Even emotions (or, according to the declaration, their “neural substrates”) are not dependent on an animal having particular brain structures, such as our cortex, after all. In fact, many other neural regions are activated when we emote and “are also critically important for generating emotional behaviors in animals,” the scientists noted.
As I was saying before, there are a few definitions of emotion. One requires only basal reactions such as fear and anger. The other such as love, admiration, etc. require a higher-order mammal to use its cortex to transpose its basal reactions into a complex emotion. Your betta doesn't "get mad at you" like a friend does when you don't share your lunch with them, they show pure aggression towards an intruder. They are not "happy" when you come into the room with food, they recognize you as the caregiver and bringer of food, so they come swimming up to the front of the tank to guarantee their chance at food. When instead you do not actually feed them, they don't sulk, they just resume their business.
As for all the things regarding "oh my betta recognizes its favourite food bottle" "oh my betta doesn't like my boyfriend". It is simply because the betta has learned -- Through simple Pavlovian learning, that you are the caregiver, and the food bottle is representative of food, or the conditioned stimulus.
I, as a scientist myself, will stand with koimaiden. One or two articles from a "science blog" don't really mean too much, IMO.