Structural Plasticity of Remote Cortical Brain Regions is Determined by Connectivity to the Primary Lesion in Subcortical Stroke

Abstract

Cortical atrophy as demonstrated by measurement of cortical thickness (CT) is a hallmark of various neurodegenerative diseases. In the wake of an acute ischemic stroke, brain architecture undergoes dynamic changes that can be tracked by structural and functional magnetic resonance imaging studies as soon as 3 months after stroke. In this study, we measured changes of CT in cortical areas connected to subcortical stroke lesions in 12 patients with upper extremity paresis combining white-matter tractography and semi-automatic measurement of CT using the Freesurfer software. Three months after stroke, a significant decrease in CT of −2.6% (median, upper/lower boundary of 95% confidence interval − 4.1%/- 1.1%) was detected in areas connected to ischemic lesions, whereas CT in unconnected cortical areas remained largely unchanged. A cluster of significant cortical thinning was detected in the superior frontal gyrus of the stroke hemisphere using a surface-based general linear model correcting for multiple comparisons. There was no significant correlation of changes in CT with clinical outcome parameters. Our results show a specific impact of subcortical lesions on distant, yet connected cortical areas explainable by secondary neuro-axonal degeneration of distant areas.

Keywords

acute stroke brain imaging diffusion tensor imaging MRI neurodegeneration

INTRODUCTION

Changes of cortical thickness (CT) examined by structural magnetic resonance imaging (MRI) is a key surrogate parameter of gray-matter plasticity in normal aging and neurologic diseases. Normal age-dependent cortical thinning is highly variable over the cortical surface,^1,2 while characteristic patterns of cortical thinning are of diagnostic and prognostic values in diseases such as Alzheimer's or the frontotemporal dementias,³ or multiple sclerosis.⁴

Ischemic stroke induces disruptive changes and dynamic adaptations in the structural and functional brain architecture supposed to be linked to clinical impairment and rehabilitation. In the neuroimaging literature, reorganization within wide-spread functional brain networks have been described.^5–8 In addition, changes of white-matter properties were observed in areas both adjacent to and remote from ischemic lesions reflecting secondary degeneration.^9–11 Only few studies have focused on changes of cortical gray matter after stroke using conventional volumetric measurements,¹² voxel-based morphometry,^13,14 or CT measurements.^15,16 In a recent study, Duering et al¹⁷ showed cortical atrophy in regions with high probability of connectivity to incident subcortical infarcts in patients with CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) showing a specific impact of subcortical lesion. Taken together, these previous studies show a general pattern of cortical atrophy mainly interpreted as secondary damage to homologous areas either via direct retrograde axonal degeneration or locally disturbed cerebral blood flow and metabolism. However, it remains largely unknown how these changes are related to the primary ischemic lesion in space and time. To a lesser extent, regional increases of CT have been observed that appear to correlate with findings from functional imaging studies. These finding have been interpreted as a hint toward adaptive structural plasticity.^15,18,19

We therefore aimed to elucidate the impact of focal subcortical stroke lesions on CT in a prospective MRI study. Our focus on patients with upper limb deficit respects the clinical impact of arm and hand paresis on the most relevant daily activities and its frequent occurrence in stroke populations. In addition, recovery of upper extremity function after stroke has been extensively studied by functional and structural MRI studies highlighting structures crucial for clinical outcome such as the pyramidal tract or connected primary and secondary motor areas.^20,21 In the present study, we define structural connectivity to the ischemic lesion by reconstruction of white-matter tracts connecting stroke regions with remote cortical surfaces using probabilistic fiber tracking.²² We then analyzed dynamics of changes in CT from the first day to 3 months after stroke onset. We hypothesized that cortical thinning would specifically occur in areas structurally connected to the primary stroke lesion. In an exploratory analysis, we further looked for patterns of CT changes within a network of brain regions connected to the original stroke lesion and the contralateral homologous network of brain areas, as well as correlations of structural plasticity with recovery of motor function in the course of stroke.

MATERIALS AND METHODS

SUBJECTS

Patients with mild to severe upper-limb paresis resulting from acute ischemic stroke were considered for inclusion in our study. Patients were recruited on days 3 to 5 after stroke via the Stroke Unit of our hospital if the following inclusion criteria were met: (1) acute stroke with motor deficit to the upper limb; (2) focal subcortical ischemic stroke lesion proven my MRI; (3) informed consent; (4) the absence of severe neurologic or non-neurologic comorbidity. Patients with previous stroke, intracerebral hemorrhage, or imaging evidence of preexisting structural brain damage were excluded from the study.

In total, 36 patients fulfilled the initial inclusion criteria. Of those, 2 patients were excluded due to insufficient image quality of the structural image data and 13 patients did not participate at follow-up examinations after 3 months. In addition, nine patients had stroke lesions involving cortical areas and were excluded from further analysis. The remaining 12 patients were included in the study.

