Post-stroke psychosis is uncommon and unexplained. Only 15 of 360 stroke patients suffered from delusions in one large case series . Such delusions usually involve those of bodily perception, and are generally transitory [7, 8, 14]. Previous studies have not used lesion overlap analysis to determine a common area of damage in patients with psychosis following stroke. Here we demonstrate the most discrete area of involvement to date: right IFG and underlying white matter was common to all three patients with persistent post-stroke psychosis. Frontal lobe involvement has been documented previously, but only in the context of widespread damage to other areas [7, 8, 14]. Our findings are consistent with observations of patients with degenerative disease, where psychotic symptoms are associated with right-sided atrophy . Delusional misidentification, manifested by two of our patients, is also associated with hypoperfusion of the inferior frontal gyri in patients with degenerative disease .
Functional imaging in psychotic patients has demonstrated an increase in right inferior frontal activity in association with auditory hallucinations . Impaired functional connectivity, especially of right inferior frontal lobe, appears to be a key factor in the pathogenesis of psychosis [1, 18] and superior longitudinal fasciculus involvement may be particularly important [6, 16]. As a corollary, our findings suggest that focal lesions to right IFG and neighbouring white matter might cause impaired anatomical connectivity in the three patients described, possibly via potentiating symptom-associated neural overactivity within the disconnected (and thus newly disinhibited) cortex. However, because all three patients had damage to right IFG as well as associated white matter, we are unable to differentiate which of these relates more closely to the generation of psychosis, or if both are required.
Why is post-stroke psychosis so rare? Structural and functional imaging studies of delusions have suggested that more than one focus of abnormality has to be present for delusional ideas to become severe enough to clinically manifest. A recent meta-analysis examined the neuroanatomical correlates of vulnerability to psychosis by comparing structural MRI findings between subjects at enhanced risk of developing psychosis and controls . Whilst a number of areas differed between these two groups, there was a specific gray matter volume reduction in right IFG only in high risk subjects who subsequently developed a psychotic episode. Functional imaging studies of schizophrenia have found that prediction error signalling (denoting the mismatch between what is expected and what is experienced) is attenuated in right frontal areas in delusional patients, and the degree of attenuation correlates with the severity of their unusual thoughts . These authors propose that inappropriate prediction error signalling associated with abnormal right frontal cortical function leads to delusions via inaccurate representations of the environment. More recently they have found that healthy people with non-clinical schizotypal beliefs have the same abnormalities in prediction error signalling in right IFG as psychotic patients, demonstrating that premorbid vulnerabilities exist in healthy people . These structural and functional imaging studies support the view that psychotic symptoms exist as part of a continuum, with schizotypy occurring as an attenuated form of clinical psychosis . Furthermore they also suggest that multiple ‘hits’ are required for psychosis to manifest clinically, and that right IFG dysfunction is a critical factor. In our study, we suggest that right IFG damage due to stroke was the second hit which triggered psychosis, the first hit being their untreated mental health problems. Although cerebral atrophy has been highlighted as a potential factor in generating post-stroke psychotic symptoms [8, 14], none of our patients had atrophy and, given the high prevalence of moderate atrophy amongst stroke patients, a much higher incidence of similar symptoms might be expected if this were contributory. We suggest that these three individuals may have had a behavioral susceptibility to psychosis, related to their pre-existing mental health problems. This would explain the relatively low incidence of post-stroke psychosis, since the large population of patients who suffer from stroke in the same vascular territory, but have no significant psychiatric history, do not develop similar symptoms.
Our findings are preliminary; the number of cases is small, due to the rarity of the condition, and premorbid schizotypy was not looked for specifically. In future large scale studies of post-stroke psychosis, it would be informative to screen for schizotypal personality as well as premorbid psychiatric disease. Ours and previous studies would predict that individuals with schizotypal features would be at a substantially higher risk of developing post-stroke psychosis if right IFG is involved. Additional cases would also be needed to determine other risk factors, behavioural or structural, that predispose to psychosis following stroke.
To summarise, our study provides the most precise lesion data to date corroborating existing structural and functional studies that right IFG damage is a critical component in the development of clinical psychosis in at-risk individuals. Moreover, from the practical point of view, in a clinical setting it might be worth making the extra effort to clarify an individual’s previous psychiatric history when they appear to be manifesting delusions in the context of hemispheric stroke.