Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2024 Nov 2;45(16):e70060. doi: 10.1002/hbm.70060

Frontoparietal Structural Network Disconnections Correlate With Outcome After a Severe Stroke

Lukas Frontzkowski 1, Felix Fehring 1, Benedikt M Frey 1, Paweł P Wróbel 1, Antonia Reibelt 1, Focko Higgen 1, Silke Wolf 1, Winifried Backhaus 1, Hanna Braaß 1, Philipp J Koch 2, Chi‐un Choe 1, Marlene Bönstrup 1,3, Bastian Cheng 1, Götz Thomalla 1, Christian Gerloff 1, Fanny Quandt 1, Robert Schulz 1,✉
PMCID: PMC11530704  PMID: 39487651

ABSTRACT

Structural disconnectome analyses have provided valuable insights into how a stroke lesion results in widespread network disturbances and how these relate to deficits, recovery patterns, and outcomes. Previous analyses have primarily focused on patients with relatively mild to moderate deficits. However, outcomes vary among survivors of severe strokes, and the mechanisms of recovery remain poorly understood. This study assesses the association between lesion‐induced network disconnection and outcome after severe stroke. Thirty‐eight ischaemic stroke patients underwent MRI brain imaging early after stroke and longitudinal clinical follow‐up. Lesion information was integrated with normative connectome data to infer individual disconnectome profiles on a localized regional and region‐to‐region pathway level. Ordinal logistic regressions were computed to link disconnectome information to the modified Rankin Scale after 3–6 months. Disconnections of ipsilesional frontal, parietal, and temporal cortical brain areas were significantly associated with a worse motor outcome after a severe stroke, adjusted for the initial deficit, lesion volume, and age. The analysis of the underlying pathways mediating this association revealed location‐specific results: For frontal, prefrontal, and temporal brain areas, the association was primarily driven by relatively sparse intrahemispheric disconnections. In contrast, the ipsilesional primary motor cortex, the dorsal premotor cortex, and various parietal brain regions showed a remarkable involvement of either frontoparietal intrahemispheric or additionally interhemispheric disconnections. These results indicate that localized disconnection of multiple regions embedded in the structural frontoparietal network correlates with worse outcomes after severe stroke. Specifically, primary motor and parietal cortices might gain particular importance as they structurally link frontoparietal networks of both hemispheres. These data shed novel light on the significance of distinct brain networks for recovery after a severe stroke.

Keywords: disconnectome, indirect, lesion, mapping, MRI, nemo, recovery

Network disconnection was analyzed in 38 severely impaired acute stroke patients using lesion masks and normative connectome data. Logistic regression models were computed to associate region‐ and pathway‐related disconnections with the outcome. Localized structural frontoparietal network disconnections were significantly linked to a worse outcome after stroke.

graphic file with name HBM-45-e70060-g002.jpg

Abbreviations

ChaCo

change of connectivity

LOOA

leave‐one‐out analysis

MNI

Montreal Neurological Institute

mRS

modified Rankin Scale

NIHSS

National Institutes of Health Stroke Scale

1. Introduction

Acute stroke treatment with thrombolysis and mechanical thrombectomy has significantly enhanced survival rates and overall outcomes after ischemic stroke. However, particularly in patients exhibiting severe initial deficits, recovery is incomplete, and survivors are left with functional disabilities that critically impede private and professional independence (Katan and Luft 2018; Jensen et al. 2023; Bonkhoff et al. 2022). In recent years, structural brain imaging has offered invaluable insights into the recovery processes following stroke, aiming to explain the intersubject variability in stroke patient outcomes (Koch, Schulz, and Hummel 2016). It not only sought to improve outcome modeling, ultimately to guide clinical decision‐making and rehabilitation strategies (Stinear et al. 2017), but structural brain imaging also aimed to enhance our mechanistic understanding of how the brain and its networks respond to stroke lesions and to identify which structural properties may contribute to a favorable outcome (Bonkhoff and Grefkes 2022; Dulyan et al. 2022; Rivier et al. 2023). While investigations of specific tracts such as the corticospinal tract have dominated the field for years (Koch, Schulz, and Hummel 2016), a network perspective, including the analysis of lesion‐induced disturbances of network connections, has been more recently incorporated with varying methodological approaches (Thiebaut de Schotten, Foulon, and Nachev 2020; Sperber, Griffis, and Kasties 2022). Notably, integrating simple lesion information into fine‐grained normative healthy connectome data for indirect disconnectome‐symptom mapping (Sperber, Griffis, and Kasties 2022) may facilitate the transfer from neuroscience to potential clinical applications.

Patterns of network disconnectivity have been associated with poststroke deficits in subacute stroke patients (Salvalaggio et al. 2020), with variability in recovery trajectories over time (Dulyan et al. 2022; Kuceyeski et al. 2016; Ding et al. 2023; Koch et al. 2021; Egger et al. 2021), with treatment gains during motor rehabilitation paradigms (D'Imperio et al. 2021), and fatigue (Schaechter et al. 2023) or cognitive functions (Kolskar et al. 2022) in the chronic stage of recovery. Here, profiles of single regions disconnected from larger brain networks and profiles of disturbed interconnecting pathways at a pairwise region‐to‐region level have been found to carry relevant information (Kuceyeski et al. 2016). Treatment‐related insights have been provided by assessing the effects of thrombolysis on network disconnections (Schlemm et al. 2021) and topology (Schlemm et al. 2022), or the influence of tract‐specific disconnections before thrombectomy for outcome inference (Koch et al. 2023). As one significant limitation, however, at least some of these disconnectome analyses have primarily focused on patients with relatively mild to moderate deficits with initial National Institutes of Health Stroke Scale (NIHSS) scores ranging between 1 and 3 (Dulyan et al. 2022; Salvalaggio et al. 2020; Ding et al. 2023; Kolskar et al. 2022) or up to 7 (Kuceyeski et al. 2016; Schlemm et al. 2022). Data regarding severely impaired patients are limited. However, particularly survivors of severe stroke are characterized by significant variability in recovery patterns and uncertainty regarding underlying, supportive mechanisms, on network neuroscience levels.

