eegviewhelp.txt

    1) Load the correct neuroscanascii file.
    2) Remember to zero out or add zeroed channels in the popup box that immediately appears. I believe that Rodriguez had quite as few - the last 3 channels were zeroed out, AT 1 and 12 (if I recall), as well as 2 missing channels - maybe P 61 and P 63? This should be obvious from the scan program, and hopefully we wrote it down somewhere too.
    3) Pick the correct .csv file. If the program warns you of missing channels, this means that the header for the .neuroscanascii program does not match the csv file for some channels. It's better to edit the header directly than to edit the csv file, in my opinion.
    4) Note: if a box pops up either before or after the csv file selection window informing you of one or multiple .amp files to choose from, this means that there are .amp files in the same directory as the .neuroscanascii file. I most certainly have to turn off this feature, as the program will load any .amp file it finds without asking first, and if the .amp file doesn't match either the csv or the .neuroscanascii, it leads to crashes and unpredictable behavior. For now, rename the .amp file to .oldamp or move it. Recall that .amp is only useful for the gamma activation sequence, and can be chosen from the grid properties box.
    5) View3 window should now pop up. Click the grid properties button on the left.
    6) Set the opacity of the items in the display, and go to the final tab in grid properties to set the pipe width. Pipes are a little too thin by default (another thing on my laundry list).
    7) If you are loading gamma activation data, go to the scalar data tab. Here, you can press open to pick a .dat and .amp file. Be sure that everything matches. However, the .amp file does not have to match the csv or .neuroscanascii header naming at this stage. The .amp file here only determines which channels from the .dat file are displayed in the gamma graph popup window that comes when you press execute.
    8) Before you press execute, should you wish to integrate the gamma graph into the view3 window, check the "add to view3" button. Notes: There is a memory leak in this function which prevents long, hires movies when the gamma window is added into view3. On my home computer I managed to get about 1000 steps before the computer crashed, so as long as the number of steps in the autopager is < 500 you should be ok.
    9) Pressing execute pops up a graph with the channels you've indicated using their .amp numbers, in its own window. Sliding this bar changes the time in the trial that they are displayed.
    10) Coherence: The "coherence options" button brings up an options menu. The "calculate over all data" checkbox allows you to set eegview to calculate over the entire data file at once. If the data file is split (epoched) by sweeps, you will want to do this. Then, set the sweep length in points to the length of one sweep (for instance, 6 seconds is 3072 points at a rate of 512. The rate is hardcoded into the header of the .neuroscanascii file). The length option chooses the length of the window of coherence. So each run of the coherence calculator will go over the chosen length, starting at the chosen offset, for each and every trial as determined by the sweep length. If you get the length wrong, your data may be off.
    11) To clarify - if you want to pinpoint coherence at a certain time (say 2.0 seconds) in a file with rate 512 and sweep length 6 seconds, and you want the coherence averaged over all trials, you should choose in the options: calculate over all data, sweep length 3072, length 128 (or 64 or 256 all seem to be reasonable values), offset [anywhere from 900 to 1000 in order to encompass point 1024 = 2.0 seconds]. Press OK, then press Compute Coherence.
    12) there is a data dump button at the bottom right that allows you to pick a filename and some basic options (that are obvious and usually useful by default as they are). This button, when checked, will dump all coherence data over all bands, and all phases, for each and every pair of electrodes. It will dump it every time you run the calculate coherence button. It will add a slightly useful header before each set of lines of data. Therefore, if you are just playing around with the coherence button, your data will be harder to parse than if you turn on the datadump and then run the autopager.
    13) Autopager: recall first that if you've opened, closed, and opened a window such as view3, the autopager will crash the whole suite.
Now, Click Autopage/Movie. The first set of options mintime & maxtime are not very useful to change - their default values will run the pager over the full set of data. If you are paging through the .neuroscanascii file, using multiple averaged sweeps, changing these values will mean cutting off sweeps.
The third option, Time intv., allows you to choose the step length. When paging the .neuroscanascii file, this option is equivalent to changing the "offset" option in the "Coherence Options" menu. Of course, in this mode, the program will continue to step by this offset amount at intervals as long as the autopager is running. Put another way, remember the coherence average window that covers every sweep, at an equal time offset from the beginning of the sweeps - the step length here will move that window by step length points, every time the autopager takes another step.
