I think this response has misunderstood the Christe paper. @Martin-マーチン doesn't contradict a 6-center-10-electron bond; in fact, his analysis is entirely consistent with a 6c10e system. The bonding of IF7 requires one 6c10e + one 3c4e bond; together these sum to 8e at the iodine. It also falls directly from group theory, where the equatorial F5 split into A1'+2E1' and the iodine s/px/py transform as A1'+E1'. That leaves 4e (one E1') nonbonding and 6e (A1'+E1') bonding (=10e over 6 centers). The 3c4e bond contributes 2 more bonding electrons (Γ(F2)=A1'+A2'', Γ(pz)=A2'') to complete the octet.