Therefore, the present work aimed to assess the relationship between lesion‐induced network disconnection profiles and subsequent outcomes after severe stroke. First‐ever supratentorial ischemic stroke patients with severe deficits underwent clinical testing and brain imaging within 2 weeks after the event. Lesion information was integrated with normative connectome data to infer individual disconnectome profiles on a localized regional and region‐to‐region level. Ordinal logistic regression modeling was conducted to link disconnectome information to the global disability as the clinical outcome, operationalized by the modified Rankin Scale (mRS) after 3–6 months. We hypothesized that regional disconnection, particularly of ipsilesional frontal and parietal brain regions (Reibelt, Quandt, and Schulz 2023) from larger networks, will significantly correlate with outcome variability. Frontal and parietal areas will differ in the extent to which specific intra‐ or interhemispheric disconnectivity will drive the overall region's importance. Finally, compared to previous publications studying less impaired cohorts, the present findings might allow novel insights into severe stroke recovery (Bonkhoff et al. 2022; Park and Kim 2023), not only aiming to improve current prediction models, but also paving the way toward patient‐tailored innovative treatment strategies, including noninvasive brain stimulation (Koch and Hummel 2017).

2. Methods

2.1. Demographic Data

The datasets of 38 severely affected stroke patients from two independent, previously published observational studies (Cohort 1, C1 (Bonstrup et al. 2019), Cohort 2, C2 (Backhaus et al. 2021)) were included in this analysis. All 38 acute stroke patients were admitted to the University Medical Center Hamburg‐Eppendorf and pooled into one single cohort by following inclusion criteria: first‐ever unilateral ischemic supratentorial stroke, upper extremity motor deficit involving hand function, mRS > 3 or Barthel‐Index (BI) ≤ 30, no history of previous neurological or psychiatric illness, and age ≥ 18 years. This approach of cohort integration, that is, pooling, is in line with our last analyses (Backhaus et al. 2021; Rojas Albert et al. 2022; Sadeghihassanabadi et al. 2022). Magnetic resonance imaging (MRI) and clinical testing were performed in the first 3–14 days after stroke onset (time point T 1). In the late subacute stage of stroke recovery (3 months poststroke), clinical testing was repeated and defined as follow‐up time point T 2. A subset of patients was not eligible for clinical testing at timepoint T 2; thus, data were only available after 6 months. In this analysis, the initial symptom burden was operationalized by the NIHSS score at T 1, and global disability at T 2 was quantified by mRS. The original studies were conducted in line with the ethical Declaration of Helsinki and were granted permission by the local ethics committee of the Chamber of Physicians Hamburg. All participants or their legal guardians provided informed consent.

2.2. Image Acquisition

A 3T Skyra MRI scanner (Siemens Healthineers, Erlangen, Germany) equipped with a 32‐channel head coil was used to obtain structural high‐resolution T 1‐weighted images applying a three‐dimensional magnetization‐prepared rapid gradient echo sequence (MPRAGE) [repetition time (TR) = 2500 ms, echo time (TE) = 2.12 ms, flip angle 9°, 256 coronal slices with a voxel size of 0.8 × 0.8 × 0.9 mm3, field of view (FOV) = 240 mm]. T 2‐weighted images were also acquired by using a fluid‐attenuated inversion recovery sequence (TR = 9000 ms, TE = 86 ms, TI = 2500 ms, flip angle 150°, 43 transversal slices with a voxel size of 0.7 × 0.7 × 3.0 mm3, FOV = 230 mm) for stroke lesion delineation.

2.3. Quantification of Structural Disconnection

Stroke lesions were semiautomatically delineated with SNAP‐ITK (Yushkevich et al. 2006) and available from our previous analyses (Backhaus et al. 2021; Rojas Albert et al. 2022; Sadeghihassanabadi et al. 2022; Nemati et al. 2022). The resulting lesion masks were registered to a 1‐mm3 Montreal Neurological Institute (MNI) template. To compute structural disconnection profiles for each subject, masks were analyzed using the network modification NEMO2 toolbox, which computes the disruption of white matter connections by overlaying an MNI‐registered lesion mask onto a normative structural connectome derived from 420 healthy Human Connectome Project participants (Kuceyeski et al. 2013). In brief, a change of connectivity (ChaCo) measure is computed, which resembles the extent of disrupted streamlines divided by total streamlines for each voxel or an atlas‐based parcellation. For our analyses, we applied the Brainnetome atlas as it is based on a connectional mapping approach and thus applicable for testing the extent of disconnection between brain regions (Fan et al. 2016). ChaCo values were computed at a regional level and a pairwise region‐to‐region level (i.e., addressing interconnecting pathways that contribute to the connectivity of each area). In line with previous recommendations, all values below 0.02 were disregarded due to noise artifacts (Kuceyeski et al. 2013). To increase comparability between stroke patients, ChaCo values were flipped to the left side when the stroke site was in the right hemisphere. We excluded all regions with a critically skewed data distribution (i.e., skew either > 1.3 or < −1.3). Out of 246 Brainnetome areas, 100 regions were eligible for further statistical analyses (Figure 1, Figure S1, Table S1). Two alternative skew thresholds (+/−1.0 and +/−1.5, respectively) were explored for sensitivity analyses. However, the extent and distribution of eligible regions for modeling remained stable (Figure S2).

FIGURE 1.

FIGURE 1

Open in a new tab

Quantification of structural disconnection: Binary lesions masks in MNI space are utilized to compute regional and pairwise region‐to‐region change of connectivity (ChaCo) scores by overlaying the lesion on a normative structural connectome. The ChaCo scores are computed by calculating the number of intersected streamlines on a regional (total % of disconnection for a given brain area) and pairwise (total % disconnection between a pair of brain regions) level. ChaCo scores were computed for 246 areas in the Brainnetome atlas. The final analysis did not consider regions exhibiting skewed data distribution (either close to 0% or 100% ChaCo). Structural disconnection of 100 regions primarily belonging to the ipsilesional hemisphere was statistically analyzed (see Figure S1). ChaCo, change of connectivity; CL, contralesional; IL, ipsilesional; MNI, Montreal Institute of Neurology.