Clinical examinations were recorded in the acute (days 3 to 5) stage and 3 months after stroke onset including the National Institute of Health Stroke Scale (NIHSS), the modified Rankin scale, the Fugl-Meyer assessment (FMA) of the upper extremity, and the Action Research Arm Test (ARAT). In addition, grip force was measured and the mean value (in kg) of three consecutive measurements was calculated. Absolute and relative changes of all clinical parameters between both time points were calculated. Patients gave written informed consent according to the Declaration of Helsinki. The study was approved by the ethic committee of the medical council Hamburg, Germany, application docket: PV 37777.

Brain Imaging

Imaging data were acquired at the same time points as clinical examinations (3 to 5 days and 3 months after stroke). A 3T Siemens Skyra MRI scanner (Siemens, Erlangen, Germany) and a 32-channel head coil was used to acquire both diffusion-weighted and high-resolution T1-weighted anatomic images. For diffusion-weighted imaging, 75 axial slices were obtained covering the whole brain with gradients (b = 1,500 mm²/s) applied along 64 noncollinear directions with the sequence parameters: repetition time = 10,000 ms, echo time = 82 ms, field of view = 256 × 204, slice thickness = 2 mm, in-plane resolution = 2 × 2 mm. The complete data set consisted of 2 × 64 b1500 images and additionally one b0 image at the beginning and one after the first 64 images. For anatomic imaging, a three-dimensional magnetization-prepared, rapid acquisition gradient-echo sequence was used with the following parameters: repetition time = 2,500 ms, echo time = 2.12 ms, field of view = 240 × 192 mm, 256 axial slices, slice thickness = 0.94 mm, and in-plane resolution = 0.94 × 0.94 mm. In addition, fluid-attenuated inversion recovery (FLAIR) sequences were acquired for delineation of ischemic lesions (repetition time = 9,000 ms, echo time = 90 ms, Inversion Time (TI) = 2,500 ms, field of view = 230 × 230 mm, slice thickness = 5 mm in-plane resolution = 0.7 × 0.7 mm).

Data Preprocessing

Diffusion-weighted images were analyzed using the FSL software package 5.1 (http://www.fmrib.ox.ac.uk/fsl). All data sets were corrected for eddy currents and head motion. Fractional anisotropy maps were calculated fitting the diffusion tensor model at each voxel. Structural T1-weighted anatomic images were processed using the Freesurfer software package 5.3.0 with standard procedures and parameters.^23–25 The semiautomatic Freesurfer (Charlestown, MA, USA) longitudinal stream has shown to create measurements of CT with high validity and reliability compared with postmortem, and test-retest analysis.^26,27 It detected longitudinal sub-millimeter changes in CT in stroke patients during a time frame of 3 months.¹⁶ All images were visually inspected and data of insufficient quality (i.e., due to motion artefacts) excluded. Special attention was directed to accuracy of brain extraction: cortical segmentations were checked visually slice by slice and manual corrections applied following recommendations from the online Freesurfer documentation (found under the section ‘fixing a bad skull strip’). Registration of structural and diffusion images was achieved using linear and nonlinear transformation tools implemented in FSL.²²

Definition of Regions of Interest, Lesion Segmentation, and Probabilistic Fibertracking

We defined original stroke lesions as region of interest (stroke regions of interest (ROIs)) and used probabilistic fiber tracking to determine ipsilesional cortical brain areas structurally connected to the stroke ROI. To study structural changes in homologous brain areas in the contralesional hemisphere, a ROI mirroring the stroke ROI was defined and connectivity of contralesional cortical brain regions to this mirrored ROI was determined. Figure 1 displays an overview of the methods used to reconstruct fiber tracts connecting subcortical lesions to cortical gray-matter areas. Stroke lesions were delineated on FLAIR images at the first examination (3 to 5 days after stroke) using the in-house software ANTONIA as described previously.²⁸ In brief, a ROI was manually drawn surrounding the FLAIR lesion including a generous margin at each effected slice. In a subsequent step, a signal intensity threshold was manually applied to refine the final lesion segmentation. For mirroring the primary lesion, the interhemispheric fissure was manually defined at two slices that were more than four slices distant from each other. Lesions were then mirrored along these axis to the healthy hemisphere (termed ‘mirrored ROI’ in the following manuscript). All lesions were delineated by experienced rater (BC) as reported previously^29,30 and masked with automated brain segmentation of the cerebrospinal fluid. Stroke and mirrored ROIs were registered to the diffusion imaging data. Connectivity of stroke ROI and mirror ROI to cortical brain areas in the corresponding hemispheres was assessed using probabilistic diffusion models and tractography (bedpost and probtrackx).^31,32 From each mask voxel, 10,000 samples were initiated through the probability fiber distribution of principle white-matter fiber directions with a curvature threshold of 0.2. Resulting tract distributions were normalized in relation to the general connectivity profile in each individual patient. We applied a threshold of 1,000 samples (1% of 10,000 samples) following recommendations from the online documentation of the FSL library. In addition, we applied two additional thresholds to investigate the reliability of our results. In general, higher thresholds yield more specific results, while being less sensitive to subordinate pathways, whereas lowering the threshold can lead to increased false-positive connections and susceptibility to noise. To determine the additional thresholds, we aimed at increasing and decreasing the initially created surface by approximately 25%. Finally, cortical ROIs were created at the intersection of white matter and the cortical surface tract using Freesurfer (see Figure 1) and the cortex of each hemisphere was binarized into connected and unconnected brain regions. Probabilistic fiber tracking was restricted to intrahemispheric connections for methodological reasons, as probabilistic tracking to contralesional cortical areas via the corpus callosum is still technically challenging. Furthermore, combination of tracking results from intra- and interhemispherical pathways would have resulted in ambiguous assignment of cortical brain areas possibly connected to both stroke ROI and mirror ROI. For this reason, we decided to focus on intrahemispherical connected brain areas tracts supplemented by the analysis of a homologous brain network in the contralesional hemisphere.

