If this is a side-affect of just the COV2 spike protein bonding (as the article seems to suggest), then we'd better hope that the mRNA from the various Covid vaccines doesn't make it to the brain after injection.
From the last paragraph of the paper:
“The current versions of the Moderna, Pfizer-BioNTech, and Johnson & Johnson SARS-CoV-2 vaccines encode the full-length spike S protein with two mutations (spike S-2P) that stabilize the prefusion conformation and inactivate its fusogenicity (39, 64, 67, 68). We used this same mutant form of spike S-2P as a negative control, demonstrating the complete lack of fusogenicity when two consecutive prolines were added at positions 986 and 987. However, our findings demonstrate that it will be critical to consider the fusogenic potential when designing any future vaccines in which viral fusogens are to be expressed in mammalian cells.”
It might require the full spike protein, not just the receptor binding domain which is what the mRNA vaccines include.
Also, I'm not sure how the coronavirus proper gets into the brain, but it seems likely it's through infecting some cell that crosses over, rather than directly getting past the blood-brain barrier. That's less relevant for the mRNA vaccines because A) mRNA doesn't replicate and B) the spike proteins that get produced mostly get chopped up before they can leave the cells (they're not budding off into viral envelopes after all.)
Also, nearly everyone in the world has had the real thing now - that's a lot more spike than the vaccines contain, orders of magnitude more.
Both the Pfizer-BioNTech and Moderna vaccines use a full length spike, look at the WHO definitions for elasomeran (Moderna) and tozinameran (Pfizer). For example, the Pfizer-BioNTech vaccine is described as "messenger RNA (mRNA) 5'[prime]-capped, encoding a full-length, codon-optimised pre-fusion stabilised[sic] conformation variant (K986P and V987P) of the SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2, GenBank: MN908947.3) spike (S) glycoprotein, flanked by 5'[prime] and 3'[prime] untranslated regions and a 3'[prime] poly(A) tail; contains N1-methylpseudouridine instead of uridine (all-U>[methylpseudouridine symbol]). immunological agent for active immunization (anti-SARS-CoV-2)". It's clearly full length, Pfizer abandoned their RBD-only candidate.