SimCLR ResNet-18 checkpoints trained on Visual Experience Dataset (VEDB) imagery under four conditions, isolating central vs. peripheral visual field