Neurobiology of Insight in Schizophrenia: Interrelationships with Symptom Domains and Genetic Influences

Thumbnail Image




Xavier, Rose Mary


Vorderstrasse, Allison

Journal Title

Journal ISSN

Volume Title

Repository Usage Stats



Problem: Insight in schizophrenia broadly defined as an awareness into illness, symptoms and need for treatment, is a clinically important phenomenon because it is associated with a range of adverse outcomes such as longer duration of untreated psychosis, poor treatment adherence, frequent hospitalizations, increased risk for involuntary commitments, increased risk for violence towards self and others, and risk for treatment resistance. Such outcomes significantly add to the morbidity in patients with schizophrenia. Both good insight and impaired insight contribute to adverse clinical outcomes which imposes significant challenges in the clinical management of such patients. There is also a lack of clarity in insight’s complex relationship with other psychopathology symptoms and cognitive domains which adds to the challenges. Thus, a neurobiological approach examining the etiological mechanisms of insight that lead to insight variations and the clinical outcomes associated with it, is critical to develop precise, effective and clinically meaningful interventions.

A systematic review of the neurobiological basis of insight identified a number of neuroimaging studies that examined the etiology of insight. Such studies identified shared and unique neural correlates for different insight dimensions. Only three studies examined insight at a cellular level and found that impaired insight was associated with reduced oligodendrocytes clusters in parietal lobe regions. No studies had examined a genetic or molecular basis of insight despite evidence supporting the heritable and trait like properties of insight. The goal of this dissertation was to develop knowledge on the neurobiological basis of insight in schizophrenia. Specifically, we examined (1) the interrelationships among insight, psychopathology symptoms and cognitive domains (2) associations of a validated schizophrenia polygenic risk score with insight, psychopathology symptoms and cognitive domains and (3) associations of candidate gene associations with insight, psychopathology symptoms and cognitive domains.

Methods: We first conducted a cross sectional analyses (N=1391) of clinical data to examine interrelationships among insight dimensions (illness and treatment insight), psychopathology symptom dimensions (positive, negative disorganized, excited and depressed) and cognitive domains using baseline data from the Clinical Antipsychotics Trial of Intervention Effectiveness (CATIE). Structural equation modeling was implemented to examine the direct and indirect effects among variables to identify causal relational paths. Next, in the sample of CATIE patients with genetic data (N= 741) we tested if genetic susceptibility to schizophrenia measured as a schizophrenia polygenic risk score (PRS) could predict insight, psychopathology symptoms and cognitive performance. Finally, we tested if there were specific candidate genes within the well validated schizophrenia autosomal loci associated with insight, psychopathology symptoms and cognitive performance. Given the association of reduced oligodendrocyte cell clusters with insight, we also examined whether a set of 11 schizophrenia related oligodendrocyte genes were associated with insight.

Results: From our cross-sectional analysis of clinical data we found that positive, depressed and disorganized symptoms were associated with illness insight, with the strongest effect for disorganized symptoms. Illness insight exerted a mediation effect between these symptoms and treatment insight. A small mediating effect of neurocognition was observed between (1) disorganized symptoms and treatment insight and (2) depressed symptoms and treatment insight. Contradictory to previous studies we did not see a relationship between negative symptoms and insight dimensions.

We found strong evidence for polygenic burden predicting insight. PRS associations were significant for overall insight (R2 = 0.005), treatment insight dimension (R2 = 0.005) and the poor insight group (Nagelkerke’s R2 = 0.032) in our analyses. We also found significant association of variants rs320073, an intergenic variant and rs1479165 in the SOX2-OT gene for poor insight. Five of the top hit SNPs among 2487 SNPs tested were in the SOX2-OT gene.

PRS associations were significant only for negative and positive symptoms explaining 0.5% and 0.7% of the variance respectively. We did not identify any significant candidate genes associated with symptoms. PRS did not predict neurocognitive composite measure and social cognitive measure in our sample. However, neurocognitive subdomains of working memory (R2= 0.006) and vigilance (R2 = 0.008) was significantly predicted by PRS but these did not survive permutation correction.

Conclusion: In our clinical data we have clarified some of the complexity of insight’s relationship with symptom and cognitive domains. Our findings suggest shared and unique paths for insight dimensions and highlights the mediation effects of neurocognition on specific symptoms and treatment insight. Further studies designed to examine these potential causal relationships are needed.

In our genetic data, we found a strong polygenic signal and a specific candidate gene association with SOX2-OT for poor insight. Polygenic signal for treatment insight but not illness insight suggests differences in biological mechanisms for different insight dimensions. To the best of our knowledge, this is the first study to provide molecular genetic evidence for insight variations in schizophrenia.

We also replicated known associations of schizophrenia PRS with negative symptoms and neurocognitive domains of working memory and vigilance in our sample of patients with chronic schizophrenia. A significant PRS prediction for positive symptoms in our chronic sample offers an avenue for further study. These findings have implications for future research and potentially clinical practice. The approach of investigating the biological basis of symptoms which are tied to long term adverse outcomes have the potential to help identify precise and effective treatments in patients afflicted with schizophrenia.






Xavier, Rose Mary (2017). Neurobiology of Insight in Schizophrenia: Interrelationships with Symptom Domains and Genetic Influences. Dissertation, Duke University. Retrieved from


Dukes student scholarship is made available to the public using a Creative Commons Attribution / Non-commercial / No derivative (CC-BY-NC-ND) license.