EEG Studio FAQ

  1. What kind of unique features does this software have? Is this another EEG/ECoG toolbox based on Matlab/IDL?
  2. This software package, which mainly based on C/C++, is different from many EEG toolboxes, which are typically based on Matlab/IDL. For example, this software package provides following outstanding features:
    • Very productive, intuitive GUI, very fast with artificial intelligence (e.g. “drag-drop” to merge two data sets, volumetric source scan without making any assumption about the number of sources or possible location).
    • Specially designed data analysis modules to utilizing both low- and high-frequency neuromagnetic/electric signals for characterize brain functions in spatial, temporal and frequency domains.
    • Volumetric source imaging with newly developed source localization algorithms (wavelet-beamformer, accumulated source imaging).
    • Dynamic magnetic source imaging (dMSI) for visualizing sources in real-time or accumulated mode.
    • Voxel multi-coding for precise determination of brain activity/activation while significantly minimizing EEG/ECOG/MEG inverse problems.
    • New signal processing algorithms which are not available in other software packages (e.g. polarity contour maps, accumulated spectrograms, etc.).
    • Intuitive graphic user interface (GUI), which provides real-time “toolbar-tips” for ease of use or better usability without memorizing all the commands.
    • High-performance and handle a huge amounts of data using assembling codes and parallel computing.
    • Optimized functions for analysis of high-frequency EEG/ECOG/MEG data (e.g. built-in re-sampling function, handle a huge amounts of EEG/ECOG/MEG data)
    • Quantitative assessment of coherence of brain activity/activation at source levels using volumetric coherence analysis, which analyze every possible voxel-pair of the entire brain.
    • Outstanding 2D and 3D data visualization tools.

  3. Where and How to download the EEG Studio software?
  4. If you are reading this file online, you may already see the EEG Studio program because this file typically comes with the EEG Studio program or software. If you read this file offline, here is the website to download the EEG Studioprogram:

    http://www.mecurer.com/

    Once you identified the software, which you are interested in, please select it and press the Right Mouse Button. Then, you can select the “Download” Menu. Once you “Click” the “Download” sub-menu, the Browser (e.g. Internet Explorer, Firefox, Chrome…) should start to download the file.

    The MEG Program has more than hundreds of files. For you convenient, all files have been “wrapped up” and compressed as a single file for fast and easy transfer. Once you downloaded the compressed file (single file), of course, you need to uncompress it (typically, unzip it). The unzipped program should unzip all files into one fold which has approximately 666 files (in 2012, it may increase or decrease in the future).

    We regularly update the software, thus, please check the website for any updates.

  5. Do I have to install the software? Is it complicated to install the software?
  6. No. The “green” version of EEG Studio can be used on Windows 7 (64 bit) without installation. The procedure is very simple: (1) copy the fold of the entire EEG Studio to your computer, where you like to store it; (2) run “EEG Studio” by double-click the “EEG Studio.exe” file.

    For your convenience, you may make a short-cut with the following steps: (1)select the “EEG Studio.exe” file; then (2) press the right Mouse Button; (3)click the “Create short cut” sub-menu in the “Pop up” Menu; (3) drag the created short-cut to the desktop. You may rename the short-cut if you like.

    If you do not like it, you may simply delete the entire fold and the software should go completely without taking any of your computer space. Noteworthy, the software does not change any of the settings of your computer. There is no virus, no malware, and no junk item. Thus, the software is pretty “healthy” and “green”.

  7. I tried to run the EEG Studio but got an error message – For example, “The program can’t start because mfc100u.dll is missing”. What is wrong?
  8. You may need to install a small update from Microsoft Company called “vcredist_64” or “vcredist_32”. We have downloaded it and put it in the same fold of the MEG Processor. You may select the correction version that matches with your windows and install it

    Once you download the file, please install it (since it is from Microsoft Inc., it should be safe) and try again.

    You may also check the Microsoft website to find any updates that are needed for running new software developed with Visual C++ 2010, which is the IDE used for compiling EEG Studio (2010-2012).

