Through conceptualizing poor insight in psychotic disorders as a form of anosognosia (neurological deficit), frontal lobe dysfunction is often ascribed a vital role in its pathogenesis. Whether non-frontal brain regions are important for insight remains to be investigated. We used a multi-method approach to examine the neural morphometry of all cortical regions for insight in first-episode psychosis. Insight was rated in 79 people with a first-episode psychosis with the awareness of illness and awareness of treatment need and efficacy items of the Scale for assessment of Unawareness of Mental Disorder. Participants were assessed with magnetic resonance imaging. Cortical thickness analysis and voxel-based morphometry were utilized to identify the possible neuroanatomical basis of insight. Cortical thickness technique revealed that poorer awareness of illness was associated with regional thinning in left middle frontal and inferior temporal gyri. Poorer awareness of treatment need and efficacy was associated with cortical thinning in left medial frontal gyrus, precuneus and temporal gyri. No significant associations emerged between any insight measure and gray matter density using voxel-based morphometry. The results confirm predictions derived from the anosognosia/neuropsychology account and assert that regional thickness in frontal cortex is associated with awareness of illness in the early phase of psychosis. The fact that prominent thickness reductions emerged in non-frontal regions of the brain in parietal and temporal cortices for both awareness of illness and awareness of treatment need and efficacy suggests that the neural signature of insight involves a network of brain structures, and not only the frontal lobes as previously suggested.
A substantial proportion of people with psychosis demonstrate lack of insight into their illness, including difficulty recognizing the pathological nature of their symptoms and acknowledging the need for treatment. One longstanding view states that impaired insight in psychosis is similar to anosognosia (Amador et al., 1991 and Babinski, 1914), the unawareness of symptoms in neurological disorders observed following right frontal, parietal or temporal lobe lesions. Research on the neural correlates of insight has been motivated by this anosognosia account, which regards poor insight to be a reflection of prefrontally mediated neuropsychological or neurological dysfunction (David, 1999 and Lewis, 1934). Consistent with this model, neuroimaging studies have established that volumetric reductions in circumscribed frontal cortical regions are at the core of poor insight in chronic schizophrenia (Bassitt et al., 2007, Flashman et al., 2001, Ha et al., 2004, Lee et al., 2006 and Sapara et al., 2007).
In two first-episode schizophrenia studies, Shad et al., 2006 and Shad et al., 2004 showed that patients with poor awareness of illness showed volumetric reductions in right dorsolateral prefrontal cortex (DLPFC), relative to patients with good insight. The authors concluded that volumetric reductions in frontal cortex are an integral part of poor insight, supporting the anosognosia account of impaired insight. However, looking only at the frontal cortex, by definition, forfeits exploration of potential structural alterations in posterior brain areas underlying insight, such as parietal and temporal cortices. This is an important consideration because there is evidence for associations between poor insight and parietal and temporal neuropsychological dysfunction (Goodman et al., 2005 and McEvoy et al., 1996). Moreover, executive dysfunction – part and parcel of poor insight in psychosis (Aleman et al., 2006) – may reflect impairment in a distributed cortical–subcortical network (Minzenberg et al., 2009). Further, morphometric studies in chronic schizophrenia have demonstrated a relationship between insight and gray matter (GM) volumes in non-frontal brain regions, including parietal and temporal cortices (Cooke et al., 2008 and Ha et al., 2004). Theoretically, as anosognosia is classically observed following parietal and temporal lesions, an anosognosic model for the neurological underpinnings of poor insight would predict structural deficits in parietal and temporal regions.
At a methodological level, the abovementioned morphometric studies have employed voxel-based analyses of magnetic resonance images (MRIs), capturing the volume of structures by the totality of voxels it encompasses or by examining gray matter density. With recent advances in neuroimaging methods, it is now possible to perform fully automated cortical thickness measurements of MRIs at a subvoxel resolution. This metric provides a direct measurement in millimeters of gray matter morphology, and moreover is anatomically meaningful, reflecting cortical laminar structure and integrity. In vivo cortical thickness measurements have not been used to address the question of whether gray matter integrity is associated with insight in psychosis.
In this study we tested predictions derived from the anosognosia account and hypothesized that poor insight in people with first-episode psychosis would associate with structural deficits in 1) DLPFC and 2) parietal and temporal cortices. We further explored whether awareness of treatment need and treatment efficacy, two insight dimensions partially independent from awareness of illness (David et al., 1992), would associate with certain neuroanatomical deficits. We applied cortical thickness analyses and voxel-based morphometry (VBM) of MRIs. The VBM analysis provided a voxel-based estimate of GM density (Ashburner and Friston, 2000 and Ashburner and Friston, 2001) and allowed us to compare directly our results to those from previous insight–neuroimaging studies. Cortical thickness measurements are performed at a subvoxel resolution and provide a direct measurement in millimeters of GM morphology, and have not been used to address the question of whether GM integrity is associated with insight in psychosis.