Figure 1.

Reconstruction of brain surface regions of interest for measurement of cortical thickness. (A) Acute fluid-attenuated inversion recovery (FLAIR) image showing an subcortical infarct (arrow); (B) stroke lesion of interest (red) and mirrored region of interest (green) after segmentation and registration to diffusion tensor images (map of fractional anisotropy, FA, shown); (C) tract reconstruction using stroke (red) and mirorred (green) region as starting points and a threshold of 1% overall connectivity to the entire hemisphere; (D) intersections of white-matter tracts and the cortical surface. Red dot marks the affected hemisphere. Connected areas are shown in yellow.

Cortical Thickness Measurements

Cortical thickness was measured and compared between brain regions connected to the stroke ROI (‘connected ROI’) and unconnected cortical areas (‘unconnected ROI’) in the stroke hemisphere. Analogous measurements were performed in the healthy hemisphere in areas connected and unconnected to the mirrored ROI. Absolute values of cortical thickness were determined by the distance between the white matter and pial surface²⁴ and relative percentage of change (PC) in time calculated for each ROI: PC = (CT_follow-up - CT_onset)/CT_onset.

Statistical Analysis

Percent change of CT was tested for significance against zero using a Wilcoxon signed-rank test for each reconstructed cortical surface. In addition, we tested for significant changes in surface areas connected to the stroke and mirrored ROI using a vertex-wise general linear model implemented in Freesurfer (mri_glmfit).³³ Individual maps of CT change were registered bilaterally to the standard template and smoothed with a Gaussian Kernel of 15 mm FWHM. A one sample group mean was tested against zero (no change in CT) and a false discovery rate of P < 0.05 applied to correct for multiple comparisons.

Relative changes of clinical parameters were calculated: Relative change NIHSS=(NIHSS_follow-up - NIHSS_onset)/NIHSS_onset. Changes were calculated for the FMA, ARAT, and grip strengths accordingly. Percentage change of CT and relative change of clinical parameters were correlated using a two-tailed Spearman's rank correlation coefficient. Stroke lesion volumes were correlated with relative thickness changes in regions connected to stroke and mirror ROI using the Spearman's Rank correlation coefficient. All statistical analysis was performed in SPSS 20.0 (IBM Co., Somers, NY, USA). For presentation of group data and group comparison, images of patients with left-hemispheric stroke were flipped across the x axis so that all stroke lesions appeared on the right hemisphere.

RESULTS

A total of 12 patients were included in the study. There were no significant differences of age or clinical impairment between patients excluded and included in our study (Supplementary Table 1). Median stroke lesion volume was 3.7 mL (3.7 to 10.2 mL). Figure 2 gives an overview of all individual stroke lesions in standard space. Clinical data for each time point are shown in Table 1. Median age at onset was 70 (56 to 81 years), median NIHSS was 5 (1 to 19). Five female patients were included (42%). Stroke lesions were located on the left side in five patients (42%). Mean duration at follow-up was 103 days (90 to 157 days) and overall, clinical improvement of motor function was seen in almost all patients. None of the included patients received systemic thrombolysis. At follow-up, median NIHSS was 1 (0 to 10).

