Non-locality of memory has been known for decades. It is one of the reasons I believe artificial intelligence is beyond my lifetime, because no one knows how it works.
Indeed, non-locality of memory does not favor Crick's "Astonishing Hypothesis." I hadn't heard of the salamander experiments, so they were particularly interesting.
None of this argues for any non-physical property of the mind. Just because memories are spread out doesn't mean they aren't embodied.
Likewise, none of this argues against a non-physical property of the mind.
Crick studied vision, which is highly localized and has structures that have an organized physical pattern. He studied vision becuse it is "easy".
As you study the brain structures that we have in common with other species, you see more specific localizations. But speech and language have well studied localizations. Damage to specific sites produce predictable deficits. Dficites of production (speech) can sometimes be trained around, but deficits of thought seldom recover. The intersting thing about people with thought deficts is that they sometimes don't know that anything is missing.