared publication
Latest Publications
- Open Access
- 20 November 2025
- Published by MDPI AG in Journal of Clinical Medicine
- Vol. 14 (22) , 8255
Are Image-Based Deep Learning Algorithms of Kidney Volume in Polycystic Kidney Disease Ready for Clinical Deployment? A Systematic Review and Meta-Analysis
Objectives: In patients with autosomal dominant polycystic kidney disease (ADPKD), total kidney volume (TKV) is the gold standard biomarker for assessing the risk of progression and the need for drug therapy. However, it is a time-consuming process. In this systematic review and meta-analysis, we evaluate the current state of deep learning (DL) algorithms for automatic kidney volume segmentation. Methods: All original research, including the search terms ADPKD, diagnostic imaging, DL, and TKV, was identified in PubMed, Embase, and Ovid MEDLINE databases from January 2000 to 13 October 2024. Articles with insufficient information to assess methodological quality were excluded. Quality was assessed using the “Quality Assessment of Diagnostic Accuracy Studies, Version 2” (QUADAS-2) and Checklist for Artificial Intelligence in Medical Imaging (CLAIM) tools. We focused on the Dice Similarity Coefficient (DSC), bias differences, and time efficiency as outcomes. Results: Nineteen studies were included, with an overall low risk of bias; however, the mean adherence to the CLAIM checklist was 64%. The pooled DSC under the random-effects model was (95% CI: ) with relatively low bias for TKV in 5622 ADPKD patients (mean age, 46.1 years; 45% male) and 9180 scans (79% MRI). The average segmentation time was decreased by 75% compared to the ground truth. Performance differences were evident among imaging modalities, MRI sequences, and 3D vs. 2D models, but not among imaging planes. The between-study heterogeneity was low (), and no statistically significant evidence of small-study effects or publication bias was detected. Conclusions: DL models for TKV in ADPKD patients demonstrated high precision compared to manual segmentation in a large, pooled sample with heterogeneous study designs and methods. While clinical implementation is not yet feasible, the current work demonstrates the technical and diagnostic efficacy of image-based DL segmentation models.- 20 November 2025
- Published by MDPI AG in Journal of Clinical Medicine
- Vol. 14 (22) , 8255
Open AccessAre Image-Based Deep Learning Algorithms of Kidney Volume in Polycystic Kidney Disease Ready for Clinical Deployment? A Systematic Review and Meta-Analysis
Objectives: In patients with autosomal dominant polycystic kidney disease (ADPKD), total kidney volume (TKV) is the gold standard biomarker for assessing the risk of progression and the need for drug therapy. However, it is a time-consuming process. In this systematic review and meta-analysis, we evaluate the current state of deep learning (DL) algorithms for automatic kidney volume segmentation. Methods: All original research, including the search terms ADPKD, diagnostic imaging, DL, and TKV, was identified in PubMed, Embase, and Ovid MEDLINE databases from January 2000 to 13 October 2024. Articles with insufficient information to assess methodological quality were excluded. Quality was assessed using the “Quality Assessment of Diagnostic Accuracy Studies, Version 2” (QUADAS-2) and Checklist for Artificial Intelligence in Medical Imaging (CLAIM) tools. We focused on the Dice Similarity Coefficient (DSC), bias differences, and time efficiency as outcomes. Results: Nineteen studies were included, with an overall low risk of bias; however, the mean adherence to the CLAIM checklist was 64%. The pooled DSC under the random-effects model was (95% CI: ) with relatively low bias for TKV in 5622 ADPKD patients (mean age, 46.1 years; 45% male) and 9180 scans (79% MRI). The average segmentation time was decreased by 75% compared to the ground truth. Performance differences were evident among imaging modalities, MRI sequences, and 3D vs. 2D models, but not among imaging planes. The between-study heterogeneity was low (), and no statistically significant evidence of small-study effects or publication bias was detected. Conclusions: DL models for TKV in ADPKD patients demonstrated high precision compared to manual segmentation in a large, pooled sample with heterogeneous study designs and methods. While clinical implementation is not yet feasible, the current work demonstrates the technical and diagnostic efficacy of image-based DL segmentation models. - Open Access
- 15 February 2023
- Published by Elsevier in Brain Stimulation
- Vol. 16 (1) , 252-253
Subjective reporting of dizziness during bi-hemispheric TDCS of the right and left motor hand area
- 15 February 2023
- Published by Elsevier in Brain Stimulation
- Vol. 16 (1) , 252-253
Open AccessSubjective reporting of dizziness during bi-hemispheric TDCS of the right and left motor hand area