Table 1

Clinical and demographic data from all patients at each time point

ID	Age	Gen	Side	Location	Vol (mL)	Days	ARAT	FMA	MRS	NIHSS	Grip (kg)
3	56	M	R	CI, CR	3.7	3	0	4	4	13	0
						99	0	7	2	5	0
5	69	F	R	CI	6.3	4	0	0	4	19	0
						157	0	5	5	10	0
7	62	M	L	CI, CR	2.5	4	38	56	3	3	7.93
						95	57	62	1	0	31.2
14	71	M	R	CR	1.7	5	57	62	1	1	15.9
						90	57	65	0	0	28
20	73	F	L	CI CR	6.3	4	0	4	4	9	0
						116	0	13	4	3	0
22	70	M	R	CR	3.7	4	57	65	2	2	27.6
						100	57	66	1	0	39
23	77	F	R	CI, CR	10.2	3	0	4	5	8	0
						103	4	2	3	2	3.67
24	53	M	R	CI, BG	3.4	4	57	66	1	4	28
						92	57	66	1	0	Missing
25	81	F	R	CR	1.3	5	57	65	1	0	12
						90	57	66	0	0	16
26	78	F	L	CR	1.6	3	57	65	1	5	15.3
						99	57	64	1	0	21
33	63	M	L	CR	1.5	4	36	42	2	3	10.7
						101	57	63	1	1	30
35	65	M	L	CI, CR	8.4	6	0	6	4	8	0
						99	0	15	3	4	1

Abbreviations: ARAT, Action Research Arm Test; BG, basal ganglia; CI, Capuslainterna; CR, corona radiata; FMA, Fugl-Meyer Assessment; MRS, Modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale. Lesion location and volume are described.

Figure 2.

Overview of lesion locations in all patients demonstrated on representative slices. Lesions are registered to normal space and displayed on the MNI brain template in coronar and axial orientations. Images are shown according to radiologic convention. Numbers represent patient IDs.

White-matter tracts connecting stroke and mirrored ROIs to the cerebral cortex were successfully reconstructed in all patients. Values resulting from different thresholds of probabilistic tracking are shown in Figure 4 and the online Supplementary Material. In the following, only findings from the initial threshold (1% overall connectivity, 1,000 of 10,000 samples) are reported.

Figure 4.

Mean percentage change of cortical thickness 3 months after stroke in areas connected to stroke lesion and mirrored region of interest (ROI). Values for three different thresholds of probabilistic tracking are shown. Error bars represent 95% confidence intervals. Horizontal line at zero signifies level of unchanged thickness. ^*P < 0.05 resulting from Wilcoxon signed-rank test against zero (i.e., no change in cortical thickness).

As shown in Figure 3A, subcortical lesions were connected to the precentral, postcentral, and dorsal parts of the superior frontal gyrus in all patients. In addition, connected cortical brain areas included the insular cortex, ventral parts of the superior temporal and angular gyrus. A similar pattern was found for connectivity to the mirrored ROI in the healthy hemisphere. Connected surface areas were comparable between the stroke (61.2 cm²; 95% CI: 53.7 cm² to 70.8 cm²) and healthy hemisphere (63.8 cm²; 95% CI: 54.5 cm² to 73.0 cm²; see also Supplementary Table 2).

Table 2

Relative values of percent change of cortical thickness after 3 months for cortical areas connected and unconnected lesion and mirrored region of interest

	Connected cortex			Unconnected cortex
	Mean	95% CI (upper/lower)	P-value	Mean	95% CI (upper/lower)	P-value
Healthy hemisphere	−0.9%	−2.9/0.9	0.27	0.1%	−1.3/1.7	0.94
Stroke hemisphere	−2.6%	−4.1/−1.1	<0.01 a	−0.5%	−1.9/1.0	0.43

Abbreviation: CI, confidence interval.

P<0.01 resulting from Wilcoxon signed-rank test against zero.

Figure 3.

Illustration of cortical areas connected to stroke and mirror region of interest (A) and change of cortical thickness after 3 months (B). Values of relative changes of cortical thickness are shown in range between 0.3% and 3%. Color bars represent number of patients (A) and percent change of cortical thickness (B). All data from stroke lesions are mapped onto the right hemisphere. (C) Composite pial representation of the statistically significant clusters of cortical thickness (87 mm², MNI coordinates: x = 13, y = - 1.5, z = 65.1), that showed a significant decrease in cortical thickness at the stroke hemisphere (P-values corrected for multiple comparisons by false discovery rate, P < 0.05).