    14) If you've added a .dat file in the scalar data tab of the grid manager and you've checked the "add to view3" button there, you will be presented with a choice of 3 radio buttons right below the step intv box - these allow you to page the .neuroscanascii file alone (as we've been discussing), only page the scalar data, or page both at the same time. Here, just to be clear, by "page" we mean "step", drive, move forward, etc.
    15) Note that changing radio buttons actually changes the default values for max and min time, so hopefully there will still be no reason to change these. Note also that for scalar data, these values will be in seconds or in points depending on how you load the scalar data from the grid manager (there is an option in the scalar data tab to load the data only in a certain time range, measured in seconds).
    16) When the program pages the scalar data, it literally moves the scroll bar in the scalar data window, which updates the gamma activation displayed in the view3 window, as well as the line on the graph embedded in the view3 window. Important note: If you wish to see the gamma activation, you may need to change the visible threshold. This can be done at the top of the same scalar data tab in the grid manager. Good values to start from are something like 0 - 5, and you can experiment with values that make sense from there.
    17) Paging both - this will only make sense when your .neuroscanascii file is epoched into multiple sweeps, and your dat file is the same length as one sweep. The autopager will scroll through the .dat file at the same step length (in points) as it moves the coherence average window over the sweeps.
    18) Before starting the autopager, click the "save images"  check box and pick a useful file name, preferably in an empty directory.
    19) Rotation is hard. Basically, you can add frames by manipulating the actual objects in the view3 window and clicking "add frame". The autopager will remember each position that the objects are in at the moment that you click add frame, and then rotate the images smoothly through these frames in the order in which you add them, such that there are an equal number of steps between each frame.
To clarify - if you have more steps than frames (as is usually the case, like say 300 steps and 5 frames), the autopager will start at the first added frame, and go through numsteps/numframes steps between each subsequent frame.
    20) Now, say that your frames are far apart, and you would rather the animation of spinning brain objects be smoother - this is where the "frames per time step" option comes in. As it implies, if you would like the autopager to interpolate frames in between each calculated time step, this is how many will be added for every step. If you've set your step interval low, this should not be a high number, or you will end up with more pictures than you need for a smooth animation. Note that 20 or 25 frames per second looks very smooth.
As an example, if your sweep length is 6 seconds = 3072 points (rate=512), and you are stepping through at an interval of 8 points per step, the program will turn 6 seconds of data into 384 pictures of data. If you want to show the movie in realtime, you already have a rate of 64 frames per second (that's 384/6), so you will not need to interpolate any extra frames in order to have a smooth animation. Thus, you can set "frames per time step" to 1. If you have many fewer pictures, you may want to set it to 3 or 5. Do the math...
    21) Clicking the "interpolate" button simply tells you how many pictures total you will end up with. That's it. It's not necessary, really. And if you click it, you can still change your mind and add more frames. You can also click "clear frames" at any time, and try again.
    22) When you are ready, put the view3 window in the foreground, move the autopager off to the side, click forward, and leave the computer alone. The autopage function takes a picture of the view3 window at each step. If the mouse is over the window, the mouse will be in the picture. If the window is minimized, the picture will come out all black (you left the lens on...). It's OK if your monitor shuts itself off though, because the computer will still be displaying in memory.
    23) Recall the datadump function. If it is running, now would be when the program is adding data to it. Thus, running the autopager over and over will result in a very confused datadump file.
    24) When the autopager is done,you can close the program and open up the picture folder. The autopager also takes a picture of the eegview herald window at the moment, which usually isn't very useful. To easily erase all of these, order the pictures in the folder by size - this should naturally separate out all of the herald images, which you can then delete.
    25) On Ubuntu, an easy way to stitch the pictures into a movie is to run the following command in the terminal at the folder where the pictures are stored: mencoder "mf://*.png" -mf fps=27 -ovc lavc -lavcopts vbitrate=15000 vcodex=mpeg4 -oac copy -o gm.stm.t1_theta_w128_4.mpg
Note that you can change the frames per second ("fps") of the movie, as well as the filename at the end. This tool will output the movie file in the same folder unless you specify otherwise.