- Open Access
Short periods of bipolar anodal TDCS induce no instantaneous dose-dependent increase in cerebral blood flow in the targeted human motor cortex
Background Anodal transcranial direct current stimulation (aTDCS) of primary motor hand area (M1-HAND) can enhance corticomotor excitability. Yet, it is still unknown which current intensity produces the strongest effect on regional neural activity. Magnetic resonance imaging (MRI) combined with pseudo-continuous Arterial Spin Labeling (pc-ASL MRI) can map regional cortical blood flow (rCBF) and may thus be useful to probe the relationship between current intensity and neural response at the individual level. Objective Here we employed pc-ASL MRI to map acute rCBF changes during short-duration aTDCS of left M1-HAND. Using the rCBF response as a proxy for regional neuronal activity, we investigated if short-duration aTDCS produces an instantaneous dose-dependent rCBF increase in the targeted M1-HAND that may be useful for individual dosing. Methods Nine healthy right-handed participants received 30 seconds of aTDCS at 0.5, 1.0, 1.5, and 2.0 mA with the anode placed over left M1-HAND and cathode over the right supraorbital region. Concurrent pc-ASL MRI at 3 T probed TDCS-related rCBF changes in the targeted M1-HAND. Movement-induced rCBF changes were also assessed. Results Apart from a subtle increase in rCBF at 0.5 mA, short-duration aTDCS did not modulate rCBF in the M1-HAND relative to no-stimulation periods. None of the participants showed a dose-dependent increase in rCBF during aTDCS, even after accounting for individual differences in TDCS-induced electrical field strength. In contrast, finger movements led to robust activation of left M1-HAND before and after aTDCS. Conclusion Short-duration bipolar aTDCS does not produce instantaneous dose-dependent rCBF increases in the targeted M1-HAND at conventional intensity ranges. Therefore, the regional hemodynamic response profile to short-duration aTDCS may not be suited to inform individual dosing of TDCS intensity. Highlights Arterial spin labeling (ASL) MRI probed regional cerebral blood flow (rCBF) during anodal TDCS. Short-duration anodal TDCS was applied to left motor hand area (M1-HAND) at 0.5, 1.0, 1.5 and 2.0 mA. Anodal TDCS produced no instantaneous dose-dependent rCBF increase in left M1-HAND. Apart from a subtle increase at 0.5 mA, rCBF was not modified in M1-HAND during anodal TDCS.
Open AccessShort periods of bipolar anodal TDCS induce no instantaneous dose-dependent increase in cerebral blood flow in the targeted human motor cortex
Background Anodal transcranial direct current stimulation (aTDCS) of primary motor hand area (M1-HAND) can enhance corticomotor excitability. Yet, it is still unknown which current intensity produces the strongest effect on regional neural activity. Magnetic resonance imaging (MRI) combined with pseudo-continuous Arterial Spin Labeling (pc-ASL MRI) can map regional cortical blood flow (rCBF) and may thus be useful to probe the relationship between current intensity and neural response at the individual level. Objective Here we employed pc-ASL MRI to map acute rCBF changes during short-duration aTDCS of left M1-HAND. Using the rCBF response as a proxy for regional neuronal activity, we investigated if short-duration aTDCS produces an instantaneous dose-dependent rCBF increase in the targeted M1-HAND that may be useful for individual dosing. Methods Nine healthy right-handed participants received 30 seconds of aTDCS at 0.5, 1.0, 1.5, and 2.0 mA with the anode placed over left M1-HAND and cathode over the right supraorbital region. Concurrent pc-ASL MRI at 3 T probed TDCS-related rCBF changes in the targeted M1-HAND. Movement-induced rCBF changes were also assessed. Results Apart from a subtle increase in rCBF at 0.5 mA, short-duration aTDCS did not modulate rCBF in the M1-HAND relative to no-stimulation periods. None of the participants showed a dose-dependent increase in rCBF during aTDCS, even after accounting for individual differences in TDCS-induced electrical field strength. In contrast, finger movements led to robust activation of left M1-HAND before and after aTDCS. Conclusion Short-duration bipolar aTDCS does not produce instantaneous dose-dependent rCBF increases in the targeted M1-HAND at conventional intensity ranges. Therefore, the regional hemodynamic response profile to short-duration aTDCS may not be suited to inform individual dosing of TDCS intensity. Highlights Arterial spin labeling (ASL) MRI probed regional cerebral blood flow (rCBF) during anodal TDCS. Short-duration anodal TDCS was applied to left motor hand area (M1-HAND) at 0.5, 1.0, 1.5 and 2.0 mA. Anodal TDCS produced no instantaneous dose-dependent rCBF increase in left M1-HAND. Apart from a subtle increase at 0.5 mA, rCBF was not modified in M1-HAND during anodal TDCS.