2.4. Statistical Analysis

Statistical analyses were carried out in R version 4.2.3 (R: A Language and Environment for Statistical Computing 2021). In the first step, 100 ordinal logistic regression models (function polr from the MASS package) were fitted to relate the regional structural disconnection value obtained at T 1 to mRS at T 2 (Venables, Ripley, and Venables 2002). Models were adjusted for lesion volume, NIHSS at T 1, and age. Lesion volumes were log10‐transformed to improve data distribution. To address multicollinearity, we included lesion volume and NIHSS at T 1 after residualization against the regional disconnection, in line with recent studies (Rojas Albert et al. 2022; Sadeghihassanabadi et al. 2022). Models were corrected for multiple comparisons via the Bonferroni–Holm false‐discovery‐rate (FDR) method and a leave‐one‐out analysis (LOOA) was applied to test the robustness of significant regional findings. Adjusted odds ratios (OR) for a one‐unit increase of 0.01 in the ChaCo value are given for significant models with respective 95% confidence intervals and p FDR values. OR values > 1 indicate a higher risk of scoring 1 point higher on mRS at T 2. In those regions whose overall level of disconnection was linked to the outcome after stroke, a pairwise region‐to‐region disconnectome analysis was added to explore which connections or networks might drive the statistical contribution of that region in the first step outcome model. Again, connections exhibiting skewed data were omitted from this second analysis step. Hence, for each region, up to 245 models (connections between the region and the remaining 245 regions of the Brainnetome atlas) were fitted. FDR correction for multiple testing and LOOA were applied. Statistical significance was set at p FDR < 0.05.

3. Results

3.1. Demographics and Clinical Data

Table 1 shows the individual demographic and clinical data. The cohort comprised 38 severely affected stroke patients with supratentorial lesions (19 females, 23 right‐hemispheric strokes, two left‐handed, median age 73 years). The median NIHSS score was 9 at T 1. The median mRS was 4 at T 1 and 3 at T 2. The mean lesion volume was 58.25 mL with a standard deviation of 70.01. Figure 2 shows a heatmap visualizing the distribution of the MNI‐registered binary stroke lesions.

TABLE 1.

Demographic overview.

ID Study Age Sex Lesion side Lesion volume, mL NIHSS T 1 mRS T 2 Thrombolysis Thrombectomy TICI
1 C1 49 Female Left 53.80 10 2 Yes Yes NA
2 C1 73 Female Left 5.82 9 4 No No No
3 C1 65 Male Left 6.57 8 3 No No No
4 C1 81 Male Left 1.73 4 3 No No No
5 C1 48 Male Left 24.44 7 2 Yes Yes 3
6 C1 87 Female Left 0.99 1 1 No No No
7 C1 43 Male Right 79.82 13 2 Yes Yes 3
8 C1 56 Male Right 2.53 13 4 No No No
9 C1 69 Male Right 25.09 3 1 Yes Yes 3
10 C1 73 Female Right 26.76 3 1 Yes Yes NA
11 C1 50 Male Right 25.52 7 2 Yes Yes 2B
12 C1 77 Female Right 9.1 8 3 No No No
13 C1 70 Female Right 74.36 5 1 Yes Yes 2A
14 C1 85 Female Right 16.75 7 4 No No No
15 C1 47 Male Right 2.58 6 3 Yes No No
16 C1 50 Male Right 50.06 4 1 Yes Yes 3
17 C2 78 Male Left 58.12 17 5 a No No No
18 C2 59 Female Left 45.24 18 4 No No NA
19 C2 83 Female Left 101.41 20 6 No No No
20 C2 84 Male Left 3.60 8 3 a No No NA
21 C2 63 Male Left 55.82 13 1 No No No
22 C2 73 Female Left 14.40 9 3 Yes Yes 2B
23 C2 80 Female Left 20.52 11 4 a No Yes 2B
24 C2 78 Female Left 33.61 10 3 No No No
25 C2 74 Male Left 303.32 24 5 a Yes Yes 2B
26 C2 67 Female Right 5.02 9 1 Yes Yes NA
27 C2 76 Male Right 100.96 11 3 Yes Yes 3
28 C2 71 Male Right 75.22 15 6 No Yes NA
29 C2 77 Female Right 286.75 11 4 a Yes No No
30 C2 71 Female Right 38.38 9 3 a Yes Yes 2A
31 C2 58 Male Right 97.96 13 5 Yes Yes NA
32 C2 67 Female Right 7.36 11 1 a Yes No No
33 C2 80 Male Right 108.41 16 6 Yes Yes 0
34 C2 79 Female Right 120.41 8 4 a Yes Yes 2B
35 C2 85 Female Right 33.49 15 5 No Yes 2B
36 C2 78 Male Right 178.07 17 4 Yes Yes 3
37 C2 73 Female Right 27.56 5 1 Yes Yes 2A
38 C2 76 Male Right 91.82 15 4 Yes Yes NA
Open in a new tab

Abbreviation: NA, data not available.

a

Follow‐up data were available after 6 months.

FIGURE 2.

FIGURE 2

Open in a new tab

Lesion heatmap. The number of patients with lesioned voxel is color‐coded with red colors, indicating high overlap between lesions. Numbers above the maps indicate the z‐value (slice) in MNI space. L, left; R, right.

3.2. Regional Disconnection and Outcome After Stroke

Ordinal logistic regression modeling revealed positive associations between lesion‐induced regional disconnections primarily affecting frontal, parietal, and temporal brain areas and worse outcomes after stroke (Figure 3A, Table 2). Highest OR values were computed for the precentral and postcentral gyrus. As an example, stacked histograms of mRS at T 2 are plotted for the level of disconnection of the primary motor cortex (after the median split of the patients into groups exhibiting lower and higher ChaCo values for illustration purposes, Figure 3B). None of the 17 subcortical regions tested showed significant associations. For sensitivity analyses, we recomputed all models with (1) sex or (2) with the factors thrombolysis and thrombectomy as additional covariates. Overall, the results remained stable, and we did not detect significant associations after adjusting for multiple comparisons between sex or acute interventions and outcomes in our models (see Figures S3 and S4).

FIGURE 3.

FIGURE 3

Open in a new tab

(A) Surface visualization of ORs corresponding to the significant associations between regional disconnection and outcome. Results are corrected for age, lesion volume, and NIHSS at T 1. Non‐colorized regions were either excluded due to a skewed data distribution or lacked a significant association. CL, contralesional; IL, ipsilesional; OR, odds ratio. (B) Shift plot illustrating the dichotomized structural disconnection of one representative Brainnetome region “A4ul_L” which represents the upper limb area of the left primary motor cortex. Outcome distribution is displayed for high and low disconnection values in yellow and red colors, indicating lower and higher mRS scores, respectively.

