Uni-directional cables aren't about "which way the electrons flow". As posted, they wiggle back and forth when AC is being conducted and we're not generating DC. The issue has to do with the shielding and these aren't standard coaxial cable, like the original RCA or 1/4" stuff. Anyone has done car audio/radio/CB/ham radio work, they know that "star grounding" is a concept that should be maintained. It means that all shielding and grounds go to one electrical ground point. This isn't necessarily a single physical point- it means that the ground points need to be at the same electrical potential, meaning that no voltage flows between those points because there's no resistance between the points. Voltage is potential energy and is often expressed as "they have a potential difference of xx Volts AC or Volts DC".
Shielding and grounding are not necessarily the same thing, either. In a coaxial cable, the shield acts as the ground but if the instrument and amplifier are each grounded, a single wire between them would do a good job of conducting the signal without noise as long as there's no resistance between the grounding points.
Twisted pair rejects "common mode" noise and if there's no shield, it still works well in environments where RF and EMI are present. Cat5e and for that matter, regular phone lines are twisted pair for this reason. The wire is cheap, it rejects noise and it works. Mic and distributed audio cabling has two conductors, twisted, with a braided or foil shield wrapped around it, sometimes with a multi-stranded wire called a 'drain' or 'sink'. This wire is in contact with the foil shield and is soldered at the head end, so the noise can be eliminated. If it's a braided shield, that's used instead. Where lowZ balanced cabling/systems have an advantage is in the fact that the chassis is grounded but the audio signal is carried on two conductors, one of which is "floated above ground"- floated meaning there's a resistance between the audio negative and the ground, so any noise that affects the grounded chassis won't be injected into the audio cabling.
The ground lift on an effect or amplifier lifts the connection from Pin1 of an XLR plug/jack so the system's common piece of equipment (preamp or mixer) can send the RF noise to ground. The disconnected end of the shield acts as an antenna for the noise and the grounded end dumps the noise to Earth. That's why unidirectional cables need to be in that orientation, not because of any device in the cable, like a cap or diode (although I have read about cables having a diode for DC voltage protection).
However, snake oil abounds and claims of "it made the sound open up", "I heard things that weren't there before" and "I couldn't believe the difference" are usually due to using a new cable in place of one that used cheap/inferior materials/construction or one that was damaged/repaired badly. Some of the claims and descriptions used by high-end audio cable manufacturers and their victims are absurd- "chocolatey"? "Eliminates the need to bias the dielectric"? "Eliminates the need for cable break-in"? W T H? Not a single electrical/communications engineer I have spoken to has heard of this stuff and at least one of them works with nuclear hardened circuits.
