In game 7 Spassky already played 1.e4 and almost lost in the Sicilian "Poisoned Pawn" line, but there he - or his team - had in mind a concrete improvement, and when he got another chance of playing it in game 11 he won - the only game in the match where Spassky managed to systematically outplay Fischer. So it makes sense that he decided to stick with 1.e4 for the rest of the match (though since, from that point on, Fischer stopped repeating lines but kept throwing at him something different every game, he could never be sure about what opening to expect). The results were mixed - he lost games 13 & 21 (in both of which the opening didn't go so well for him), but managed to put Fischer under considerable pressure in games 15, 17 & 19. Any other player in the world than Fischer, at that time, would almost certainly have lost to Spassky at least one of the middlegame positions that arose from the openings of these games.

So you say, and you may be right, but it's only fair to note that Spassky himself claimed he came up with Nb1 at the board. Gligoric said of this that "nobody believes him", but I think I do, mainly because the move really isn't that great, and only worked because Fischer had an off day.

Yeah, I forgot about that part in Gligoric's book (also that Spassky spent 30 minutes on Nb1) - I was going by "Russians vs. Fischer", where Bondarevsky is quoted as attributing the move to team home analysis. Btw, even if the move doesn't objectively deserve the "!!" it gets in some commentaries, I still think it was a very good practical idea, because it posed Fischer some serious problems to solve at the board, in a line where he usually outprepared his opponents and hardly faced any problems. At any rate, if Spassky himself came up with the move at the board it actually fits even better with the more general point I was trying to make about his decision to stick with 1.e4 from that game on - in the sense that he might have been encouraged by the feeling that he managed at last to outplay Fischer "on his own".

Either way, I can see why Spassky played it. 3 points down, he needed to try the sharpest line he could find, and this was the sharpest one in Fischer's repertoire.

Maybe they suggested it to him beforehand (Hey, maybe Nb1 is worth a try), but he spent a long time at the board convincing himself it was playable. Some of the players who praised it the most didn't play it themselves when they had the chance a year or so later, so I don't know if I'd call it an "improvement". It certainly didn't deserve the shower of exclamation points annotators heaped on it at the time.

Other lines are worse for white.

Fischer knew the rules, and he knew when he could break them.

At this point in the match, Spassky's brain was still a sieve. I wouldn't be surprised if he did review it with Bondarevsky, but completely forgot about it when he was at the board.

Another kibitzer mentioned 7.Qd3, which was played against me in Burnham-A Shaw, 2nd Monadnock Marathon 1979. There followed 7....Nxe4 8.Qxe4 Bxc3+ and my tired brain doesn't remember the rest of the game, except that my opponent somehow grovelled a draw from this mess.

That is the mark of a champion-give as good as you get.

I know in Spasskys case that it's well known that he forgot aloth of the prep Geller was providing.