Model

I think schizophrenia can be understood in some sense as modeling the brain like a statistical predictor.

First, let us consider how statistical models typically work. A numerical model gives us a number that we use to make a decision. For example, if a model is trying to predict whether an image contains a dog or does not contain a dog, we might get a large number from a model if the model has evidence that a dog present, and the model might give a small number if there is not evidence that a dog present. This number is a reference or in statistics is called a threshold.

There is a tradeoff in decision making called sensitivity and specificity. If we require the model to give a very large number before we decide there is a dog, we may lose sensitivity. Because we may not sense a dog in images where there is a dog present. That same large number requirement may also mean that we gain specificity. That means, we will be very specific in choosing images that we decide have a dog. On the other hand, we may only require that the model gives a very low number to decide that a dog is in the image. In this case we gain sensitivity, because almost all images that have dogs, we will decide that there is a dog, and say our model sensed a dog. We lose specificity however, because we are not being specific enough, and sometimes an image that does not have a dog, we will decide there is a dog, because even a low number from the model will be considered to tell us that.

This seems similar to schizophrenia. There may be a region of the brain, where the neurons don't require much input to fire. The “threshold” of their activation is low. This would mean that weak signals are able to activate a region of the brain. This is experienced as follows. I might walk into a room, and there are several conversations going on, and I have trouble focusing on my work, because it is hard to ignore any of those conversations. Or perhaps, there is no one in the room, and it is quiet. But I can remember an interaction with a person from earlier with so much emphasis that I hear them talking in my mind. Or, maybe I get enough composure to start my work. I might read a passage in a book, and it seems like the passage is a message from God to me. Because the signal from that passage seems very strong. Or I might make a connection between two ideas in a book, where the connection is coming from something weak, like both sentences begin with the letter ‘b’, and then weak evidence would seem to give a strong connection, because my brain fires too much given very non important signals.

The problem is, there is not a test that can say a person is too sensitive and not specific enough. However, if you have this condition, you will find that the people in your life know there is something wrong. You won't be able to ever prove it to yourself, but everyone can tell you, because they have stronger specificity relative to sensitivity, they will be able to tell you that you have schizophrenia.

This also helps explain why a psychedelic or marijuana can trigger schizophrenia. If parts of the brain become more connected (as some drugs increase signal flow at neuroreceptor sites) they may deliver more signals to parts of the brain with a low threshold for firing, causing the brain to notice things that are not significant or even not present in the environment, but just within the consciousness.

Subsequent Treatment Idea

I also think a statistical idea can be used in treating schizophrenia. Often, we won't know if something is true, but we will have a level of confidence about whether we think something is true (or worthy of attention). So, I could do the following practice. Instead of trying to decide if what I am experiencing is real or a delusion, I can get in the habit of asking myself, how likely is what I believe to be true? This soft approach may be a part of the healing process and teaching my brain to be careful in the gray areas of life.