    If you downloaded the new version of EEG Studio VS2013, you may also check the Microsoft website to find any updates that are needed for running new software developed with Visual C++ 2013, which is the IDE used for compiling EEG Studio (2013-2014).

  9. I tried to run the EEG Studio on Windows XP and Window 7 (32 bit), but got an error message about compatibility. What is the problem?
  10. Our work currently focuses on Windows platform, in particularly, 64-bit Window 7 and Window 8. Therefore, the updated version of EEG Studio has only been tested on 64-bit of windows. If you run our new software on 32-bit windows, it may not work. You need to download the EEG Studio 32 bit version.

    Though EEG Studio 32 bit version does work on 32 bit windows. However, please use 64 bit version of our software (we may also move to 64 bit Mac or other OS, but not 32-bits). We consider that 64 bits OS is the future of computer for MEG technology because 64 bits OS supports more memory and can process a huge amount of data efficiently.

    Please also note that Magnetic Source Locator (MSL) and BrainX can run on both 32 and 64 bits of Windows. However, BrainX needs different drivers for 32 and 64 bit windows if you would like to access or control the parallel ports.

  11. What kind of computer do I need to run MEG Processor?
  12. Though the software package (Dec. 2012) was developed on a workstation with very high configuration (e.g. it has 144 GB memory), the software package has been tested on office desktop computer as well as notebook/laptop with 3 GB memory. Here is an example for desktop computer:
    • Windows 7 or 8 (64 bits)
    • 3rd generation Intel(R) Core(TM) Processor (2.7 GHz, 8MB L3 Cache)
    • 24 GB DDR3 System Memory
    • 1 TB Hard Drive
    • NVIDIA(R) GeForce(R) GT 650M Graphics with 2GB GDDR 5 video memory
    • 20-inch LED Display (1920 x 1080, 32 bit color)
    • 8X DVD+/-R/RW
    • Standard Keyboard

  13. The software works well for daily tasks, but it crushes daily for spectral-Cxc and some source localization modules. What is the problem? Why don’t you fix the bugs?
  14. EEG Studio was preliminarily developed and optimized for analyzing high-frequency brain activity or high-frequency MEG/EE signals. We are aware that this is a problem in the studies of high frequency brain activity, because the study of high-frequency brain activity requires a high-sampling rate of EEG/ECOG/MEG data, which produces an unusually high number of MEG data points. Consequently, considerable computing power and time are required to analyze these data points. To ensure it is practically usable, the program loads all the data into memory to ensure the work can be done in a timely manner. The tradeoff is obvious that the program can easily run out of memory and crash!

    The solutions we have developed are: (1) re-sampling data in each module for low-frequency analysis in high-sampling rate; (2) utilizing parallel computing to improve performance and limit the use of memory. However, as the computer technology is developing rapidly, we anticipate this problem will be easily fixed in the near future (e.g. using GPU/SDD to improve performance or use Cloud computing to get work done).

    Please note that many advanced functions of EEG Studio have been tested on computer workstations with more than 200 GB memory (typical desktop computer has 3-8 GB memory). You may ignore those functions if your computer has limited memory.

    All functions with “…Spectral Cxc..” require a lot of memory.

  15. Why you use some strange terms such as “CxC”? No one can understand them. You need to change those terms.
  16. The development of EEG Studio aims to take the advantages of high-spatiotemporal resolution of MEG with novel signal processing algorithms and/or methods. At least, we have not found any better terms to replace those “strange terms”.

