Post History

Trace width is a silly thing to worry about at such low currents and high impedances.

You've got 1 MΩ resistances in the signal path, and the opamp input is presumably much higher impedance than that.  You didn't say what the upper limit of your signal frequency is, but lets say 1 kHz since it's from the body.  At 1 kHz, the 1 µF caps in your signal path have an impedance magnitude of 160 Ω.  A few mΩ are not relevant.

You should really be focusing on dealing with the common mode signal.  That's where your noise will come from.  One technique for electrically floating bodies is to drive a spot on the body far from the measurement to try to null out the common mode signal.  For heart signals, this is often done by driving the right leg in a feedback loop inversely to the common mode voltage.  The reason for the right leg is that it's far away from the heart, so signals induced there will be common mode at the measurement site.

Another possible technique is to float the entire signal front end circuit on the common mode voltage.  Then have it pass on digital measurement via opto-couplers to the rest of the device.  That might also help with safety certification.