Results from cortical measurements showed a significant decrease of cortical thickness in areas connected with the primary lesion side of −2.6% over the observation period (95% CI: - 4.1%/ - 1.1%, P = 0.006, Wilcoxon signed-rank test against zero). Using alternative thresholds to identify connected regions did not significantly alter values of mean cortical thinning. There was, however, a trend toward more pronounced cortical thinning for more conservative thresholds identifying regions with higher likelihood of being connected to the stroke region (Figure 4, see also online Supplementary Table 3). Absolute values of mean CT in all areas are shown in Table 3. Lesser and nonsignificant changes of CT were observed in white-matter areas connected to the mirrored ROI in the healthy hemisphere (-0.9%; 95% CI: −2.9%/ 0.9%, P = 0.272). No significant changes were detected in areas unconnected to the lesion and mirrored ROI (Tables 2 and 3). In the surface-based statistical analysis of CT change, an area with significant thinning (tested against zero, False Discovery Rate (FDR) < 0.05) was found at the superior frontal gyrus, at the lateral border of the supplemental motor area (SMA), in the stroke hemisphere (P < 0.0005; surface area: 87 mm², MNI coordinates of peak voxel: x = 13; y = −1.5; z = 65.1). No significant changes were found on the healthy hemisphere. Figure 3B illustrates uncorrected values of thickness changes in areas connected to the primary lesion and mirrored ROI at the stroke and healthy hemisphere. The area of significant change is shown in Figure 3C. In the exploratory visual analysis, we observed a tendency of increased CT at the precentral and postcentral gyrus in the healthy hemisphere (Figure 3B). No significant correlations between lesion volume and percent change of cortical thickness in regions connected to stroke or mirror ROI were found (stroke ROI: r = 0.15, P = 0.65; mirror ROI: r = - 0.24, P = 0.46).

Table 3

Absolute values of cortical thickness at each time point after stroke for cortical areas connected and unconnected to primary lesion or mirrored region of interest

	Time point	Mean (mm)	Median (mm)	s.d.
Healthy hemisphere
Connected cortex	Onset	1.97	2.01	0.19
	FU	1.95	1.99	0.19
Unconnected cortex	Onset	2.20	2.23	0.13
	FU	2.20	2.20	0.13

Stroke hemisphere
Connected cortex	Onset	1.93	1.97	0.22
	FU	1.88	1.90	0.22
Unconnected cortex	Onset	2.19	2.22	0.14
	FU	2.18	2.18	0.14

Abbreviations: FU, follow-up; s.d., standard deviation.

We did not find any significant correlations between values of clinical performance and CT changes. In detail, relative changes of motor performance (NIHSS, ARAT, FMA, and grip strength) were not significantly correlated with changes in CT in cortical areas connected to stroke and mirrored ROIs. Detailed results of correlation statistics are shown in Supplementary Table 4.

DISCUSSION

In this study, a significant decrease in CT was exclusively found in brain areas connected to subcortical focal stroke lesions, while unconnected areas did not display remote cortical changes. In the healthy hemisphere, less pronounced cortical thinning was found in cortical regions connected to brain regions homologous to the stroke lesions. In an exploratory analysis, a more differentiated pattern of CT change emerged within connected ipsilesional and homologous brain regions demonstrating focal increases of CT predominantly in the precentral gyrus of the healthy hemisphere. These changes, however, did not reach statistical significance and have therefore to be interpreted with caution.

Longitudinal changes of CT in stroke patients have been examined by few cross-sectional and longitudinal studies. There is only one single study reporting CT changes as early as 3 months after stroke, and in this study an overall decrease in CT of 0.5% was observed in a heterogeneous sample of patients (n = 16) with both cortical and subcortical strokes.¹⁶ On a much longer time scale, Duering et al¹⁷ examined cortical thinning in areas connected to incidental subcortical infarcts in nine patients with CADASIL.¹⁷ After a mean follow-up of 34 months, thickness of cortical areas connected to stroke lesions decreased about 9%, whereas no significant changes were detected in the corresponding contralesional gray-matter areas. Cortical atrophy post-stroke has also been a recurrent theme in volumetric studies either using voxel-based morphometry or more conventional methods.³⁴ Decrease of cortical matter volumes were observed exceeding the primary infarct lesion,¹² especially involving the ipsilesional hemisphere.¹⁴ In a similar time span poststroke, Dang et al¹³ reported reduced volumes of ipsilesional cortical areas 3 months after subcortical stroke, mainly involving the ipsilesional precentralgyrus and SMA at the superior frontal gyrus.¹³ Interestingly, significant cortical thinning has also been found at this region at the lateral border of the SMA in our study. Finally, evidence from animal studies showed similar trends of morphologic changes after experimental stroke involving reduced CT, volume and neuronal density in areas remote from primary lesions.³⁵ Cortical thinning was also the prominent finding in our study. In addition, we provide novel data by demonstrating atrophy in specific cortical areas that are in distance, yet connected to subcortical lesions. In contrast, CT of unconnected areas remained relatively unchanged over 3 months. We thus conclude in line with the previous studies that cortical thinning 3 months after subcortical stroke does not represent a general mechanism of damage affecting the entire brain like accelerated aging, but rather is a specific consequence of secondary degeneration in structurally connected brain areas. ‘Dying back’ of neurons proximally to injury has been shown by imaging studies³⁶ and potentially induces cortical thinning in patients with diffuse white-matter hyperintensities.³⁷ In line with these findings, our results suggest a specific impact of the subcortical lesion on cortical atrophy that is most likely explained by retrograde degeneration after axonal injury.