- 11 July 2016
- Published by SAGE Publications in Cephalalgia
- Vol. 37 (2) , 114-124
Calcitonin gene-related peptide induced migraine attacks in patients with and without familial aggregation of migraine
Background: Calcitonin gene-related peptide provokes migraine attacks in 65% of patients with migraine without aura. Whether aggregation of migraine in first-degree relatives (family load) or a high number of risk-conferring single nucleotide polymorphisms contributes to migraine susceptibility to calcitonin gene-related peptide infusion in migraine patients is unknown. We hypothesized that genetic enrichment plays a role in triggering of migraine and, therefore, migraine without aura patients with high family load would report more migraine attacks after calcitonin gene-related peptide infusion than patients with low family load. Methods: We allocated 40 previously genotyped migraine without aura patients to receive intravenous infusion of 1.5 µg/min calcitonin gene-related peptide and recorded migraine attacks including headache characteristics and associated symptoms. Information of familial aggregation was obtained by telephone interview of first-degree relatives using a validated semi-structured questionnaire. Results: Calcitonin gene-related peptide infusion induced a migraine-like attack in 75% (12 out of 16) of patients with high family load compared to 52% (12 out of 23) with low family load ( P = 0.150). In addition, we found that the migraine response after calcitonin gene-related peptide was not associated with specific or a high number of risk-conferring single nucleotide polymorphisms of migraine without aura. Conclusion: We found no statistical association between familial aggregation of migraine and hypersensitivity to calcitonin gene-related peptide infusion in migraine without aura patients. We also demonstrated that the currently known single nucleotide polymorphisms conferring risk of migraine without aura have no additive effect on calcitonin gene-related peptide induced migraine-like attacks.
- 11 July 2016
- Published by SAGE Publications in Cephalalgia
- Vol. 37 (2) , 114-124
Calcitonin gene-related peptide induced migraine attacks in patients with and without familial aggregation of migraine
Background: Calcitonin gene-related peptide provokes migraine attacks in 65% of patients with migraine without aura. Whether aggregation of migraine in first-degree relatives (family load) or a high number of risk-conferring single nucleotide polymorphisms contributes to migraine susceptibility to calcitonin gene-related peptide infusion in migraine patients is unknown. We hypothesized that genetic enrichment plays a role in triggering of migraine and, therefore, migraine without aura patients with high family load would report more migraine attacks after calcitonin gene-related peptide infusion than patients with low family load. Methods: We allocated 40 previously genotyped migraine without aura patients to receive intravenous infusion of 1.5 µg/min calcitonin gene-related peptide and recorded migraine attacks including headache characteristics and associated symptoms. Information of familial aggregation was obtained by telephone interview of first-degree relatives using a validated semi-structured questionnaire. Results: Calcitonin gene-related peptide infusion induced a migraine-like attack in 75% (12 out of 16) of patients with high family load compared to 52% (12 out of 23) with low family load ( P = 0.150). In addition, we found that the migraine response after calcitonin gene-related peptide was not associated with specific or a high number of risk-conferring single nucleotide polymorphisms of migraine without aura. Conclusion: We found no statistical association between familial aggregation of migraine and hypersensitivity to calcitonin gene-related peptide infusion in migraine without aura patients. We also demonstrated that the currently known single nucleotide polymorphisms conferring risk of migraine without aura have no additive effect on calcitonin gene-related peptide induced migraine-like attacks.