Nineteen patients with ALS and 20 healthy control subjects underwent clinical assessment, and MRI examinations, including DTI. Detailed demographic date and clinical characteristic of the entire study cohort are shown in Table 1. All subjects were right-handed as assessed by the Edinburgh handedness inventory (27). At the time point of examination 4 patients showed definitive ALS, 8 probable ALS and 4 possible ALS according to the revised El Escorial criteria (1). Three patients displayed lower motor neuron variant. Mean disease duration was 34.8 ± 34.8 months. Clinical assessment consisted of neurological examination with special respect of handedness; MMs were evaluated by sequential finger tapping of one hand without optic control and observation of MMs according to the procedure of Woods and Teuber (28), and the evaluation of the revised ALS functional rating scale (-R) (29). Furthermore, to address the upper motor neuron (UMN) involvement an UMN “burden” (UMNB) was calculated by totalling the number of pathological UMN signs on examination (maximum score 16) (30). For cognitive screening we used the “Montreal Cognitive assessment” (MOCA) (31). The MOCA score ranged between 18 and 29 with an average of 24.5 ± 4 for the ALS patients, and between 27 and 29 with an average of 27.6 ± 0.7 for the controls (P = 0.006). Controls did not have cognitive complaints and scored within 1.5 standard deviations of the MOCA age and education adjusted norm value.

Patients and control individuals were only included in the study if written consent was given. The study was approved by the Institutional Review Board of the Medical Faculty, University of Rostock (A-2011-0026, A 2012-0083).

We performed TMS in all patients and in 10 of 20 healthy controls. Central motor conduction time (CMCT), motor evoked potentials (MEP) amplitudes, and contralateral silent period (cSP) were determined in all ALS patients and control subjects. MEPs were recorded from the first dorsal interosseus muscle (FDI) and from the anterior tibial muscle (TA) using a standard circular coil (outside diameter 9 cm) connected to a Magstim 200 stimulator (Magstim Co., Whitland, Dyfeld, UK). For data acquisition a commercially available MEP system was employed (Brain Quick System Plus, Inomed, Erlangen, Germany). Investigation of the ipsilateral silent period (iSP) was performed with a focal coil (external loop diameter 7 cm). The coil was oriented to induce a posterior-anteriorly directed current flow to the hand area of the motor cortex; the point of optimal excitability (POE) of the FDI muscle was determined over the contralateral motor cortex; TMS was applied at the POE with 1.5 times resting motor threshold (RMT) while subjects performed a maximum tonic activation of the ipsilateral FDI muscle and while they kept the contralateral FDI muscle relaxed as published in detail previously (3). For determination of iSP latency and duration 10 trials for each hemisphere were used. After offline rectification and averaging of EMG signals, iSP parameters were analyzed.

MRI acquisitions of the brain were conducted using a 3-Tesla MRI scanner with a 32-channel phased-array head coil and parallel imaging capabilities (Magnetom Verio, Siemens, Erlangen, Germany, software syngo MR B17). Subjects were scanned in a single session without changing their position in the scanner. The following sequences were used: We acquired a sagittal high-resolutionT1-weighted magnetization prepared rapidly acquired gradient echo (MP-RAGE) 3D-sequence, matrix size of 256 × 256 × 192, isometric voxel size 1.0 mm³), TE/TI/TR of 4.82 ms/1,100 ms/2,500 ms, flip angle 7°. To identify white matter lesions a two-dimensional T2-weighted sequence was performed (fluid attenuation inversion recovery FLAIR, matrix size of 384 × 187, 24 slices with slice thickness of 5.0 mm, TE/TI/TR of 94 ms/2,500 ms/9,000 ms, flip angle 150°). Diffusion-weighted imaging was performed with an echo-planar-imaging sequence (TE/TR 81 ms/12,700 ms) Diffusion gradients were applied in 30 different spatial directions. The b values were 0 and 1,000 s/mm². The images had a matrix size of 128 × 128 mm, slice thickness of 2 mm, the resulting voxel size was 2.0 × 2.0 × 2.0 mm².

DTI data were pre-processed using the DTI tool box of the FSL software (http://www.fmrib.ox.ac.uk/fsl/, FMRIB, Oxford, UK, Version 4.1) (32). We first applied corrections for eddy currents and head motion. The skull was stripped using Brain Extraction Tool for differentiation of brain tissue and non-brain tissue using a binary mask (threshold 0.15–0.3) and the diffusion tensors were fitted to the data with DTI fit (FMRIB Image Analysis Group, Oxford, UK). Fractional anisotropy (FA), and mean diffusivity (MD) maps as well as axial diffusivity (AD) and radial diffusivity (RD) maps were created.