Interestingly, we also observed cortical thinning in areas connected to the contralesional area mirroring the stroke lesion in a pattern largely symmetrical to the injured hemisphere. Although decreased gray-matter density contralesional to stroke lesions has been described,¹⁹ the underlying mechanisms are still unclear. However, it might be speculated that focal lesions exert an influence on contralesional motor areas strongly connected between both hemispheres via trans- or intersynaptic pathways.³⁸ Additionally, in an exploratory analysis, a regional increase of CT in spite of generally atrophic processes in connected cortical areas was observed in the motor hand area (‘hand knob’)³⁹ of the ipsi-and contralesional precentral cortex. These findings have to remain preliminary, since no significance (testing against zero) was reached in the vertex-based analysis. It is however tempting to speculate that dynamic reorganization of functional networks after stroke⁴⁰ is linked to structural plasticity.⁴¹ This would be in line with previous findings from other groups showing a increase of gray-matter volumes in a similar time frame.^13,15

The pattern of brain areas connected to the primary lesion and within 3 months affected by changes in CT mainly revealed a network of brain areas involved in primary and secondary motor function control such as the pre- and postcentral sensorimotor area and the SMA located at the superior frontal gyrus.⁷ This, of course, results from our inclusion criteria with enrolment of subcortical strokes with upper extremity motor symptoms. The underlying subcortical stroke lesions consistently involved the pyramidal tract to some extent either in the corona radiata or at the level of the internal capsule. The localization of regional CT increase in the contralesional primary motor cortex in a sample of patients with upper limb paresis as characteristic clinical feature also points toward a potential functional relevance of this observed increase in CT. Further studies are needed to elucidate the interplay between structural and functional plasticity of stroke patients and motor deficits in a larger population of stroke patients.

In the present study, we did not detect significant correlations between the degree of motor recovery and relative change in cortical thickness. Thus, we are unable to link findings of structural plasticity to clinical parameters. The lack of correlation with clinical scores may result from the fact that we examined structural changes over a relatively short period of time (3 months). Given the dynamics of functional reorganization and recovery within the first months after stroke, 3 months after stroke might be too early for capturing the final association of structural changes and functional rehabilitation. In previous studies, association between changes in CT and clinical outcome has so far been reported in a much longer time frame of at least 1 year after stroke.^14,19 To the best of our knowledge, only one published study has so far reported significant correlations of motor performance with cortical volume using a different method Voxel-Based Morphometry (VBM) in a comparable period of 3 months¹³ with reduction in the ipsilesional SMA volume showing a moderate correlation with worse clinical outcome measured by the FMA (r = 0.52, P = 0.048). Of note, we similarly detected highest rates of atrophy in this region using a surface-based method to measure CT. The relatively small sample size represents an additional limitation that contributes to missing correlations between structure and function. From a technical point of view, probabilistic tractography offers the advantage of modeling multiple fiber orientations to detect a range of subordinate pathways missed by deterministic fiber tracking.³¹ However, normalization of tractography results to remove false-positive results is not standardized and remains arbitrary to some extent. In this work, we chose an initial threshold of 1% (of overall connectivity) to eliminate erroneous results as recommended by the online documentation and applied in a comparable study of CT measurements in patients with CADASIL.¹⁷ This threshold yielded plausible results as to the anatomic locations of connected brain areas (see Figure 3A). Significant cortical thinning with comparable extent was also demonstrated applying alternative thresholds that decreased and increased the connected area by about 25%, respectively (Figure 4). In addition, we observed more pronounced effects of cortical thinning with higher probability thresholds, which is most likely due to the higher probability of these regions being connected to the primary lesion (Supplementary Table 5).

In conclusion, we have showed the direct impact of focal subcortical ischemic lesions on cortical atrophy in a well-defined group of patient combining white-matter tractography with surface-based cortical thickness analysis. Our results are in line with axonal retrograde or anterograde (transsynaptic) degeneration as the underlying mechanism of cortical thinning in distant and primarily unaffected regions and point toward adaptive structural plasticity in homologous area of the injured brain network during recovery from stroke. Further work is needed to elucidate the impact of structural changes on clinical performance in a larger population of stroke patients.

