Can a brain be modeled as a simplified interaction of different states and their triggers?

Question

I've long been interested in the concept of "states of mind", which influence the perception of the outside world and outlook on past, present and future. They can be thought of as "colored lenses" through which the world is perceived. Each state has a certain trigger. Some example states are:

• Anxiety - world is a terrible place full of danger
• Sexual arousal - brain notices sexual stimuli more and reduces inhibitions
• Creative inspiration - ideas fly and there's a drive to create

Today I've read about an experiment to replicate a part of a rat brain in supercomputer. The following quote jumped at me as rather significant:

The researchers wrote, that the slow synchronous waves of neuronal activity, which have been found in the brain during sleep, were ‘triggered’ during the simulations, suggesting that neural circuits may have the unique ability to able to switch into different ‘modes’ that could explain critical behaviours.

“An analogy would be a computer processer that can reconfigure to focus on certain tasks. The experiments suggest the existence of a spectrum of states, so this raises new types of questions, such as ‘what if you’re stuck in the wrong state?” said Markram.

I read that the project is criticized due to it's complexity, seems like they are working from the bottom up, which makes me ask:

Are there projects out there that attempt to model the brain from the higher levels of abstraction (discrete states and their triggers) down to more detail?

To use a computer analogy - instead of writing very low level binary code, I can take a high level programming library and work with that. Is there research in this direction?

This image is an example of a state machine - a system is modelled in terms of discrete states and their interactions. The author does not concern themselves with interaction of individual neurons, instead with higher level states:

Answer 1

My answer is probably a weird hodgepodge of sometimes poorly explained stuff, but hopefully it's coherent enough :P

For many decades in psychology, we've had a mechanistic stimulus-organism-response understanding of the brain. That is, a stimulus triggers an internal psychological process, which produces some behavioral response. One of the major limitations of this kind of thinking is that it assumes that the mind is at rest until it's stimulated by something in the environment (even if we add a recursive component to it). However, this is fundamentally untrue.

Instead, the brain is a predictive organ (e.g., Clark, 2013). It's loaded with prior knowledge about past experiences, which is constantly used to predict incoming sensory information. If there is a discrepancy between incoming information and past experience (i.e., prediction error), then our knowledge is updated. This is most clearly seen in the literature on vision, which shows how top-down expectations and prior knowledge robustly bias early visual activity, even in the absence of a visual stimulus (Summerfield & de Lange, 2014). The trigger-state model could not easily account for this.

Moreover, dividing up the brain into mental states may not actually represent how the brain functions. The brain isn't really faithful to the kinds of distinctions we make between cognition and emotion, for example (e.g., Barrett, 2009; Pessoa, 2008). And not only that, those higher order states like anxiety and sexual arousal can be represented in the brain in many different ways. Indeed, there is no dedicated mechanism in the brain or body for producing states like anxiety or creative inspiration (e.g., Wager et al., 2015). They're merely concepts for organizing and communicating our experience. And even then, our brain may not really "have" concepts (see Laurence Barsalou's work; but see also Blouw, Solodkin, Thagard, & Eliasmith, 2015), although it may have a conceptual system that allows us to conceptualize (e.g., Barsalou, 2005).

So I guess what I'm getting at is that brain function is nonlinear, spatiotemporally dynamic, and doesn't adhere to folk psychological categories. Thus, it would probably be difficult to understand the brain by attempting to reduce linear trigger-state relationships to the level of networks, regions, and cells. However, maybe (probably) someone has a better informed, more balanced answer than me! :)

Answer 2

I agree with the previous answer/comments that seeking a simplified abstract model of the brain when it is so complex is probably asking too much. We would need to know a lot more about the "states" you are talking about in order to model them, and in reality the set of "triggers" etc is going to be far too long.

However, given your interest and analogy from computing, you may be interested in reading about cognitive modelling architectures, which try to model some tasks/behaviours/processes at an abstract but tractable level. One of the most well known is ACT-R.