Association of Extrapyramidal Tracts’ Integrity With Performance in Fine Motor Skills After Stroke

Twenty patients (aged 65.0±2.4 years; range 44–81; 11 females; 4 left-handed) with first-ever ischemic strokes (12 in the dominant hemisphere) and impaired motor function were recruited after approval of the local ethics committee (Ethik-Kommission der Ärztekammer Hamburg) and written informed consent of the patients were obtained. Pontine, mesencephalic, or large hemispheric ischemic lesions of over two-thirds of the medial cerebral artery’s territory had been excluded previously. The patients underwent clinical examination in the chronic stage of recovery (370±6.9 days after the stroke incident). We chose this time point of clinical examination because of the reported delayed development of dystonia, tremor, and parkinsonism after ischemic lesions in the basal ganglia within 1 to 12 months.⁶ The examination comprised assessment of the modified Rankin Scale score, National Institutes of Health Stroke Scale, grip force for both hands determined by dynamometer (GF_AH/UH in kg) and the nine-hole peg test (NHP) performance as a measure for fine motor skills for both hands (NHP_AH/UH in pegs per second). We chose the NHP to assess fine motor skills for its focus on fine manual dexterity, which fitted to the deficits of the cohort of well-recovered patients more precisely than the Action Research Arm Test or the Box and Block Test.¹³ As a baseline of the NHP we used the unaffected hand, since Schulz et al¹⁴ reported of no significant difference in the performance in both hands of healthy subjects.

At the subacute stage of recovery (108±6.5 days) diffusion-weighted images were acquired using a 3-T Siemens Skyra magnetic resonance imaging scanner (Siemens, Erlangen, Germany). For calculation of the probable trajectories of the nigro-pallidal system, we used diffusion-weighted data from 26 healthy participants of comparable age and sex to our patients that had been recruited for a previous study.¹⁴ FSL software package 5.1 (http://www.fmrib.ox.ac.uk/fsl) was used for preprocessing accordingly to protocols of recent studies.¹⁵ Masks of the SN, GP internus, and GP externus were derived from a probabilistic atlas of subcortical anatomy.¹⁶ To account for the size of the mask in relation to the resolution of our data, we combined both pallidal masks to 1 seed and target mask, respectively. Acute fluid-attenuated inversion recovery images were used to characterize stroke lesions.

As described recently,¹⁷ probabilistic tractography was conducted to reconstruct the connections between the SN and the GP in the healthy subjects sending reciprocally 5000 streamlines per seed voxel to the according target mask. After applying an individual threshold of 1% of successful streamlines the voxels, which were found in at least 65% of the subjects were considered representing the probable course of the nigro-pallidal fibers. Subsequently, the overlap of the reconstructed tracts with the seed and target masks was excluded. Finally, these binarized templates were used in the stroke patients to calculate individual tract-specific levels of fractional anisotropy (FA_SN-GP) as a marker for microstructural properties. For the CST we used templates from the same healthy controls calculated for a previous study¹⁴ in the same manner (FA_CST). For both we calculated ratios of affected and unaffected hemisphere.

Data were analyzed using RStudio 1.0.143 (http://www.rstudio.com/) with R statistical package 3.4.1 (http://www.r-project.org/). Statistical significance was considered at P values ≤0.05 and all descriptive results are presented as mean±SEM. Paired sample t test was used for pair-wise comparisons of absolute FA values and comparison of the affected and unaffected hand within the patients. For the final statistical modeling, we used proportional FA_AH/UH values for taking in account the intersubject variability in general levels of FA. Linear regression models were fitted for NHP_AH. Initially, besides the FA_SN-GP, we included age, sex, whether or not the dominant hemisphere was affected, FA_CST and the lesion volume in the model. For final model selection, we removed stepwise insignificant variables. The R-package effects¹⁸ were used for statistical figures. For tract and lesion illustrations we used MRIcroGL v1.0.20170714 (https://github.com/neurolabusc/MRIcroGL/).

Our group of patients was well recovered (median modified Rankin Scale score of 1.0 and median National Institutes of Health Stroke Scale of 0). Of the 20 patients recruited, only 1 patient was not able to perform the NHP_AH. The mean values for the NHP on the affected hand (NHP_AH) was 0.36±0.03 pegs/s, significantly slower than the mean NHP_UH of 0.42±0.02 pegs/s (P=0.007). The mean GF_AH was 34.23±4.12 kg and was not significantly different from the GF_UH with 37.84±2.96 kg (P=0.23). The ischemic lesions were of heterogeneous localization including the striatum, cortex, capsula interna, and externa. Please see Figure 1 for the distribution of stroke lesions and for detailed patient characteristics Table I in the online-only Data Supplement.

