Mechanical damage to parts of the brain can cause personality changes. Multiple neurological syndromes are directly linked to macroscopic brain changes. This includes schizophrenia, traumatic brain injuries dramatically increase its risk.
We also know that schizophrenia almost never happens along with epilepsy. They are almost mutually exclusive.
It's fair to say there is no single macroscopic trigger for schizophrenia, but that's it.
Schizophrenia is directly linked to dopamine. So is ADHD, but in the opposite way (not enough dopamine). Both were demonstrated on fMRI.
Psychosis is a physical issue in the brain (too much dopamine making the brain fire when it shouldn't) that's solved by administering dopamine antagonists (= antipsychotics).
https://sci-hub.ru/10.1001/jamapsychiatry.2020.1941
The citation is:
Weinberger, D. R., & Radulescu, E. (2020). Structural Magnetic Resonance Imaging All Over Again. JAMA Psychiatry. DOI:10.1001/jamapsychiatry.2020.1941
The article is a follow-up editorial on an older (2016), more substantive critique of MRI methods in psychiatry, which is also worth digesting slowly:
https://ajp.psychiatryonline.org/doi/10.1176/appi.ajp.2015.1...
The citation is:
Weinberger DR, Radulescu E. Finding the elusive psychiatric “lesion” with 21st-century neuroanatomy: a note of caution. Am J Psychiatry. 2016;173(1):27-33.
The situation is about as bad as it can be: there are so many potential sources of confounding in MRI studies that they simply cannot be trusted to say anything about the basic cellular elements of brain tissue in either healthy or diseased patients. All they can validly talk about is themselves; in other words "rather than referring to differences as evidence of brain structural abnormalities, they should be called differences on MRI measurements."
This is without getting into other hugely problematic aspects of MRI studies in mental health: tiny samples; failure to replicate; and most interesting to me, the fact that experimental subjects (i.e. the people in the study "with ADHD" or schizophrenia or whatever) are not screened for psychiatric medication use, which, because medications can shrink brain tissue, can make it look like the condition is affecting brain structure when in fact the drug is. (So another source of confounding, then.)
It's like trying to diagnose and treat a political crisis by observing telephone connections. We see stark changes in telephone use in a crisis. If we meddle with the telephone exchanges we might improve or harm various types of event. It's not really about the telephone calls though, we are just hacking at the messaging of an external event we are ignorant of.
A conspiracy theory induced panic might be superficially "solved" but cutting all the the phone lines. We have not identified or solved the cause though and people don't find life without telephones (dopamine) worth living so they plug them back in (quit their anti-psychotics).
But currently we're unable to image live people at that resolution, not even dead people yet.