Despite the high clinical role of delusions as a transnosological psychopathological phenomenon, the number of experimental studies on the different types of delusions across schizophrenia spectrum is still relatively small, and their results are somehow inconsistent. We aimed to understand the current state of knowledge regarding the structural and functional brain alterations in delusions to determine whether particular types of delusions are associated with specific brain changes and to identify common alterations underlying the formation and persistence of delusions regardless of their content. For this systematic review, we followed PRISMA guidelines to search in PubMed for English papers published between 1953 and September 30, 2023. The initial inclusion criteria for screening purposes were articles that investigated delusions or subclinical delusional beliefs in schizophrenia spectrum disorders, high clinical or genetic risk for schizophrenia using fMRI, sMRI or/and dwMRI methods. Exclusion criteria during the screening phase were articles that investigated lesion-induced or substance-induced delusions, delusions in Alzheimer's disease and other neurocognitive disorders, single case studies and non-human studies. The publication metadata were uploaded to the web-tool for working on systematic reviews, Rayyan. For each of the studies, a table was filled out with detailed information. We found 1752 records, of which 95 full-text documents were reviewed and included in the current paper. Both nonspecific and particular types of delusions were associated with widespread structural and functional alterations. The most prominent areas affected across all types of delusions were the superior temporal cortex (predominantly left language processing areas), anterior cingulate/medial prefrontal cortex and insula. The most reproducible findings in paranoia may be alterations in the functioning of the amygdala and its interactions with other regions. Somatic delusions and delusional infestation were mostly characterized by alterations in the insula and thalamus. The data are ambiguous; however, in general the predictive processing framework seems to be the most widely accepted approach to explaining different types of delusions. Aberrant prediction errors signaling during processing of social, self-generated and sensory information may lead to inaccuracies in assessing the intentions of others, self-relevancy of ambiguous stimuli, misattribution of self-generated actions and unusual sensations, which could provoke delusional ideation with persecutory, reference, control and somatic content correspondingly. However, currently available data are still insufficient to draw conclusions about the specific biological mechanisms of predictive coding account of delusions. Thus, further studies exploring more homogeneous groups and interaction of diagnoses by types of delusions are needed. There are also some limitations in this review. Studies that investigate delusions induced by lesions, substance abuse or neurodegeneration and studies using modalities other than fMRI, sMRI or dwMRI were not included in the review. Due to the relatively small number of publications, we systematized them based on a certain type of delusions, while the results could also be affected by the diagnosis of patients, the presence and type of therapy, illness duration etc.