TABLE 2.

Results of ordinal logistic regressions between regional structural disconnection of 28 cortical brain regions and outcome after severe stroke.

Lobe Anatomical region Atlas region Adjusted OR (95% CI) p FDR
Frontal Inferior frontal gyrus IFS_L 1.03 (1.01–1.06) 0.034
Middle frontal gyrus A46_L 1.05 (1.02–1.08) 0.025
Middle frontal gyrus A6vl_L 1.04 (1.02–1.07) 0.019
Middle frontal gyrus A8vl_L 1.04 (1.02–1.07) 0.011
Middle frontal gyrus A9/46d_L 1.05 (1.02–1.08) 0.006
Middle frontal gyrus A9/46v_L 1.04 (1.02–1.07) 0.009
Middle frontal gyrus IFJ_L 1.04 (1.02–1.06) 0.027
Paracentral lobule A1/2/3ll_L 1.06 (1.02–1.11) 0.050
Precentral gyrus A4hf_L 1.05 (1.02–1.08) < 0.001
Precentral gyrus A4ul_L 1.06 (1.03–1.10) < 0.001
Precentral gyrus A6cdl_L 1.06 (1.03–1.09) 0.009
Precentral gyrus A6cvl_L 1.04 (1.02–1.06) 0.025
Superior frontal gyrus A8dl_L 1.05 (1.02–1.09) 0.022
Parietal Inferior parietal lobule A39rd_L 1.05 (1.03–1.08) < 0.001
Inferior parietal lobule A39rv_L 1.04 (1.02–1.07) < 0.001
Inferior parietal lobule A40c_L 1.04 (1.02–1.07) < 0.001
Inferior parietal lobule A40rd_L 1.05 (1.03–1.08) < 0.001
Inferior parietal lobule A40rv_L 1.04 (1.02–1.07) 0.009
Postcentral gyrus A1/2/3tru_L 1.06 (1.02–1.10) 0.034
Postcentral gyrus A1/2/3ulhf_L 1.05 (1.03–1.08) < 0.001
Postcentral gyrus A2_L 1.06 (1.03–1.09) < 0.001
Superior parietal lobule A5l_L 1.05 (1.03–1.08) < 0.001
Temporal Middle temporal gyrus A21c_L 1.04 (1.02–1.07) 0.020
Middle temporal gyrus A37dl_L 1.05 (1.02–1.07) < 0.001
Posterior superior temporal sulcus cpSTS_L 1.04 (1.02–1.06) < 0.001
Posterior superior temporal sulcus rpSTS_L 1.03 (1.01–1.06) 0.027
Superior temporal gyrus A22c_L 1.04 (1.02–1.06) 0.007
Superior temporal gyrus A41/42_L 1.04 (1.02–1.06) 0.010
Open in a new tab

Note: Results are corrected for age, lesion volume, and NIHSS at T 1. Adjusted OR with CI 95% and p FDR values are given. OR was adjusted so that a one‐unit increase in the ChaCo value would relate to an increase of 0.01.

Abbreviations: _L, left hemisphere, corresponding to the ipsilesional hemisphere; OR, odds ratio.

3.3. Pairwise Disconnection Patterns and Outcome After Stroke

Pairwise disconnection analyses provided additional insights into possible drivers of significant disconnectome‐outcome relationships on a pathway or region‐to‐region level, results are depicted in Figure 4. The primary finding was the involvement of sparse intrahemispheric disconnections in frontal and prefrontal areas. Only labels contributing to the primary motor and dorsolateral premotor cortices showed involvement of stronger frontoparietal and interhemispheric disconnectivity. Similarly, temporal regional disconnections were also driven by rather sparse intrahemispheric pathways except in posterior middle temporal areas. In contrast to these frontal and temporal brain regions, the importance of parietal cortical disconnectivity for the outcome appeared to be driven not only by the structural state of extensive strong frontoparietal intrahemispheric but also extensive interhemispheric connections. On visual inspection, these connections appeared to be most pronounced for posterior parietal cortices and inferior temporal cortices (Figure 4).

FIGURE 4.

FIGURE 4

Open in a new tab

Glass brain plots showing significant pairwise region‐to‐region disconnection profiles for every significant regional disconnection per lobe. Edges that showed a significant structure‐outcome relation are colorized according to their adjusted OR values. Frontal and temporal regions exhibit a predominantly intrahemispheric disconnection profile, while parietal regions show a strong edge disconnection not only to frontal areas of the ipsilesional, but also to frontal and parietal brain areas of the contralesional hemisphere. Results are corrected for age, lesion volume, and NIHSS at T 1. Adjusted OR are color‐coded. A list of region names can be found in the Table S1. CL, contralesional; IL, ipsilesional; OR, odds ratio.

4. Discussion

The main finding of the present study was that lesion‐induced disconnections of frontal, parietal, and temporal cortical brain areas, adjusted for the initial deficit, lesion volume, and age, are significantly associated with outcomes after severe stroke. The spatial distribution of the underlying pathways mediating the disconnection‐outcome association was not diffuse or unspecific. It revealed location‐specific results: For frontal, prefrontal, and most temporal brain areas, it was primarily driven by relatively sparse intrahemispheric disconnections. In contrast, the ipsilesional primary motor cortex, the dorsal premotor cortex, and various parietal brain regions showed a remarkable involvement of either frontoparietal intrahemispheric or additionally interhemispheric disconnections. These results indicate that localized disconnection of multiple regions embedded in the structural frontoparietal network correlates with outcome after severe stroke and that primary motor and parietal cortices might gain particular importance as they structurally link frontoparietal networks of both hemispheres.

These findings extend previous data regarding the functional importance of lesion‐induced disconnectome profiles for outcome inference after stroke. One study mapped cognitive functions, basic mobility, and daily activity, assessed after 6 months, to disconnectomes in 40 acute stroke patients. Intrahemispheric network disconnections primarily predicted functional scores in that cohort with a rather moderate symptom burden. These were localized in the frontal lobes for applied cognitive tests and basic mobility, and in frontal and frontoparietal networks for daily activities, respectively. Notably, for the latter, a pairwise region‐to‐region approach was found to be particularly informative for outcome prediction when compared to lesion volumes or regional disconnectivity values alone (Kuceyeski et al. 2016). In contrast to that study, the present cohort included more severely impaired patients. The current analyses suggest that interhemispheric parietal disconnectivity might play a crucial role in the recovery of stroke patients with severe impairments, although variations exist in methodologies, such as statistical approaches or clinical assessment tools utilized.