AUTHOR CONTRIBUTIONS

BC has contributed substantially to the conception of this study, acquisition, analysis, and interpretation of data as well as drafting the article. MB, JS, and RS have contributed substantially to the acquisition of data for this study and GT, CG, FH, and JF have contributed substantially to the conception and design of this study, interpretation of data as well as revising the article critically for important intellectual content. All authors have approved the final version of this manuscript for publication.

DISCLOSURE/CONFLICT OF INTEREST

GT has received funding from the German Research Foundation (DFG), SFB 936 ‘Multi-site Communication in the Brain’ (Project C2) and has received fees as consultant or lecturer from Bayer, Bristol-MyersSquibb/Pfizer, BoehringerIngelheim, and Covidien. CG has received funding from the German Research Foundation (DFG), SFB 936 ‘Multi-site Communication in the Brain’ (Project C1). FH has received funding from the German Research Foundation (DFG), SFB 936 ‘Multi-site Communication in the Brain’ (Project C4). JF has received funding from the German Research Foundation (DFG), SFB 936 ‘Multi-site Communication in the Brain’ (Project C2) and has received fees as consultant or lecturer from Bayer, BoehringerIngelheim, Codman, Covidien, MicroVention, Penumbra, Philips, and Siemens.

Footnotes

Notes

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website ()

References

Salat

Buckner

Snyder

Greve

Desikan

RSR

Busa

. Thinning of the cerebral cortex in aging. Cereb Cortex 2004; 14: 721–730.

Tamnes

Ostby

Fjell

Westlye

Due-Tønnessen

Walhovd

. Brain maturation in adolescence and young adulthood: regional age-related changes in cortical thickness and white matter volume and microstructure. Cereb Cortex 2010; 20: 534–548.

Hartikainen

Räsänen

Julkunen

Niskanen

Hallikainen

Kivipelto

. Cortical thickness in frontotemporal dementia, mild cognitive impairment, and Alzheimer's disease. J Alzheimers Dis 2012; 29: 1–18.

Calabrese

Rinaldi

Mattisi

Grossi

Favaretto

Atzori

. Widespread cortical thinning characterizes patients with MS with mild cognitive impairment. Neurology 2010; 74: 321–328.

Ward

Newton

Swayne

OBC

Lee

Thompson

Greenwood

. Motor system activation after subcortical stroke depends on corticospinal system integrity. Brain 2006; 129: 809–819.

Carter

Shulman

Corbetta

. Why use a connectivity-based approach to study stroke and recovery of function? Neuroimage 2012; 62: 2271–2280.

Rehme

Eickhoff

Rottschy

Fink

Grefkes

. Activation likelihood estimation meta-analysis of motor-related neural activity after stroke. Neuroimage 2012; 59: 2771–2782.

Wang

Chen

Qin

Fan

. Dynamic functional reorganization of the motor execution network after stroke. Brain 2010; 133: 1224–1238.

Crofts

Higham

Bosnell

Jbabdi

Matthews

Behrens

TEJ

. NeuroImage Network analysis detects changes in the contralesional hemisphere following stroke. Neuroimage 2011; 54: 161–169.

10.

Auriel

Edlow

Reijmer

Fotiadis

Ramirez-Martinez

. Microinfarct disruption of white matter structure: a longitudinal diffusion tensor analysis. Neurology 2014; 83: 182–188.

11.

Thomalla

Glauche

Weiller

Röther

. Time course of wallerian degeneration after ischaemic stroke revealed by diffusion tensor imaging. J Neurol Neurosurg Psychiatry 2005; 76: 266–268.

12.

Kraemer

Schormann

Hagemann

Witte

Seitz

. Delayed shrinkage of the brain after ischemic stroke: preliminary observations with voxel-guided morphometry. J Neuroimaging 2004; 14: 265–272.

13.

Dang

Liu

Xing

Xie

Peng

. Longitudinal cortical volume changes correlate with motor recovery in patients after acute local subcortical infarction. Stroke 2013; C: 1–7.

14.

Fan

Zhu

Chen

Qin

Wang

. Dynamic brain structural changes after left hemisphere subcortical stroke. Hum Brain Mapp 2013; 34: 1872–1881.

15.

Brodtmann

Pardoe

Lichter

Ostergaard

Cumming

. Changes in regional brain volume three months after stroke. J Neurol Sci 2012; 322: 122–128.

16.

Pardoe

Lichter

Werden

Raffelt

Cumming

. Cortical thickness estimation in longitudinal stroke studies: A. Neuroimage (Amst) 2014, doi:10.1016/j.nicl.2014.08.017.

17.

Duering

Righart

Csanadi

Jouvent

Hervé

Chabriat

. Incident subcortical infarcts induce focal thinning in connected cortical regions. Neurology 2012; 79: 2025–2028.

18.

Sterr

Dean

PJa

Vieira

Conforto

Shen

Sato

. Cortical thickness changes in the non-lesioned hemisphere associated with non-paretic arm immobilization in modified CI therapy. NeuroImage Clin 2013; 2: 797–803.

