Hi at all,
I started to experiment with the crp toolbox (thanks for this nice toolbox btw) and want to apply RP to the analysis of fMRI data.
I have two time series. The first consists of 1's and 0's and represents a stimulus on/offset and the second is a continuous time series consisting of fMRI data. Since those two time series can not be considered part of the same system, JRP should be the choice here I figured. I attached a picture of my JRP for which I used fan at 0.1.
I'm a little bit confused about how to interpret the bottom part (<50) where there seems to be a higher amount of recurrence points.
More over, I'm not sure if it is even possible to JRP those two time series, since one of them is binary. Has anyone done something like this before and is it a valid approach?
I appreciate any helpful commments and opinions on this matter
David
JRP between binary and continuous time series

 Junior
 Posts: 6
 Joined: Wed Oct 8, 2014 13:00
 Affiliation (Univ., Inst., Dept.): University of Muenster
 Location: Muenster, Germany
 Research field: fMRI, EEG research, restingstate
JRP between binary and continuous time series
 Attachments

 jrp.png (72.03 KiB) Viewed 4822 times
 Norbert
 Expert
 Posts: 173
 Joined: Wed Jan 4, 2006 11:03
 Affiliation (Univ., Inst., Dept.): Potsdam Institute for Climate Impact Research, Germany
 Location: Potsdam, Germany
 Location: Potsdam Institute for Climate Impact Research, Germany
Re: JRP between binary and continuous time series
Hi David,
thanks for this interesting example that is a quite challenging one, in my opinion.
First, you are right to try to use the JRP. However, in my opinion the binary data requires an alternative recurrence criterion. As a start, you can try to do it as you started. The dense region at the bottom comes from your recurrence criterion (FAN). You can use the criterion to fix the RR (option RR in the GUI) that is an alternative to FAN but ensures the symmetric plot.
As a next step I suggest to implement a similarity measure that can work with spike trains, i.e., point process based RP or event synchronization (BTW, I suggest to calculate the JRP by not using the function jrp but by simply multiplying the RPs):
Norbert
thanks for this interesting example that is a quite challenging one, in my opinion.
First, you are right to try to use the JRP. However, in my opinion the binary data requires an alternative recurrence criterion. As a start, you can try to do it as you started. The dense region at the bottom comes from your recurrence criterion (FAN). You can use the criterion to fix the RR (option RR in the GUI) that is an alternative to FAN but ensures the symmetric plot.
As a next step I suggest to implement a similarity measure that can work with spike trains, i.e., point process based RP or event synchronization (BTW, I suggest to calculate the JRP by not using the function jrp but by simply multiplying the RPs):
 S. Suzuki, Y. Hirata, K. Aihara: Definition of distance for marked point process data and its application to recurrence plotbased analysis of exchange tick data of foreign currencies, International Journal of Bifurcation and Chaos, 20(11), 36993708p. (2010).
 R. Quian Quiroga, T. Kreuz, P. Grassberger: Event synchronization: A simple and fast method to measure synchronicity and time delay patterns, Physical Review E, 66, 041904p. (2002).
Norbert

 Junior
 Posts: 6
 Joined: Wed Oct 8, 2014 13:00
 Affiliation (Univ., Inst., Dept.): University of Muenster
 Location: Muenster, Germany
 Research field: fMRI, EEG research, restingstate
Re: JRP between binary and continuous time series
Thank you Norbert . Did not look into this matter any further, but as soon as I do, I will post an update into this thread.