Assuming already that emotion exists in games and that it's a good thing, of course. Having my baby led to a train of thought that I'd like to share: finally, after twenty hours of labor (but only two hours and ten minutes of pushing, so it's not as bad as it sounds), when we finally had our baby, and hit that emotional moment that parents on Birth Day claim they're so proud of, we didn't care. We went through the motions, taking the photos and cutting the cord and holding our little blood-spattered extra-terrestrial looking baby-thing and generally feigning happiness, but really we were just exhausted. Two days later, we were able to look back at the birth and how nice the doctors and nurses were and how Cathy was able to push a poorly-aligned baby out through sheer force of will without having to resort to a C-section and actually feel some emotion.
My point? In one of my psych classes way back when I found out that when you feel emotion, a chemical gets released. This chemical is in limited supply. Which means you can only feel emotion for so long before you run out. Which means, if a game is about making you feel a constant stream of intense emotion - and most survival horror games fall into this category - it only works if the player plays the game in small doses, or if the game itself consciously varies the intensity of the experience.
So: vary the intensity of the experience, right? But this comes into conflict with another good guideline for game design, which is to focus on the game's strength and trim away everything else. For obvious reasons: less is more, focus is good, lower quality stuff will dilute overall quality, blah blah blah.
The Zeldas vary intensity excellently: you have your intense dungeons where you're feeling a lot of fear, frustration, engagement--and often the joy of discovery; then you have the calmer, sometimes idyllic, surface level where you can talk to people and shop and so on. I always thought the shopping and talking portion of games was fairly boring and wanted to get back to the hacking and slashing and puzzle solving - I always thought they should make a Zelda that was all dungeons. I now think that I was wrong. If you could trust the player to quit playing the game after they were 'drained', and come back refreshed, it would be a good thing - but if your players are like me, you can count on them to binge until the game's over, even if they're not enjoying it as much as they could if they spread it out. So you do what Zelda does: you include low-intensity, low-emotion stuff between the high-intensity areas so the player can recharge.
Now, it doesn't necessarily have to be a hub or an overworld, (although hubs and overworlds have so many other benefits you need to ask yourself why *not* have a hub or overworld rather than why have one). If you're doing a fairly linear funhouse walkthrough sort of game you can vary the intensity level-by-level: explore, then fight, explore, then fight. Prince of Persia is a pretty good example of that. (And so are movies: good horror movies don't try to scare the crap out of you every second; they have sequences where they let you catch your breath and recover before they start building the tension again.)
I was asked not long ago if and how I use my psych background in games. Well, this is an example of me using it. Yay! My four years of college weren't totally wasted!