So, looking over the 2014 workshop manual a bit more I believe I may have it figured out. I may actually buy the PAC just to use its cables to make a "middle-man" cable that I can use to monitor each line while it is still plugged in (and ultimately tie RCA jacks onto if this works out) so I can test out this theory.
Here are my findings
First off, here is a link to the 2014 workshop manual: Wiring Diagram - Body Accessories
Everything within the scope of my findings also applies to the 2015/16 since the connections between the TAU and Bose amp appear to be the same on all three.
It seems to be that the head-unit interacts with the TAU to generate the four channel, low-level signals as all of the media input options don't ever connect to the headunit directly (i.e. radio, the bluetooth receiver, the XM receiver, etc), which makes sense given the name of the latter, "Tuner & Amplifier Unit". The headunit seems to simply command the TAU which input to use at any given time and connects to it with data only. The connector in question that seems to carry the low-level signals from the TAU to the Bose amp is labeled as "0920-515A" within the workshop manual, and on schematic "0920-5d" it shows the lines from the one end of this connector labeled "AUDIO UNIT" (the TAU) to the other, labeled "AUDIO AMPLIFIER" (the Bose Amp):
The page is named "CAR-NAVIGATION SYSTEM (Bose TYPE AUDIO)".
The brown and black/red wires on this connector (1G and 1H) are for the AudioPilot mic and can be ignored but the other 8 are shown in the second image where the TAU connects to the Bose amp. Since the user in that other thread I linked mentioned using the grey and blue wires (1J and 1I), I'm assuming that all 8 of the wires in this one connector group are related to the low-level audio signals. I believe that he just took one of the channels since all he was hooking up to it was a sub. Given that he used two wires and that it would be incredibly stupid to connect the reference (or negative) terminals of each channel all together, I think we can safely say that the grey and blue wires are the positive and negative of one channel. It then is reasonable to assume that since there are eight wires total, these annoyingly unlabeled lines in the second image are indeed the 4-channel low-level signals that we've been looking for:
Grey and Blue / 1J and 1I = Channel 1 +/-
Light Green and Pink / 1L and 1K = Channel 2 +/-
Green and Red / 1N and 1M = Channel 3 +/-
White and Black / 1P and 1O = Channel 4 +/-
This is further supported by the fact that these wires are shielded, while many of the other wires connecting to the TAU or Bose amp that are clearly labeled as something data related, are not.
Now I have no idea which of each pair is the positive or negative, and no clue of all the channels which ones are front left, front right, rear left, and rear right, but I am fairly confident these are the wires I want to tap into right here. I believe my scope has a capture function so I can always load an output sample into MATLAB, amplify it, and play it to confirm I got the right wires as I should literally hear whatever source material I was playing at the time Also, I can do this to help figure out which of the wires are positive (although it shouldn't matter as long as they are all hooked up in phase).
Figuring out the channels would just be a matter of disconnecting only one set at a time and listening to which portion of the car has the audio drop out.
Now, if you look at the rest of the connections going into the Bose Amp there are the speakers and then a number of data lines; however, most of the data seems to be unrelated to Audio so I am assuming that Mazda just used the Bose AMP as a junction box for other electronics in the car (for example the push start button is fed through the AMP <img src="http://forum.mazda6club.com/images/Mazda6Club_2014/smilies/tango_face_smile_big.png" border="0" alt="" title="Laugh" class="inlineimg" />). So, I think there is a good chance that everything audio related goes through those four channels and that I won't lose anything by splicing in at that point. Either way, I can easily test if anything like chimes or the like are sent to the amp from elsewhere by monitoring the 4 channels while causing those sounds to play, or even disconnecting those 4 channel lines and seeing if I can get any sound to be produced at all (since if it did then the amp would be getting that sound from somewhere other than these 8 wires).
At some point I am going to sniff these wires by accessing the cable at the amp side under the seat and I'll see what I can figure out.