A recent study that derived connectome data from tract‐related microstructure and aimed at predicting the natural recovery in acute stroke patients corroborates our findings (Koch et al. 2021). The authors reported that connections within the parietal cortex, including the intraparietal sulcus and the superior parietal gyrus on both hemispheres, with a specificity of the unaffected hemisphere for severely impaired patients, improved prediction of recovery (Koch et al. 2021). Various electrophysiological, functional, and structural imaging studies have already indicated that interhemispheric connections and the contralesional hemisphere might be particularly important for stroke recovery of severely impaired patients. For instance, stimulation studies have revealed that suppressing regional excitability (Bradnam et al. 2012) and the perturbation of information processing (Lotze et al. 2006; Tscherpel et al. 2020) on the contralesional hemisphere could impair motor functions in chronic stroke patients. Functional imaging has indicated widespread increases in contralesional brain activation from frontal to occipitoparietal brain areas (Reibelt, Quandt, and Schulz 2023; Rehme et al. 2012; Favre et al. 2014), particularly after a severe stroke. It has been argued that attentional processes or adaptive motor learning strategies to compensate for motor impairment might drive such changes in activation (Ward et al. 2003; Buma et al. 2016; Marshall et al. 2009). Coupling analyses based on resting‐state functional data found that interhemispheric rather than intrahemispheric connection profiles were significantly related to acute deficits (Carter et al. 2010) and later outcomes (Lee et al. 2018; Wang et al. 2010). Another study investigated 18 chronic stroke patients and reported that patients with a good outcome exhibited a stronger interhemispheric parietal coupling between the anterior intraparietal sulcus than patients with a poor outcome (Hensel et al. 2022). Structural analyses have complemented this largely coherent picture. For instance, one study showed that contralesional cortical thickness assessed early after a severe stroke could inform about subsequent recovery (Rojas Albert et al. 2022). Another study found that white matter disconnections affecting parietal cortices and the posterior corpus callosum were correlated with motor rehabilitation outcomes in chronic stroke patients (D'Imperio et al. 2021).

Apart from interhemispheric connections, the importance of the regional disconnection of the primary motor cortex, the dorsal premotor cortex, and multiple parietal cortices along the intraparietal sulcus was additionally driven by intrahemispheric reciprocal frontoparietal disconnectivity. This finding corroborates studies that already emphasized that ipsilesional frontoparietal network integrity, characterized by regional structural or connectivity analyses, is an important factor for a favorable outcome. Of note, this might particularly apply to severe stroke (Reibelt, Quandt, and Schulz 2023) as severity‐specific roles of frontoparietal network involvement were statistically indicated by subgroup analyses (Hordacre et al. 2021) or, at least with a solid statistical trend, by previous interaction modeling (Schulz et al. 2017).

Collectively, when considering both inter‐ and intrahemispheric findings, particularly in the parietal regions, the current disconnectome data emphasize that parietal brain regions may hold particular significance for recovery after a severe stroke. This is because they are involved, at least structurally, in extensive frontoparietal networks that span both hemispheres.

Interestingly, only sparse interhemispheric disconnections mediated the behavioral effect of most frontal brain regions, except for the upper limb area of the ipsilesional primary motor cortex. At least from this structural disconnectome perspective, secondary motor areas of the frontal lobe might, therefore, be primarily involved in recovery processes after stroke via intrahemispheric connections. This interpretation would find compelling parallels in various functional connectivity studies reporting changes in intrahemispheric premotor–motor couplings and significant coupling‐outcome relationships more frequently than interhemispheric connections. Consistent and functionally relevant alterations in interhemispheric connectivity were found for the primary motor cortices of both hemispheres. Preserved or increased connectivity was positively related to a better outcome, in the general stroke population (Rehme and Grefkes 2013), and patients with severe hand weakness (Min et al. 2020). In agreement, previous structural studies (Koch, Schulz, and Hummel 2016) have linked interhemispheric connectivity with corticospinal tract damage, bilateral brain activation, and motor impairment (Wang et al. 2012; Paul et al. 2023) or functional improvement during motor rehabilitation (Lindenberg et al. 2012). The present structural disconnectome study adds cross‐sectional imaging and longitudinal clinical data to this existing research and underlines that the primary motor cortex does not act solely as a key area within the ipsilesional motor network, but also as a crucial hub maintaining functionally important structural connections to the contralesional hemisphere. Their disturbances impede recovery after severe stroke beyond otherwise important influential factors such as initial symptom burden and mere lesion volume.

Given these localized disconnectome‐outcome relationships in severe stroke, the question emerges how such insights can add to stroke recovery research. One idea arises from recent ambitions to develop noninvasive brain stimulation toward patient‐tailored treatment strategies (Koch and Hummel 2017; Ovadia‐Caro et al. 2019) to overcome difficulties translating one‐suits‐all approaches from neuroscience to clinical applications (Tedesco Triccas et al. 2016; Schulz, Gerloff, and Hummel 2013). For instance, patients with preserved interhemispheric connections between primary motor cortices or parietal brain regions might be particularly susceptible to bilateral brain stimulation. First attempts to uncover possible patterns suitable for patient stratification have been recently published in very small cohorts for bilateral motor cortex stimulation combined with motor therapy (Lindenberg et al. 2012) and contralesional parietal cortex stimulation (Hensel et al. 2022) in chronic stroke.

