Figure 1 appears to use the Signal Processing Toolbox envelope (link) function to isolate the bursts. I would then use findpeaks (link) to define the peaks of the envelope output. That data will let you define the IBI data.
However, the authors apparently filter their data to get the envelopes. If you want to follow their lead in that respect, and you have R2018a or later, use the bandpass (link) function to design the filter and filter your signal. (Note that if you use the ‘d’ digital filter output to filter your signals, use filtfilt to filter them, not filter as in the documentation. That only works with linear-phase FIR filters, and bandpass designs an elliptical IIR filter by default.) There are other ways to design filters if you do not have R2018a or later. They simply require more effort.
That is the approach I would use.
