Nantonos said in an earlier post:

There is a related question, did Epona start out as a 'divine mare' who became progressively anthropomorphised, first to a woman sitting on a horse and then to a woman seated on a throne attended by horses (this model assumes the Gaulish sidesaddle type is earlier than the Imperial throned type) or did Epona start out as a more generic mother goddess of fertility and abundance, gradually becoming more specialised onto horses (this is argued for by the cornucopiae, baskets or patera of food, nursing mares, and so forth).

This is something I've often wondered. From the reading I've done, it makes sense that Epona would have been horse-related initially.

There are three clusters of artifacts represented on the distribution map, representing areas of greatest concentration. Could they also define where her cult has been around the longest?

One clusters falls within the area of the Aedui, who were known horse-breeders.[1] Do these larger clusters mean that she was honored there longer? Or just that it was more prevalent?

Lastly, if there were representations of Epona as just a mare, without humans, without inscription - how would we know and how would we distinguish this type from ay other votive horses which could have been offered to various deities both Romano-Celtic (eg 'the Celtic Mars') and indeed Roman (cf for example the sacrifice of the Equus October).

As you said, would we know a votive offering if we saw it? Maybe the horse as a representation of Epona was the standard image for her before Roman influence changed to a more Roman-approved icon of her.[2]

[1] I've read this some where, and have to locate the reference. I think it's from Gallic Wars, and also from Ann Hyland's Training the Roman Cavalry.

[2] Greg Woolf, Becoming Roman: The Origins of Provincial Civilization in Gaul. Cambridge Press. 1998