Matlab r2012b

The EEG signals were recorded using a Neuro Scan NuAmps Express system (Compumedics USA Inc., Charlotte, NC, USA) with 32-channels EEG cap, as shown in Figure 2. All EEG signals were examined using EEGLAB software (10.2.2.4b Version, UC San Diego, Swartz Center for Computational Neuroscience (SCCN), La Jolla, CA, USA) and MATLAB R2012b (The MathWorks Inc., Natick, MA, USA) [10 (link)]. In EEG cap, all 32-channels were positioned according to the international 10–20 system. The EEG signals were acquired at a sampling frequency rate of 500 Hz. We use an infinite impulse response (IIR) filter to eliminate linear trends in EEG signals. The EEG signals were filtered with a 1–50 Hz band pass IIR filter. The filter configuration was set to 1 Hz high pass and 50 Hz low pass to eliminate high frequency noise. The EEG signals that were considerably contaminated by artifacts, such as muscle activity, eye blinking, eye movement, and environmental noise were first removed manually and then by independent component analysis (ICA) to minimize their influence on the analysis of the EEG signal. In the artifact removal analysis, we found that about 12% of the epochs (trials) were so noisy in the raw EEG signals. Consequently, 12% noisy epochs were rejected from the raw EEG signal. We removed various artifacts, like muscle activities and eye blinking [10 (link),20 (link),21 (link)].

Data were down-sampled to 1000 Hz and filtered using a windowed since type I linear phase FIR filter (band pass: IAF ± 3.5 Hz, filter order: 6002) and did not induce a phase shift in the alpha range (6–14 Hz). The identical filter was used for the stimulation signal and the EEG data (Widmann and Schröger, 2012 (link)). Filtered EEG data are shown in Figure 3 (Box 2). For each condition, the first and last 2 s were discarded, since the steady-state condition has to build up and to avoid edge effects (Halbleib et al., 2012 (link)). All analysis steps were performed in MATLAB R2012b (The MathWorks Inc., Natick, MA, USA) and EEGLAB 11.0.4.3 (Delorme and Makeig, 2004 (link)).

The processing of the structural brain volume images was conducted using SPM12b software (Statistical Parametric Mapping, https://www.fil.ion.ucl.ac.uk/spm/), implemented in Matlab R2012b (The MathWorks Inc.). We used the unified segmentation approach [39 (link)] to segment grey matter, white matter, and cerebrospinal fluid. Volumes were further assessed for total grey matter, total white matter, total brain tissue (grey + white matter), and total intracranial volume (TIV: grey + white matter +cerebrospinal fluid). The Freesurfer image analysis suite version 5.1 (Martinos Center for Biomedical Imaging, Harvard-Massachusetts Institute of Technology) was used to measure hippocampal volume through automatic volumetric segmentation as previously described [40 (link)]. WMH were manually delineated on the FLAIR images [41 (link)].
All volumes were corrected for TIV using the analysis of covariance approach [42 (link)]. Total brain tissue volume was used for the creation of global models, whereas grey- and white-matter volumes, hippocampal volume, and WMH volume were used in forming the specific models.

I used SPM12 (Wellcome Department of Cognitive Neurology) implemented in MATLAB R2012b (The MathWorks Inc., Natick, MA, USA) for data preprocessing and GLM. Each participant underwent four fMRI-runs, each comprising 477 volumes. After having discarded the first four volumes of each run, all images were corrected for head movements. Slice-acquisition delays were corrected using the middle slice as reference. All images were normalized to the standard SPM12 EPI template, resampled to 2 mm isotropic voxel size, and spatially smoothed using an isotropic Gaussian kernel of 8 mm FWHM. Time series at each voxel for each participant were high-pass filtered at 220 s and pre-whitened by means of autoregressive model AR(1).

LRI utilizes a 10 × 10 × 0.5 mm CdWO₄ scintillator crystal (MTI Corporation), which is in contact with the tissue of interest, to convert ionizing radiation from emitted beta particles of tracers in the tissue into visible-range photons detectable in a sensitive microscope. The collected photons pass through a wound fiber bundle (IG-567, Schott AG) and are mapped onto a camera chip via a relaying lens (Schott AG). The wound fiber bundle allows motion of the imaging tip while the camera stays in place. The fiber bundle has a 40% transmission between 500 and 1200 nm, a numerical aperture of 0.63, and length 1.720 m. The 5 by 6.7 mm fiber bundle has 10 μm elements in a 6 × 6 array. The camera is a EMCCD (ProEM, Princeton Instruments, Trenton, NJ) with a 512 × 512 imaging chip with a pixel size of 16 μm. The imaging setup yields a visible field of view of 4.80 × 6.51 mm with 407 × 300 pixels. Images were analyzed using Matlab R2012b (Mathworks).

The raw MALDI-TOF MS spectra were processed in Matlab R2012b using the Mathworks bioinformatics tool box (MathWorks, Natick, MA, U.S.A). The workflow consisted of spectra resampling followed by baseline subtraction, smoothing and normalization of the total area under the curve (i.e. normalizing based on the total amount of sample protein ionized per spectrum). Normalization for total area under the curve insures that differences in saliva sample protein concentration, or in the protein concentration of the spotted sample, are compensated for. Reference spectra were used to align batches of spectra analyzed on different days and to compensate for inter-session instrument drift. Duplicate spectra were then averaged into one sample spectrum. Subsequently, peak detection was performed, followed by peak binning (peak coalescing) using a hierarchical clustering algorithm to calculate a common m/z reference peak vector. The final result was a 3-dimensional database of salivary peptide profiles consisting of sample ID’s, peak m/z values, and peak intensities. Peak identification was attempted using MALDI-MS/MS or by matching the peak m/z ratios to literature values from previous studies [19 (link), 20 (link)]. The spectra processing workflow is exemplified in S1 Fig.