TBSS analysis (tract based spatial statistics) were performed using FMRIB software (version 4.1 www.fmrib.ox.ac.uk/fsl) (33). TBSS allows spatial reorientation of FA maps into a standard space without systematic effects of spatial transformation on fiber tract directionality and without the need to select a spatial smoothing kernel that may impact upon the desired effects (22). Processing of diffusivity maps consisted of the following steps: FA maps were transformed to MNI (Montreal Neurologic Institute) space. Normalized FA maps of all subjects were averaged and skeletonized with a standardized threshold of 0.2. Based on the study specific with matter skeletons an individual skeleton for each subject was created.

For specific analysis of changes in diffusivity of individual fiber tracts we conducted a region of interest (ROI) analysis. The selection of ROIs from JHU was done as described before (34). Based on the JHU white matter atlas (35) we defined seven regions: corpus callosum with subregions genu, truncus, and splenium, corticospinal tract, and crus posterior of the internal capsula on both sides. Data processing was done by Matlab software 2013a (MathWorks Inc., MA, Natrick, USA) (Figure 1).

For analysis of correlation between TMS values and DTI measures we performed a region of interest (ROI) analysis using MarsBaR (36) to extract mean values from corpus callosum and left and right corticospinal tract, respectively.

Statistical analysis was conducted for the clinical scores, TMS measurements and ROI data in SPSS (version IBM SPSS statistics 20). For the whole brain analysis using the tool box “Randomize” under TBSS in FMRIB (version 4.1 www.fmrib.ox.ac.uk/fsl) was used. All clinical scores and TMS parameters showed a normal distribution, which was tested by the use of Kolmogorov Smirnov test (P > 0.1). Differences in TMS parameters were compared between groups using Student's t-test. The results of iSP measurements were dichotomized in pathological vs. non-pathological, and analyzed using Chi² test.

For ROI analysis, group comparisons of FA of patients and controls were performed according to the findings of pathological iSP and non-pathological iSP in an univariate variance analysis and subsequent post-hoc tests. Furthermore, correlations between iSP findings and ALSFRS and FA parameters of all investigated ROIs was calculated. For correlation analysis Spearman-Rank-correlation were used and significance was set at P < 0.05.For whole brain analysis, we investigated group differences in fractional anisotropy and mean diffusivity using the general linear model on a voxel basis. Whole brain analyses were performed for all diffusion parameters. FA, MD, RD, and AD maps were compared by univariate variance analysis (factor: group, covariate: ALSFRS-R and UMNB). To avoid overestimation of significance, a test of multiple comparisons was applied (family wise error method) and significance was set at P < 0.05.

TMS investigation in ALS patients revealed a mean iSP latency of 41.3 ± 5.4 ms/40.9 ± 5.7 ms (right/left), which was longer than that of the healthy controls 39.0 ± 4.6/37.7 ± 6.3 (right/left) (p = 0.006/0.012). No differences were found for the iSP duration between the ALS patients and control subjects. Twelve ALS patients (63.2%), but none of the controls showed loss of iSP. cSP was numerically shorter in ALS patients than in controls without reaching significance. ALS patients had a significantly lower mean RMT compared to controls (TMS results are shown in detail in Table 2 and Figure 2).

Concerning the analyses of pre-specified ROIs FA values of the CST were significantly lower in ALS patients compared with controls. FA values of the different ROIs of the corpus callosum (genu, splenium, and body) did not differ between both groups (Table 3). AD and RD values showed similar findings. Whole brain analysis did not show any differences in FA, MD, RD, and AD between ALS patients and controls at an FWE corrected level of significance.

We found significant negative correlations of FA values of the CST with RMT values on both sides in ALS patients, but not in controls (Table 4). There was no significant correlation between FA values and iSP parameters (Table 4). In contrast, there was a significant correlation of ALS-FRS-R scores with FA values of the CST on both sides (r CST-right, ALS-FRS-R = 0.58, p = 0.009 and r CST-left, LS-FRS-R = 0.05, p = 0.02), but not with FA values of the corpus callosum.

ST has done the listed works below (all in Germany): MSD Sharp & Dohme GmbH, Lindenplatz 1, 85540 Haar

11.09.2018—Quality circle for physicians in Kühlungsborn, Talk: “Dementia and Diabetes—current report”

14.11.2018—MSD Expert-forum: NAB Alzheimer in Munich, participator as consultant

13.08.2019—Event “Diabetes and Dementia” in Rostock, Talk: “Dementia and Diabetes—current report” ROCHE Pharma AG, Emil-Barell-Str. 1, 79639 Grenzach-Wyhlen

12.09.2019-−3. National Advisory-Board in Frankfurt am Main, participation as consultant

27.09.2019—ROCHE Symposium at the DGN Congress in Stuttgart, Talk: “Amyloid as target for diagnosis and treatment in Alzheimer's disease.” The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.