You actually have a major issue. You're triggering off noise created by dorking with your leads, and not on anything related to the ecg. This is going to be very difficult to debug without an oscilloscope. What are you using for electrodes? Also, a real circuit diagram for the analog portion would be of huge help.
I suggest your input stage should be a modest gain, maybe 10, with an instrumentation amp, then a high pass filter, then a big gain with an op amp and a low pass filter.
The leads don't belong near your heart. Hold one lead in each hand, and ground your ankle to the same ground you use for vref on the instrumentation amp.
UPDATE:
1) The circuit in the video is not perfectly suitable to amplify biopotentials. You want a DIFFERENTIAL AMPLIFIER, and this amp is single-ended. Interestingly enough, even if this were never intended to be used for biopotentials (there's no suggestion in the vid that it is), its still sort of a textbook example of students using textbook examples and expecting them to work. They're in the right direction, but there's practical aspects of the circuit that will have a ton of problems. The 7 MegaOhm resistors will prove to big noise sources if you try to crank up the gain, and also those big resistors will cause large offset voltages to to input bias currents that will saturate gain stages beefy enough to measure ecg signals, which are about .1-5mV in amplitude (so you'd like a gain of about 500 or so).
2) Even if the circuit were OK to measure biopotentials, it wasn't implemented right in this case.
a) Although the op-amps you used are fine to power with a single power supply, you should have a positive and a negative power supply for the purposes you're using them for here. You can alter the circuit to use a single supply, but that's more work than you want to take on.
b) You can't swap out resistor and capacitor values willy nilly and expect them to work. As luck would have it, the combo you use on the high pass filter stage seems OK, but you haven't provided enough info for me to know about the low pass stage.
To do this right, you should have a modest gain differential stage that won't saturate for an input offset of about 200mV. An instrumentation amp (like your INA128) with a gain of 10 would work if you can supply +/- power. If you need to work with one power supply, then you need an instrumentation amp like the AD623, who's inputs can go below the negative rail, with Vref set at about 2.5V. Then you need a high pass filter to get rid of the amplified offset. After that, a stage with a gain of about 50-100 and a low pass cutoff of about 50Hz. After that, you need some sort of comparator stage to use as a heartbeat detector, but perhaps that's getting a bit too far ahead.
(As an aside, the more modern approach is probably to just have the modest gain first stage and a low-pass filter, then sample it at 24 bit resolution, and do everything after that digitally).
So, bottom line is the circuit you're using won't work, and there's no great way to fix it. If you're going to try this without an oscilloscope to debug, you're going to need to start with an absolutely bulletproof design, and implement it correctly.
I suggest looking at http://www.eng.utah.edu/~jnguyen/ecg/instructions.html (with the schematic at http://www.eng.utah.edu/~jnguyen/ecg/bigsch.gif) for a differential amplifier that works for this purpose.
I that think you need to add a driven right leg circuit at the first for patient safety and protection to bypass the excessive current in case of contact with live wire.
Also for proper reproduction of waveforms on PC make sure to isolate your analog ground and your controller ground.
Also for Ecg, gain of around 10000 is enough.You can split this gain between two stages to avoid saturation and add a 60hz notch filter between the two.
Eg. stage1 amplifier(Instrumentation amp)----notch filter------stage2 amplifier.
Best Answer
The most advisable to obtain an ECG signal is to use conventional electrodes. The alternative to the conventions is to use a textile ECG electrode, but this is experimental. You can put "ecg textile electrodes" in google and you will find many PDF files that contain research. This type of electrode is usually more expensive.
If you want to make a conventional ECG you should use general purpose electrodes, if you want to implement a wearable small size if you can try to work with a textile electrode.
I am currently working with the AD8232 and trying to use only 2 electrodes is to have an ECG signal with a lot of noise. I recommend that you use the configuration of 3 electrodes, the electrode of the right leg is there to reduce the noise of the electrical network. Techniques to reduce noise in your ECG signal: have the electrodes as short as possible, that the electrode cable is shield to GND, that the area of the electrode is as large as possible. Using quality electrodes significantly improves the quality of the signal. Also, applying a pre-skin preparation helps to diminish the impedance of the skin, which allows to reduce the noise and have a signal with greater amplitude.
When working on the design of an ECG you must be very careful, and respect the electrical standards. In the case of using a self-developed electrode, you should request the certification of the electrode you are going to use. I mention this as a little advice.