    Specifically, “CxC” indicates operations applying to two channels or a channel-pair. An operation of two channels is a channel-cross-channel (“CxC”) operation that can produce many possible results. Here are some powerful usages:
    1. Virtual channels: similar to many software programs, EEG Studio can produce virtual channel by subtracting or adding two physical channels. Instead of manually defining a virtual channel by subtracting two physical channels or a channel-pair, “CxC” can generate all possible virtual channels by using available physical channels with a set of operations. All the operations can be done by a few clicks. Once you try it, you will the power and usefulness of the “CxC” function. It is amazing, subjective and very efficient.
    2. Coherence at sensor levels: coherence, correlation and association of two-sensor can be done by simply selecting “Covariance” or “Correlation” operation. “CxC” can analyze the relationship of all possible physical channel pair according to the setting of operations.
    3. Covariance matrices for source localization: many EEG/ECOG/MEG source localization algorithms require covariance matrices from sensor data. The “CxC” functions enable users to compute, preview and check the covariance matrices for source localization.
    Text Box: A channel does not necessarily mean a sensor. In this manual, channels includes sensor channels which detect brain signal, reference channels which are used for references or detecting noise, EKG channels and EMG channels.

  17. I tried to run the EEG Studio on Windows 8, but I got an error message about “missing D3dx9_34.gll” something like that, what’s the problem?
  18. EEG Studio uses Direct3D to render 3D images. If your computer does not have Direct3D dlls which are included in DirectX from Microsoft Inc., the program will show error messages. Please download the latest version of DirectX from the website or copy the “dxwebsetup.exe” file coming with MEG Processor. You may also download the file from the following link:

    http://www.mecurer.com/

    Once you download the file, please install it (since it is from Microsoft Inc., it should be safe) and try again. Please note that EEG Studio supports both Direct3D and OpenGL. EEG Studio (~2012) currently supports Direct3D 9.0.

  19. EEG/ECOG/MEG signals are generated by the neurons in the brain. As far as I know, neurons fire at a low frequency range (< 100 Hz). How can the brain generate EEG/ECOG/MEG signals above 100 Hz?
  20. EEG/ECOG/MEG does not detect electric or magnetic signals from a single neuron. In fact, MEG detects signals from approximately 50,000 neurons, which activate “simultaneously” in a synchronized manner. Therefore, EEG/ECOG/MEG signals are the spatiotemporal summation. The frequency components may reflect the spatiotemporal changes as well as orientation changes. High frequency signals may be generated by “out of phase” activation from a group neurons. From our point of view, high-frequency MEG signals does not generated by single neuron. Instead, high-frequency EEG/ECoG signals reflect the spatiotemporal organization of neural activation in a variety of directions.

  21. It is well-known that the brain generates brain signals below 100 Hz such as alpha (8-12 Hz), typically below 70 Hz. How can the software detect signals in high-frequency ranges?
  22. Low-frequency brain signals such as alpha activity, K-complex and spike-wave discharges do provide very important information about the brain function. Those low-frequency signals have large amplitude and therefore can be easily identifiable in waveforms. However, according to our observation with thousands of MEG datasets

    http://clinicaltrials.gov/show/NCT00600717

    Low-frequency MEG signals are typically generated from a large area or from extended cortical sources. High-frequency MEG signals are typically generated from a small area or focal cortical sources. In other words, low-frequency signals are generated a large group of neurons, while high-frequency signals are probably generated by a small group of neurons. If the high-frequency signals and low-frequency signals are generated from a similar area or a closed region, low-frequency signals may be spatiotemporal summation of high-frequency signals. Therefore, high-frequency signals encode more spatial information than that of low-frequency signals.

    From source localization point of view, high-frequency signals are highly localized while low-frequency signals are diffused. Therefore, to analyze brain activation/activity at source space or source levels, high-frequency neuromagnetic signals would be the key in the future. Yes, I believe MEG can detect brain activity/activation above 100 Hz.

  23. According to the textbook, the neuron generates signals below 100 Hz. How can you use MEG to detect brain signals above 100 Hz? Do you know that MEG signals are magnetic parts of the electrical signal generated by neurons in the brain? It’s an ill focused area.
  24. MEG does not detect signals from a single neuron. Instead, MEG detects the spatiotemporal summation of magnetic signals from a group of neurons (> 50,000 neurons) which are firing simultaneously in a similar orientation. The frequency of MEG data does not represent the frequency of a neuron. The frequency characteristics of MEG data reflect the spatiotemporal patterns of the activation of a group of neurons or other “cells” which generate electromagnetic signals.