There are several important limitations to note. First, the cohort comprised severely impaired stroke patients with rather large lesions primarily localizing in the vascular territory of the middle cerebral artery. This introduces a systematic bias that is likely to explain (1) a relevant portion of the disconnectome profiles, both on the regional and region‐to‐region level, and (2) why any subcortical region, of which all showed very high disconnection values, could be linked to outcome variability due to reduced variance across the individual patients. Second, all stroke lesions were flipped to the left hemisphere. Disconnectome values in the contralesional hemisphere exhibited either a too high or low skew and were thus not eligible for statistical analysis. Only 100 regions were included in the first step of the analysis. A larger sample size might result in a better data distribution on both hemispheres, which might, in turn, lead to the detection of further disconnection‐outcome associations, particularly in the light of potentially more sensitive hemisphere‐specific analyses (Rohrig et al. 2023). Third, the present cohort comprised a relatively small group of early subacute ischemic stroke patients. Regression models were corrected for multiple testing, and a LOOA was additionally carried out to improve the specificity of the results at the cost of reduced sensitivity. The analyses were correlative. Future studies combining larger samples and predictive modeling, including out‐of‐sample validation, would be needed to further test the present findings' robustness and the clinical significance in terms of additive effects in the combination of multiple disconnectivity measures. Finally, follow‐up clinical data were acquired in the late subacute stage of recovery. The primary outcome was global disability, operationalized by mRS, dominated by motor functions. Hence, the extent to which the findings might generalize to other time points or other more precise functional clinical scores or domains such as language or cognition remains an exciting topic for upcoming studies.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1.

HBM-45-e70060-s001.pdf (422.5KB, pdf)

