Superposition of distinct psychotic and affective axes explains affective psychosis: evidence from cross-diagnostic connectomics.
Whether affective and non-affective psychoses reflect a severity continuum or distinct neurobiological mechanisms remains unresolved. We compared whole-brain and subnetwork efficiency across this spectrum.
Resting-state functional magnetic resonance imaging data were analyzed from 78 patients with affective psychosis, 197 with non-affective psychosis, 164 with non-psychotic primary affective disorders, and 165 healthy controls. Graph analysis quantified whole-brain and subnetwork global and local efficiency. We tested symptom and cognitive associations and performed sensitivity analyses stratifying schizophrenia by depressive-symptom status and affective psychosis by diagnosis.
Non-affective psychosis showed lower whole-brain efficiency than controls and both affective groups, and lower sensory/somatomotor efficiency than controls. Conversely, whole-brain efficiency was preserved in both affective groups; affective psychosis showed higher cingulo-opercular, dorsal attention, and salience network efficiency, whereas the non-psychotic affective group showed higher dorsal attention and salience network efficiency. Sensitivity analyses identified shared increases in dorsal attention global efficiency and salience network local efficiency across psychotic major depression and psychotic bipolar disorder, and reduced sensory/somatomotor efficiency in schizophrenia regardless of depressive-symptom status. After false discovery rate correction, higher default mode network efficiency was associated with Positive Symptoms, while higher whole-brain, cingulo-opercular, salience, and sensory/somatomotor efficiency was associated with Activation symptoms; no cognitive associations survived correction.
These opposing, subgroup-robust patterns favor a dichotomous-mechanism framework comprising an affective axis of increased control-network efficiency and a psychosis axis of whole-brain and sensorimotor inefficiency. Affective psychosis predominantly expressed the affective axis with a partial shift toward the psychosis-axis profile.
Resting-state functional magnetic resonance imaging data were analyzed from 78 patients with affective psychosis, 197 with non-affective psychosis, 164 with non-psychotic primary affective disorders, and 165 healthy controls. Graph analysis quantified whole-brain and subnetwork global and local efficiency. We tested symptom and cognitive associations and performed sensitivity analyses stratifying schizophrenia by depressive-symptom status and affective psychosis by diagnosis.
Non-affective psychosis showed lower whole-brain efficiency than controls and both affective groups, and lower sensory/somatomotor efficiency than controls. Conversely, whole-brain efficiency was preserved in both affective groups; affective psychosis showed higher cingulo-opercular, dorsal attention, and salience network efficiency, whereas the non-psychotic affective group showed higher dorsal attention and salience network efficiency. Sensitivity analyses identified shared increases in dorsal attention global efficiency and salience network local efficiency across psychotic major depression and psychotic bipolar disorder, and reduced sensory/somatomotor efficiency in schizophrenia regardless of depressive-symptom status. After false discovery rate correction, higher default mode network efficiency was associated with Positive Symptoms, while higher whole-brain, cingulo-opercular, salience, and sensory/somatomotor efficiency was associated with Activation symptoms; no cognitive associations survived correction.
These opposing, subgroup-robust patterns favor a dichotomous-mechanism framework comprising an affective axis of increased control-network efficiency and a psychosis axis of whole-brain and sensorimotor inefficiency. Affective psychosis predominantly expressed the affective axis with a partial shift toward the psychosis-axis profile.
Authors
Wang Wang, Wang Wang, Zhong Zhong, Zhang Zhang, Yang Yang, Liu Liu, Yang Yang, Palaniyappan Palaniyappan
View on Pubmed