As you're using ffe and need a similar thing to me, I'll give you what I use and try and explain it.
The important parts are highlighted in yellow. As you can see, all the settings boxes are blank apart from "extra parameters". This is because a) we're not actually converting any of the data, just extracting it and b) "extra parameters" can contain any parameters ffmpeg can use.
Important stuff
Input: The input file. In our case, the mp4 created by OBS.
Output: The output file. For this, it needs to have a .wav extension, so that ffmpeg knows we're creating a WAV file.
Extra Parameters:
Bear with me on this one.
-filter_complex tells ffmpeg we're going to use a complex filter, which will be described by the text in quotes following it.
My mp4 files contain all four of the audio tracks OBS can create. In this case, in ffmpeg-speak they're contained in streams [0:1] to [0:4]. ([0:0] contains the video stream, which we don't care about right now)
Resolve will happily make use of the first audio track in the mp4 file, it just can't see the other three, so I need to extract [0:2] through [0:4]. I use the "amerge" filter to merge the stereo tracks together into a single multichannel stream, so I tell it that the amerge filter needs three inputs and the result should be placed in a stream I've labelled "aout".
Finally, the '-map "[aout]"' parameter tells ffmpeg that what we want to map into the output file is the "aout" stream we've just created.
The result is that we take the last three stereo audio tracks from our OBS mp4, merge them into a single six-channel audio stream, and then write that out to a WAV file. I've found this works nicely in Resolve, where you can tell it that the six-channel audio file you've imported is in fact three sets of two-channel audio.
This is almost certainly information overload again, but hopefully some of it is comprehensible.
