ThoughtStream Monitor Windows Beta Software

[neuroasis]

At some point performance and memory will become a consideration. Over a long session there would be tens of thousands of data points. Most aren't very useful to display. The overall curve and trend is what is useful. I'm thinking of ways of doing smart filtering and zooming could be one way to do that.

The spreadsheet data table that I will use to load and compare sessions should have some kind of 'dynamic' frame where it only loads one page of the current view and maybe one above and one below. So it is a similar concept. The graph and the temp file on disk are the storage places right now but I could create buffers that fill up when called upon.

It can get rather complex really. It just depends on what is most useful. I also plan to store sessions in a database and tie in with our other software, like the KBS editor, so that you can compare the result of a light and sound session with the biofeedback readings generated.

I wonder what is the average length of most user sessions? Probably similar to light and sound from 15 minutes to 45 minutes. Anonymous statistics would be helpful to learn this.

Anyway, I am open to ideas and suggestions. We'll evolve the software over time.

An Android version is a real possibility. I already have read TS data on my Android tablet. And I know of a graphing framework that seems not too hard to implement. It's good to have this desktop app out so that people can try it and give feedback so we learn what is most needed.

[Andy]

At some point performance and memory will become a consideration. Over a long session there would be tens of thousands of data points.

At least in Windows, there's nothing to concern yourself with there. Even if you got into the 10's of billions of data points, Window's memory management will take care of it without any noticeable delay. Consider this - the sample rate is extremely slow when compared to audio or video capture, which can easily be done on any modern computer.

Most aren't very useful to display. The overall curve and trend is what is useful. I'm thinking of ways of doing smart filtering and zooming could be one way to do that.

You really should be capturing everything then allowing the user to choose how to display it. If you truncate the data when it's being captured, you can't get it back afterwards. Perhaps someone wants to zoom in on a specific few seconds of data. No reason to not capture it all and then manipulate the display to show it however you feel.
Just my 2 cents...

[neuroasis]

The performance costs come in the loading of the file from disk into the data grid. Try a long session now it takes tens of seconds for around 30 minutes of data. Not too bad really. I am mainly running on a VM too.

The data would still be on disk, I wouldn't truncate it on capture, just the display altered (maybe) so the graph and grid are more responsive. Like finding min and max points and spanning them, basically display smoothing. Anyway, that is an idea of how to approach it. As you say the size never compares to audio or video files but they are also handled differently whereas the software and codecs are optimized.

I want to implement some kind of 'marker' system, where regions of interest or events can be notated. This would help with navigation and identifying the key events of a session. Metrics are really of interest as well. Like how long did it take you to reach each level.. what is your fastest? average? last month? last year? That is what a database is good for.

[Andy]

Sounds ideal.

I'm curious why it would take so long to load, though. If this becomes an issue it might be worth looking at how data is stored and loaded (but again, only if it becomes a problem).

[neuroasis]

Mainly because it is set to 'just do it' at the moment without any optimization.

[krosskin]

Very strange - why so few software downloads? I would like to hear more reviews

[krosskin]

Yesterday I could not test the program in detail, but I did it today - and I repeat it's great. It's beyond my expectations. How do you think what settings to use for my purposes?

[neuroasis]

The ThoughtStream crowd is usually pretty quiet. Hopefully with renewed releases that will all change.

For ideomotor use you will want a very small window to view the data. I would say do a session at the default settings until you relax and progress down a couple of levels at least. Then you can change the number of samples to 1000 (you can change it while the session is still playing) then start to ask yourself questions. Start with 'show me yes' then 'show me no' then 'show me I don't know'... do this a number of times until you feel like that you have a good idea of how those answers are being represented on the graph. I have been working with this more and I get some results.

What I plan to do is build a gauge type of graph for the purpose of ideomotor communication. Or at least something more specialized for this use. I think using biofeedback this way is very interesting.

[krosskin]

Is it possible to call the signals received TS ideomotor? I think it's close but still not quite ideomotor signals (maybe I'm wrong). By the way, I last week with the help of TS and the official version of program's won the lottery about $ 120 (guessing the number 4 of 6). I still wanted to ask - you have written that you have a version for Android OS if you can put it to the test?

[neuroasis]

It is accurate to say that galvanic skin response signals are ideomotor in the way it is at least described at Wikipedia: http://en.wikipedia.org/wiki/Ideomotor_phenomenon They classify 'tears' as part of the ideomotor phenomenon, so I would say 'sweating' would qualify. Pretty cool that you won some extra money. That helped pay for the ThoughtStream and made it worth your investment.

An Android version will take a while. I have never built a full app for that platform. I want to but I have no idea how long it would take.