19.

Gauthier

Taub

Mark

Barghi

Uswatte

. Atrophy of spared gray matter tissue predicts poorer motor recovery and rehabilitation response in chronic stroke. Stroke 2011; 43: 453–457.

20.

Schulz

Park

Boudrias

Gerloff

Hummel

Ward

. Assessing the integrity of corticospinal pathways from primary and secondary cortical motor areas after stroke. Stroke 2012; 43: 2248–2251.

21.

Schulz

Zimerman

Timmermann

Wessel

Gerloff

Hummel

. White matter integrity of motor connections related to training gains in healthy aging. Neurobiol Aging 2014; 35: 1404–1411.

22.

Jenkinson

Beckmann

Behrens

TEJ

Woolrich

Smith

. Fsl. Neuroimage 2011; 62: 782–790.

23.

Dale

Fischl

Sereno

. Cortical surface-based analysis. I. Segmentation and surface reconstruction. Neuroimage 1999; 9: 179–194.

24.

Fischl

Dale

. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci USA 2000; 97: 11050–11055.

25.

Fischl

. FreeSurfer. Neuroimage 2012; 62: 774–781.

26.

Han

Jovicich

Salat

van der Kouwe

Quinn

Czanner

. Reliability of MRI-derived measurements of human cerebral cortical thickness: the effects of field strength, scanner upgrade and manufacturer. Neuroimage 2006; 32: 180–194.

27.

Reuter

Rosas

Fischl

. Highly accurate inverse consistent registration: a robust approach. Neuroimage 2010; 53: 1181–1196.

28.

Forkert

Cheng

Kemmling

Thomalla

Fiehler

. ANTONIA perfusion and stroke. A software tool for the multi-purpose analysis of MR perfusion-weighted datasets and quantitative ischemic stroke assessment. Methods Inf Med 2014; 53: 1–13.

29.

Cheng

Brinkmann

Forkert

Treszl

Ebinger

Köhrmann

. Quantitative measurements of relative fluid-attenuated inversion recovery (FLAIR) signal intensities in acute stroke for the prediction of time from symptom onset. J Cereb Blood Flow Metab 2012; 223153: 1–9.

30.

Cheng

Forkert

Zavaglia

Hilgetag

Golsari

Siemonsen

. Influence of stroke infarct location on functional outcome measured by the modified rankin scale. Stroke 2014; 45: 1695–1702.

31.

Behrens

TEJ

Berg

Jbabdi

Rushworth

MFS

Woolrich

. Probabilistic diffusion tractography with multiple fibre orientations: What can we gain? Neuroimage 2007; 34: 144–155.

32.

Smith

Jenkinson

Woolrich

Beckmann

Behrens

TEJ

Johansen-Berg

. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 2004; 23: S208–S219.

33.

Fischl

Sereno

Tootell

RBH

Dale

. High-resolution inter-subject averaging and a surface-based coordinate system. Hum Brain Mapp 1999; 8: 272–284.

34.

Jouvent

Viswanathan

Chabriat

. Cerebral atrophy in cerebrovascular disorders. J Neuroimaging 2010; 20: 213–218.

35.

Karl

Alaverdashvili

Cross

Whishaw

. Thinning, movement, and volume loss of residual cortical tissue occurs after stroke in the adult rat as identified by histological and magnetic resonance imaging analysis. Neuroscience 2010; 170: 123–137.

36.

Liang

Zeng

Liu

Ling

. A prospective study of secondary degeneration following subcortical infarction using diffusion tensor imaging. J Neurol Neurosurg Psychiatry 2007; 78: 581–586.

37.

Seo

Lee

J-M

Park

J-S

Kim

S-H

Kim

. Cortical thinning related to periventricular and deep white matter hyperintensities. Neurobiol Aging 2012; 33: 1156–1167.

38.

Wahl

Lauterbach-soon

Hattingen

Jung

Singer

Volz

. Human motor corpus callosum : topography, somatotopy, and link between microstructure and function. Differences 2007; 27: 12132–12138.

39.

Yousry

Schmid

Alkadhi

Schmidt

Peraud

Buettner

. Localization of the motor hand area to a knob on the precentral gyrus. A new landmark. Brain 1997; 120: 141–157.

40.

Rehme

Fink

von Cramon

Grefkes

. The role of the contralesional motor cortex for motor recovery in the early days after stroke assessed with longitudinal FMRI. Cereb Cortex 2011; 21: 756–768.

41.

Schaechter

Moore

Connell

Rosen

Dijkhuizen

. Structural and functional plasticity in the somatosensory cortex of chronic stroke patients. Brain 2006; 129: 2722–2733.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB