I’m wondering if you’re reverse biasing the optoisolators (i.e. + is connected to - and vice versa). You might want to try measuring resistance with the switch ‘on’ with the DMM probes reversed.
I do not suggest connecting a high-ish voltage supply across the switches until you’ve worked out what’s going on.
My thoughts exactly. These are diodes essentially… they have a specific polarization. Forcing voltage backwards through them will result in this kind of behavior, and may damage the isolators in the process… though when that happens typically you get 0v and magic smoke.