Society for Neuroscience Annual Meeting, San Diego, CA, 2018.
Neuroimaging studies are accumulating fast. A significant number of these studies use functional magnetic resonance imaging (fMRI) and report stereotactic brain coordinates. In the last 15 years meta-analytic software tools have been developed to identify over-arching data agreement across studies (e.g., http://www.brainmap.org/). Meta-analytic studies help establish statistical concordance and quantitatively summarize large amounts of evidence. To date there are 944 papers on fMRI meta-analyses, as indexed by Web of Science (WOS; 28/04/18). Before analyzing coordinates researchers have to compile, systematically review relevant literature and extract stereotaxic coordinates. One process of pooling information from the articles requires manual search of the articles and manual extraction the relevant data, such as coordinates (i.e., foci), contrasts (i.e., experiments) and types of analyses (whole-brain or region of interest). Another available approach is offered by software with pre-extracted information, such as Sleuth (http://brainmap.org/sleuth/), Neurosynth (http://neurosynth.org/) and other open-source programs. Critically, these methods do not have up to date datasets covering only a limited number of studies (e.g., 11406 papers in the Neurosynth and 3294 papers in the Sleuth 2.4 at the 28/04/2018), whereas, a WOS search for the keyword (“fMRI”) yields 61976 papers. To improve the quality of the manual search for area-based meta-analyses and increase the speed of the identification of the foci of interest, we developed CoordsFinder - standalone graphical interface software for addressing the challenge of processing multiple fMRI articles reporting data in coordinate space. The software is written using WPF (C# and XAML), based on .NET Framework 4.5.2, and it supports Microsoft Windows 7 operating system or higher. The CoordsFinder estimates the foci uploaded in the software manually and searches for it inside the specified folder, which contains the pdf files of the papers, as this is the most common file format for articles. Foci coordinates can be found both in tables and in a plain text of the articles. The foci file uploaded could contain MNI or TAL space coordinates, and the software can indicate each type. In the current version, CoordsFinder can explore only files stored at the user’s computer, and process 274 papers per minute for a typical computer. Practically this software provides a solution for automatically extracting coordinates from multiple articles for effectively organizing and further analyzing data already available in the literature.
Cortical mapping with transcranial magnetic stimulation (TMS) is a promising approach for non-invasive investigation of the motor cortex in humans. However, apart from the parameters reflecting general corticospinal excitability, e.g., resting motor threshold (RMT) and mean amplitude of the motor evoked potentials, the reproducibility of other TMS motor maps metrics remains controversial or unknown. We studied a test-retest reliability of TMS cortical maps corresponding to multiple hand muscles. We took into account both standard size parameters such as map`s areas as well as novel parameters such as intricate muscle-specific excitability profiles. The study included 18 young healthy right-handed male volunteers. We used MRI-navigated TMS to stimulate left motor cortex in two mapping sessions separated by 5-10 days. For the mapping we used a grid of 53-58 points each being stimulated in a pseudo-random order five times. Second day TMS session was an exact repetition of the first day session. An analysis was performed using custom-made software TMSmap (http://tmsmap.ru/). We used intra-class correlation coefficient (ICC) to assess reliability of map areas, volumes and the extent of the different muscles overlap. For the quantative comparison of the cortical excitability profiles of individual muscles we utilized earth mover's distance metrics (EMD). We found that RMT remained the same across two testing sessions in all but two subjects in whom it changed by one percent. ICC for the same muscle representation could be considered as good (0.73-0.85) for areas, moderate for the extent of the different muscles maps overlap (0.7) and poor (0.45-0.49) for volumes. An average shift for hotspots was ~10 mm and for centers of gravity it was ~3 mm. When assessing individual excitability profiles, we found significantly smaller normalized EMD (higher reproducibility) for the same muscle representations across days than for the different muscle representations across days (P<0,0001). The obtained results provide evidence that not only general excitability but also other specific features including standard characteristics (areas, volume) and even excitability profiles of the cortical muscle representation can be reliably traced with TMS motor mapping. This in turn indicates that the existence of the complex TMS cortical representations doesn’t simply indicate stochastic fluctuations in the corticospinal excitability during TMS mapping procedure but rather demonstrates a possibility to probe with TMS cortical organization reflecting intricate descending projections relating to specific muscles.