Funding: This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) SFB 936—178316478, projects C1 to C.G., C2 to G.T., and SFB TRR169 project A3 with the National Science Foundation of China (NSFC) in project Crossmodal Learning to C.G., and the Else Kröner‐Fresenius‐Stiftung (2016_A214 to R.S.). R.S. and C.U.C. are supported by an Else Kröner Exzellenzstipendium from the Else Kröner‐Fresenius‐Stiftung (2020_EKES.16 to R.S., 2018_EKES.04 to C.U.C.). F.Q. is supported by the Gemeinnützige Hertie‐Stiftung (Hertie Network of Excellence in Clinical Neuroscience). Open Access funding enabled and organized by Projekt DEAL. We acknowledge financial support from the Open Access Publication Fund of UKE ‐ Universitätsklinikum Hamburg‐Eppendorf.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  1. Backhaus, W. , Braass H., Higgen F. L., Gerloff C., and Schulz R.. 2021. “Early Parietofrontal Network Upregulation Relates to Future Persistent Deficits After Severe Stroke—A Prospective Cohort Study.” Brain Communications 3, no. 2: fcab097. 10.1093/braincomms/fcab097. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bonkhoff, A. K. , and Grefkes C.. 2022. “Precision Medicine in Stroke: Towards Personalized Outcome Predictions Using Artificial Intelligence.” Brain 145, no. 2: 457–475. 10.1093/brain/awab439. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bonkhoff, A. K. , Hope T., Bzdok D., et al. 2022. “Recovery After Stroke: The Severely Impaired Are a Distinct Group.” Journal of Neurology, Neurosurgery & Psychiatry 93, no. 4: 369–378. 10.1136/jnnp-2021-327211. [DOI] [PubMed] [Google Scholar]
  4. Bonstrup, M. , Krawinkel L., Schulz R., et al. 2019. “Low‐Frequency Brain Oscillations Track Motor Recovery in Human Stroke.” Annals of Neurology 86, no. 6: 853–865. 10.1002/ana.25615. [DOI] [PubMed] [Google Scholar]
  5. Bradnam, L. V. , Stinear C. M., Barber P. A., and Byblow W. D.. 2012. “Contralesional Hemisphere Control of the Proximal Paretic Upper Limb Following Stroke.” Cerebral Cortex 22, no. 11: 2662–2671. 10.1093/cercor/bhr344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Buma, F. E. , van Kordelaar J., Raemaekers M., van Wegen E. E. H., Ramsey N. F., and Kwakkel G.. 2016. “Brain Activation is Related to Smoothness of Upper Limb Movements After Stroke.” Experimental Brain Research 234, no. 7: 2077–2089. 10.1007/s00221-015-4538-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Carter, A. R. , Astafiev S. V., Lang C. E., et al. 2010. “Resting Interhemispheric Functional Magnetic Resonance Imaging Connectivity Predicts Performance After Stroke.” Annals of Neurology 67, no. 3: 365–375. 10.1002/ana.21905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. D'Imperio, D. , Romeo Z., Maistrello L., et al. 2021. “Sensorimotor, Attentional, and Neuroanatomical Predictors of Upper Limb Motor Deficits and Rehabilitation Outcome After Stroke.” Neural Plasticity 2021: 8845685. 10.1155/2021/8845685. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ding, L. , Liu H., Jing J., et al. 2023. “Lesion Network Mapping for Neurological Deficit in Acute Ischemic Stroke.” Annals of Neurology 94, no. 3: 572–584. 10.1002/ana.26721. [DOI] [PubMed] [Google Scholar]
  10. Dulyan, L. , Talozzi L., Pacella V., Corbetta M., Forkel S. J., and Thiebaut de Schotten M.. 2022. “Longitudinal Prediction of Motor Dysfunction After Stroke: A Disconnectome Study.” Brain Structure & Function 227, no. 9: 3085–3098. 10.1007/s00429-022-02589-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Egger, P. , Evangelista G. G., Koch P. J., et al. 2021. “Disconnectomics of the Rich Club Impacts Motor Recovery After Stroke.” Stroke 52, no. 6: 2115–2124. 10.1161/STROKEAHA.120.031541. [DOI] [PubMed] [Google Scholar]
  12. Fan, L. , Li H., Zhuo J., et al. 2016. “The Human Brainnetome Atlas: A New Brain Atlas Based on Connectional Architecture.” Cerebral Cortex 26, no. 8: 3508–3526. 10.1093/cercor/bhw157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Favre, I. , Zeffiro T. A., Detante O., Krainik A., Hommel M., and Jaillard A.. 2014. “Upper Limb Recovery After Stroke Is Associated With Ipsilesional Primary Motor Cortical Activity: A Meta‐Analysis.” Stroke 45: 1077–1083. 10.1161/STROKEAHA.113.003168. [DOI] [PubMed] [Google Scholar]
  14. Hensel, L. , Lange F., Tscherpel C., et al. 2022. “Recovered Grasping Performance After Stroke Depends on Interhemispheric Frontoparietal connectivity.” Brain 146, no. 3: 1006–1020. 10.1093/brain/awac157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hordacre, B. , Lotze M., Jenkinson M., et al. 2021. “Fronto‐Parietal Involvement in Chronic Stroke Motor Performance When Corticospinal Tract Integrity is Compromised.” NeuroImage: Clinical 29: 102558. 10.1016/j.nicl.2021.102558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Jensen, M. , Sehner S., Cheng B., et al. 2023. “Patient‐Reported Quality of Life After Intravenous Alteplase for Stroke in the WAKE‐UP Trial.” Neurology 100, no. 2: e154–e162. 10.1212/WNL.0000000000201375. [DOI] [PubMed] [Google Scholar]
  17. Katan, M. , and Luft A.. 2018. “Global Burden of Stroke.” Seminars in Neurology 38, no. 2: 208–211. 10.1055/s-0038-1649503. [DOI] [PubMed] [Google Scholar]
  18. Koch, P. , Schulz R., and Hummel F. C.. 2016. “Structural Connectivity Analyses in Motor Recovery Research After Stroke.” Annals of Clinical Translational Neurology 3, no. 3: 233–244. 10.1002/acn3.278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Koch, P. J. , and Hummel F. C.. 2017. “Toward Precision Medicine: Tailoring interventional Strategies Based on Noninvasive Brain Stimulation for Motor Recovery After Stroke.” Current Opinion in Neurology 30, no. 4: 388–397. 10.1097/WCO.0000000000000462. [DOI] [PubMed] [Google Scholar]
  20. Koch, P. J. , Park C. H., Girard G., et al. 2021. “The Structural Connectome and Motor Recovery After Stroke: Predicting Natural Recovery.” Brain 144, no. 7: 2107–2119. 10.1093/brain/awab082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Koch, P. J. , Rudolf L. F., Schramm P., et al. 2023. “Preserved Corticospinal Tract Revealed by Acute Perfusion Imaging Relates to Better Outcome After Thrombectomy in Stroke.” Stroke 54, no. 12: 3081–3089. 10.1161/STROKEAHA.123.044221. [DOI] [PubMed] [Google Scholar]
  22. Kolskar, K. K. , Ulrichsen K. M., Richard G., et al. 2022. “Structural Disconnectome Mapping of Cognitive Function in Poststroke Patients.” Brain and Behavior: A Cognitive Neuroscience Perspective 12, no. 8: e2707. 10.1002/brb3.2707. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kuceyeski, A. , Maruta J., Relkin N., and Raj A.. 2013. “The Network Modification (NeMo) Tool: Elucidating the Effect of White Matter Integrity Changes on Cortical and Subcortical Structural connectivity.” Brain Connectivity 3, no. 5: 451–463. 10.1089/brain.2013.0147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kuceyeski, A. , Navi B. B., Kamel H., et al. 2016. “Structural Connectome Disruption at Baseline Predicts 6‐Months Post‐Stroke Outcome.” Human Brain Mapping 37, no. 7: 2587–2601. 10.1002/hbm.23198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lee, J. , Park E., Lee A., Chang W. H., Kim D. S., and Kim Y. H.. 2018. “Alteration and Role of Interhemispheric and Intrahemispheric Connectivity in Motor Network After Stroke.” Brain Topography 31, no. 4: 708–719. 10.1007/s10548-018-0644-9. [DOI] [PubMed] [Google Scholar]