The nigro-pallidal tracts could be reconstructed successfully in all healthy subjects. Leaving the SN, the fibers ascend in the ventral mesencephalon spatially separated and medio-ventral of the CST. Passing as 1 bundle through the area of the nucleus subthalamicus lateral of the mammillary bodies and medial of the tractus opticus, they fan out to a cranial and a caudal branch toward GP internus and GP externus (Figure 1; Figure I in the online-only Data Supplement).

The mean volume of stroke lesions was 18.8±6.4 mL (range 0.6 to 92.1 mL). In 5 subjects we observed a marginal overlap (mean 0.10±0.02 mL) with the mask of the nigro-pallidal tracts at the termination toward the striatum. Overlapping voxels were excluded from further analysis to avoid any direct influence of the ischemic lesions on the calculated FA. For all 20 patients included in this study, FA_SN-GP (ratio AH/UH, 0.96±0.02) and FA_CST (ratio AH/UH, 0.90±0.02) were significantly reduced in the affected hemisphere (P=0.021 and P<0.001, respectively; paired t test; Table). The lesion volume had no influence on the degree of tract integrity (FA_SN-GP, P= 0.77; Pearson’s correlation).

In the final model, FA_SN-GP (P=0.040), FA_CST (P<0.001), and age (P<0.001) remained as significant predictors of performance in the NHP 1 year after stroke. Please see Figure 2 and Table II in the online-only Data Supplement for details. In advance we excluded stepwise the factors sex (P=0.085) and dominant hand affected (P=0.224). A cross-correlation of the analyzed factors did not exist and there were no high-leverage points (Cook’s distance > 0.5).

We successfully reconstructed the ascending and descending tracts between the SN and the GP externus and GP internus robustly. The anatomic course suggests that they include the tracts connecting the SN with the striatum and the ones connecting the nucleus subthalamicus with the GP, and it matches the previously reconstructed tracts by postmortem diffusion-weighted imaging.¹² Except the feedback tracts to the cortex through the thalamus, and the interstriatal connections between putamen, nucleus caudatus, and GP, they represent the main part of the extrapyramidal system of the basal ganglia. That these tracts do not overlap with the CST supports this classification.

We determined microstructural properties of the nigro-pallidal tracts by calculating the FA 3 months after stroke onset in expectance of a full amount of neuronal degeneration at that point of time according to literature.¹⁹ In patients with ischemic stroke the nigro-pallidal tracts’ FA was reduced on the ipsilateral side significantly compared with the healthy hemisphere without the tracts being directly involved by ischemic lesions. We therefore interpret their decreased integrity by secondary (Wallerian) degeneration consecutive to ischemic death of basal ganglia cells. The fact that the tracts’ integrity was not influenced by lesion size supports the assumption of a region-specific degeneration instead of a mere mass effect related to the amount of the overall ischemic issue damage distal to the tracts.

Integrity of nigro-pallidal tracts was a significant predictor of performance in the NHP measuring fine motor skills at 1 year after stroke in addition to and beyond the predictive value of age and CST integrity. This can be explained by the functional role of the pathways between the striatum and the SN, and between the striatum, GP, and nucleus subthalamicus in modulating complex movements,^20–22 discovered in animal models. That their loss of integrity is associated with diminished fine motor skills in patients with fully recovered strength supports their described role in selection, initiation, performance, and evaluation of action sequences and might hint to their importance for long-term recovery.

The positive association of CST integrity with motor performance reflects its executive role in the motor network for performing coordinated action sequences necessary for the NHP next to its role in developing strength and accords with current data.^23,24 Younger age is associated with better recovery from stroke symptoms for a wide range of functions^25,26 and may results from the larger brain capacity for reorganization in younger age.

That the examined patients had a normal grip force (Table) may be an argument for a specific contribution of the nigro-pallidal tracts to fine motor skills and gives rise to the idea of relating different tracts to different (motor) functions, as cerebellar tracts were associated with tremor²⁷ and prefrontal tracts with executive functions.²⁸ As a future perspective DTI-tractography may be developed to a helpful tool for connecting qualitatively and quantitatively severe and complex clinical deficits to the different involved networks.

Our study has limitations. We analyzed a small group of patients experiencing relatively mild impairment, with a special containment of localization and size of ischemic lesions, so the results cannot be generalized to the overall population of heterogeneous stroke patients. Moreover, results have to be considered exploratory and need to be validated in a larger cohort. Commonly known technical limitations of DTI as of today further limit the resolution so that we were not able to separately study different tracts belonging to the extrapyramidal system.

To conclude, we successfully reconstructed nigro-pallidal tracts representing a relevant part of the extrapyramidal system. We found DTI-metrics indicating reduced structural integrity of nigro-pallidal tracts ipsilateral but distant to the primary stroke lesion 3 months after stroke, which was associated with worse performance in fine motor skills. These results indicate specific contributions of extrapyramidal tracts to residual motor function after stroke beyond the well-known contribution of the CST. The study of damage to extrapyramidal tracts may deliver supportive information for prediction of motor recovery after stroke.