    From our point of point of view, high-frequency signals are highly localized while low-frequency signals are diffused. Therefore, to analyze brain activation/activity at source space or source levels, high-frequency neuromagnetic signals would be the key in the future. We believe the study of high-frequency neuromagnetic signals represent the future of MEG research.

  25. The 3D rendering in this program is fast and very powerful. However, as I was playing with it, I saw a black window occasionally. Do you know why?
  26. The 3D rendering engine supports both Direct3D and OpenGL. Direct 3D seems very fast on windows computer.

    To visualize the structural and functional components, this program provides three kinds of selections for precisely imaging the data of interest. If the selections are not used appropriately, you will see a “black window” without anything visible.

    Spatial Selection: a 3D selector has been designed to spatially select volumetric data. The selector provides three modes: (1) Select- all (no selection, all voxels or data are visible); (2) Select-out: voxels out of the selected region are visible; (3) Select-In: voxels within the selected region are visible. If you used the 3D selector to select all the data, and then pick the “Select-out” mode”, you will see a “black window” because there is nothing to show. In this case, pick the “Select-All” option.
    • Data Marks: This program provides at least three kinds of marks (red, blue and yellow) for image segmentation, voxel (pixel) label and analysis/measurement. You may selectively visualize voxels with: (1) special mark(s); and (2) voxels. If you have not mark or label any voxels and try to visualize special mark(s), you will see a “black window” because there is no mark to be visualized. In this case, pick the view “All Voxels”.
    • Range of Data Value: This program enables the visualization of volumetric data in certain range (from minimum value to maximum value in 3D View Control). For example, MRI/CT images typically have gray scale data in limited ranges. If you select very high or low gray scale value, you will see a “black window” because all available MRI/CT voxels are out of the selected range. In this case, please click the “Autofit” button to automatically adjust the range for visualization.
    In fact, another “back window” typically appears when users hide all 3D objects or there are no 3D objects available.

    Please note that, this program supports EEG/ECOG/MEG sensor, sources, MRI/CT images and neural networks. If you perform source scan without MRI/CT image (structural data), the program will use templates to simulate structural data. In this case, it seems the program has structural data but there is nothing visible.

  27. Are there any publications that have used this software? As this software has unique functions to detect high-frequency oscillation, can you show me any example?
  28. Yes, many papers used this software. You may find some sample publications related to high-frequency oscillation in the following website:

    http://clinicaltrials.gov/ct2/show/NCT00600717

  29. I see clear responses in averaged waveform. However, I do not see any high-frequency component with band-pass filter (100~1000 Hz). Where are the high-frequency components?
  30. The conventional averaging reveals “time-locked” signals. Once you averaged the waveforms, the high-frequency components have more than likely gone because high-frequency oscillations might not “time-locked”.

    In addition to the well-known trigger jitter; there is natural variation of brain responses. High-frequency oscillations may vary in latency and the conventional averaging cannot reveal it. One approach is to use “accumulated spectrogram” and “accumulated source imaging” to reveal those high-frequency components (e.g. Xiang et al. Front Neuroinform. 2014 May 21;8:57).

  31. The size of the file appears to be bigger than I anticipated. It takes considerable hard disk space. Is it possible to compress the file and use less space?
  32. To ensure the quality of the results and extract useful information from high-frequency signals, the software uses double-float numbers to represent the numbers. The detailed discussion can be found in previous reports (e.g. Xiang et al. Front Neuroinform. 2014 May 21;8:57). Though double-float generates outstanding results and can accurately get high-frequency components from multi-frequency signals (the amplitude of low-frequency signals are typically much larger than that of the high-frequency signals), the size of the file can be huge.

    One solution is to compress the file (or fold of the files) to save space. EEG Studio can open the compressed file. One of the well-tested software is “WinRAR”, which can be freely downloaded. Here are the steps of the compressing and drag-and-drop opening.

  33. What shall I do if I found some bugs? Is there anyone who can answer my question(s)? May I send comments and suggestions to the programmers?
  34. Please feel free to send emails to: support@mecurer.com