  26. Lindenberg, R. , Zhu L. L., Rüber T., and Schlaug G.. 2012. “Predicting Functional Motor Potential in Chronic Stroke Patients Using Diffusion Tensor Imaging.” Human Brain Mapping 33, no. 5: 1040–1051. 10.1002/hbm.21266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lotze, M. , Markert J., Sauseng P., Hoppe J., Plewnia C., and Gerloff C.. 2006. “The Role of Multiple Contralesional Motor Areas for Complex Hand Movements After Internal Capsular Lesion.” Journal of Neuroscience: The Official Journal of the Society for Neuroscience 26, no. 22: 6096–6102. 10.1523/JNEUROSCI.4564-05.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Marshall, R. S. , Zarahn E., Alon L., Minzer B., Lazar R. M., and Krakauer J. W.. 2009. “Early Imaging Correlates of Subsequent Motor Recovery After Stroke.” Annals of Neurology 65, no. 5: 596–602. 10.1002/ana.21636. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Min, Y. S. , Park J. W., Park E., et al. 2020. “Interhemispheric Functional Connectivity in the Primary Motor Cortex Assessed by Resting‐State Functional Magnetic Resonance Imaging Aids Long‐Term Recovery Prediction among Subacute Stroke Patients With Severe Hand Weakness.” Journal of Clinical Medicine 9, no. 4: 975. 10.3390/jcm9040975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Nemati, P. R. , Backhaus W., Feldheim J., et al. 2022. “Brain Network Topology Early After Stroke Relates to Recovery.” Brain Communications 4, no. 2: fcac049. 10.1093/braincomms/fcac049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Ovadia‐Caro, S. , Khalil A. A., Sehm B., Villringer A., Nikulin V. V., and Nazarova M.. 2019. “Predicting the Response to Non‐Invasive Brain Stimulation in Stroke.” Frontiers in Neurology 10: 302. 10.3389/fneur.2019.00302. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Park, C. H. , and Kim M. S.. 2023. “Stratified Predictions of Upper Limb Motor Outcomes After Stroke.” Frontiers in Neurology 14: 1323529. 10.3389/fneur.2023.1323529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Paul, T. , Wiemer V. M., Hensel L., et al. 2023. “Interhemispheric Structural Connectivity Underlies Motor Recovery after Stroke.” Annals of Neurology 94, no. 4: 785–797. 10.1002/ana.26737. [DOI] [PubMed] [Google Scholar]
  34. R: A Language and Environment for Statistical Computing . 2021. “R Foundation for Statistical Computing.” https://www.R‐project.org/.
  35. Rehme, A. K. , Eickhoff S. B., Rottschy C., Fink G. R., and Grefkes C.. 2012. “Activation Likelihood Estimation Meta‐Analysis of Motor‐Related Neural Activity After Stroke.” NeuroImage 59, no. 3: 2771–2782. 10.1016/j.neuroimage.2011.10.023. [DOI] [PubMed] [Google Scholar]
  36. Rehme, A. K. , and Grefkes C.. 2013. “Cerebral Network Disorders After Stroke: Evidence From Imaging‐Based Connectivity analyses of Active and Resting Brain States in Humans.” Journal of Physiology 591, no. 1: 17–31. 10.1113/jphysiol.2012.243469. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Reibelt, A. , Quandt F., and Schulz R.. 2023. “Posterior Parietal Cortical Areas and Recovery After Motor Stroke: A Scoping Review.” Brain Communications 5, no. 5: fcad250. 10.1093/braincomms/fcad250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Rivier, C. , Preti M. G., Nicolo P., Van De Ville D., Guggisberg A. G., and Pirondini E.. 2023. “Prediction of Post‐Stroke Motor Recovery Benefits From Measures of Sub‐Acute Widespread Network Damages.” Brain Communications 5, no. 2: fcad055. 10.1093/braincomms/fcad055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rohrig, L. , Rosenzopf H., Wohrstein S., and Karnath H. O.. 2023. “The Need for Hemispheric Separation in Pairwise Structural Disconnection Studies.” Human Brain Mapping 44, no. 16: 5212–5220. 10.1002/hbm.26445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Rojas Albert, A. , Backhaus W., Graterol Perez J. A., et al. 2022. “Cortical Thickness of Contralesional Cortices Positively Relates to Future Outcome After Severe Stroke.” Cerebral Cortex 32, no. 24: 5622–5627. 10.1093/cercor/bhac040. [DOI] [PubMed] [Google Scholar]
  41. Sadeghihassanabadi, F. , Frey B. M., Backhaus W., et al. 2022. “Structural Cerebellar Reserve Positively Influences Outcome After Severe Stroke.” Brain Communications 4, no. 6: fcac203. 10.1093/braincomms/fcac203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Salvalaggio, A. , De Grazia M. D. F., Zorzi M., Thiebaut de Schotten M., and Corbetta M.. 2020. “Post‐Stroke Deficit Prediction From Lesion and Indirect Structural and Functional Disconnection.” Brain 143, no. 7: 2173–2188. 10.1093/brain/awaa156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Schaechter, J. D. , Kim M., Hightower B. G., Ragas T., and Loggia M. L.. 2023. “Disruptions in Structural and Functional Connectivity Relate to Poststroke Fatigue.” Brain Connectivity 13, no. 1: 15–27. 10.1089/brain.2022.0021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Schlemm, E. , Ingwersen T., Konigsberg A., et al. 2021. “Preserved Structural Connectivity Mediates the Clinical Effect of Thrombolysis in Patients With Anterior‐Circulation Stroke.” Nature Communications 12, no. 1: 2590. 10.1038/s41467-021-22786-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Schlemm, E. , Jensen M., Kuceyeski A., et al. 2022. “Early Effect of Thrombolysis on Structural Brain Network Organisation After Anterior‐Circulation Stroke in the Randomized WAKE‐UP Trial.” Human Brain Mapping 43, no. 16: 5053–5065. 10.1002/hbm.26073. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Schulz, R. , Gerloff C., and Hummel F. C.. 2013. “Non‐Invasive Brain Stimulation in Neurological Diseases.” Neuropharmacology 64: 579–587. 10.1016/j.neuropharm.2012.05.016. [DOI] [PubMed] [Google Scholar]
  47. Schulz, R. , Park E., Lee J., et al. 2017. “Interactions Between the Corticospinal Tract and Premotor–Motor Pathways for Residual Motor Output After Stroke.” Stroke 48, no. 10: 2805–2811. 10.1161/STROKEAHA.117.016834. [DOI] [PubMed] [Google Scholar]
  48. Sperber, C. , Griffis J., and Kasties V.. 2022. “Indirect Structural Disconnection‐Symptom Mapping.” Brain Structure & Function 227, no. 9: 3129–3144. 10.1007/s00429-022-02559-x. [DOI] [PubMed] [Google Scholar]
  49. Stinear, C. M. , Byblow W. D., Ackerley S. J., Barber P. A., and Smith M. C.. 2017. “Predicting Recovery Potential for Individual Stroke Patients Increases Rehabilitation Efficiency.” Stroke 48, no. 4: 1011–1019. 10.1161/STROKEAHA.116.015790. [DOI] [PubMed] [Google Scholar]
  50. Tedesco Triccas, L. , Burridge J. H., Hughes A. M., et al. 2016. “Multiple Sessions of Transcranial Direct Current Stimulation and Upper Extremity Rehabilitation in Stroke: A Review and Meta‐Analysis.” Clinical Neurophysiology 127, no. 1: 946–955. 10.1016/j.clinph.2015.04.067. [DOI] [PubMed] [Google Scholar]
  51. Thiebaut de Schotten, M. , Foulon C., and Nachev P.. 2020. “Brain Disconnections Link Structural Connectivity With Function and Behaviour.” Nature Communications 11, no. 1: 5094. 10.1038/s41467-020-18920-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Tscherpel, C. , Hensel L., Lemberg K., et al. 2020. “The Differential Roles of Contralesional Frontoparietal Areas in Cortical Reorganization After Stroke.” Brain Stimulation 13, no. 3: 614–624. 10.1016/j.brs.2020.01.016. [DOI] [PubMed] [Google Scholar]
  53. Venables, W. N. , Ripley B. D., and Venables W. N.. 2002. “Modern Applied Statistics With S.” In Statistics and Computing, 4th ed., 495. New York: Springer. [Google Scholar]
  54. Wang, L. , Yu C., Chen H., et al. 2010. “Dynamic Functional Reorganization of the Motor Execution Network After Stroke.” Brain 133, no. 4: 1224–1238. 10.1093/brain/awq043. [DOI] [PubMed] [Google Scholar]
  55. Wang, L. E. , Tittgemeyer M., Imperati D., et al. 2012. “Degeneration of Corpus Callosum and Recovery of Motor Function After Stroke: A Multimodal Magnetic Resonance Imaging Study.” Human Brain Mapping 33, no. 12: 2941–2956. 10.1002/hbm.21417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Ward, N. S. , Brown M. M., Thompson A. J., and Frackowiak R. S. J.. 2003. “Neural Correlates of Motor Recovery After Stroke: A Longitudinal fMRI Study.” Brain: A Journal of Neurology 126, no. 11: 2476–2496. 10.1093/brain/awg245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Yushkevich, P. A. , Piven J., Hazlett H. C., et al. 2006. “User‐Guided 3D Active Contour Segmentation of Anatomical Structures: Significantly Improved Efficiency and reliability.” NeuroImage 31, no. 3: 1116–1128. 10.1016/j.neuroimage.2006.01.015. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1.

HBM-45-e70060-s001.pdf (422.5KB, pdf)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES