Specific Frontal Lobe Subregions Correlated With Unawareness of Illness in Schizophrenia

Previous literature suggests that the majority of patients with schizophrenia demonstrate unawareness of their illness, including difficulty identifying their symptoms and recognizing that they have a mental disorder. This unawareness can significantly affect treatment compliance. It has been noted that unawareness of symptoms in neurological disorders (i.e., anosognosia) may be comparable to unawareness in schizophrenia.¹ Anosognosia has often been associated with bilateral frontal lobe and right parietal lesions.² Consistent with this link, there is evidence of a relationship between unawareness of illness in schizophrenia and frontal and parietal lobe neuropsychological deficits.^3,4

Given the critical role the frontal lobes and related circuitry play in self-monitoring and self-awareness, we hypothesized that deficits in awareness in patients with schizophrenia would be associated with frontal lobe structural abnormalities. Flashman et al.⁵ previously found no differences in global left and right frontal lobe volumes between aware and unaware patients with schizophrenia. However, the frontal lobes are not homogeneous in function, and subtle changes in the circuitry underlying unawareness of illness might be lost with such an approach. Thus, there might be differences in subregions of the frontal lobe that correlate with unawareness.

METHODS

Subjects were 15 patients (13 inpatients, 2 outpatients) with a diagnosis of either schizophrenia (n=12) or schizoaffective disorder (n=3) recruited from New Hampshire Hospital and Dartmouth-Hitchcock Medical Center. All patients were receiving neuroleptics, but not other medications, at the time of the study. Demographic information is presented in Table 1. After complete description of the study to the subjects, written informed consent was obtained from them. All participants completed a comprehensive clinical assessment, including the Structured Clinical Interview for DSM-IV, Brief Psychiatric Rating Scale (BPRS), Schedule for Assessment of Positive Symptoms (SAPS), and Schedule for Assessment of Negative Symptoms (SANS) with at least two of the authors present (L.A.F., T.W.M.).

The Scale to Assess Unawareness of Mental Disorders (SUMD; Amador et al.⁶) is a semistructured interview and scale that assesses awareness of illness and symptoms. There are general questions (e.g., “Do you have a mental illness?”) and items that assess awareness of specific symptoms (e.g., “Could the voices you hear be hallucinations?”). The latter specific questions are asked only for items where the patient scored 3 or higher on the SANS and SAPS. Items are rated on a five-point Likert-type scale, with 1 indicating full awareness and 5 indicating complete unawareness of a symptom. When items are rated 1, 2 or 3 (indicating some level of awareness), subjects are asked how they explain the symptom, and ratings of attribution are rated in the same way (1 indicating correct attribution of a symptom to a mental illness and 5 indicating complete misattribution of a symptom).

All subjects received a structural MRI using a 1.5-T GE Signa scanner. Volumetric scan parameters were a Spoiled Grass sequence (TE=13 ms, TR=38 ms, flip angle=45, NEX=1; in-plane resolution=0.9375), which yielded a series of 124 contiguous 1.5-mm coronal slices. Volume estimates of eight distinct frontal lobe regions (frontal pole, superior frontal gyrus, middle frontal gyrus, inferior frontal gyrus, orbital frontal gyrus, precentral gyrus, gyrus rectus, and cingulate) were derived in both hemispheres by the same technician (J.R.) using manual tracing on slices 1.5 mm thick for each region. Boundary designations of each region for each subject were then reviewed by two additional investigators (S.C.J, R.L.G.). These investigators were blind to the awareness status of the patients. Determination of the boundaries of regions was made according to the atlas of Duvernoy;⁷ a detailed manual with boundary descriptions is available from the authors on request.

Pearson correlations were used to examine the relationship between unawareness and misattribution and the eight frontal lobe subregions. Because Flashman et al.⁵ indicated that brain size itself is associated with awareness of illness, intracranial content (brain tissue plus cerebrospinal fluid) was partialed out of all correlational analyses of selected subregions.

RESULTS

Large inverse correlations (P≤0.01) were found between level of unawareness and bilateral middle frontal gyrus volume (–0.92, right; –0.72, left); smaller volumes were correlated with greater unawareness of illness. In addition, significant inverse correlations were found between unawareness and the right gyrus rectus (–0.87) and the left anterior cingulate gyrus (–0.73). In contrast, significant inverse correlations were found between misattribution of symptoms and superior frontal gyrus (–0.73, bilaterally); smaller volumes were associated with greater misattribution. No other significant relationships were seen. This pattern of results did not change when the patients with schizoaffective disorder were eliminated from the analyses. Although left middle frontal gyrus, left anterior cingulate gyrus, and bilateral superior frontal gyrus correlations were not significant following Bonferroni adjustment due to a reduction of power, the effect sizes were large and were significant at P<0.01. Replication will be important to confirm the observed relationships.

DISCUSSION

These results suggest that characteristics of specific frontal lobe subdivisions are associated with distinct aspects of unawareness in patients with schizophrenia (i.e., awareness/unawareness of symptoms, correct/incorrect attribution of symptoms). In all cases, unawareness and misattribution were associated with smaller volume estimates (i.e., the more unawareness, the smaller the brain region).

While the exact meaning of these preliminary findings is unclear, they appear consistent with suggestions made previously by Stuss and Benson^2,8 on the role of the frontal lobe in self-monitoring and awareness. The strong correlations between bilateral middle frontal gyri and unawareness suggest involvement of dorsolateral prefrontal cortex (DLPFC). This is also consistent with neuropathological studies^9–13 implicating abnormalities in DLPFC in schizophrenia.

One hypothesis that requires further corroboration is that the association between unawareness and the smaller subregions may be mediated by deficits in working memory. The DLPFC and cingulate gyrus are important components of working memory circuitry.^14–16 The gyrus rectus has been conceptualized as an orbitofrontal extension of the cingulate complex¹⁷ that receives projections from the mediodorsal nucleus of the thalamus (the major thalamic relay to the prefrontal area). Partial resection of the gyrus rectus following anterior communicating artery aneurysm has been associated with problems in short-term memory.¹⁸ Patients with schizophrenia show impairment on many memory tasks, including those assessing working memory (e.g., the Wisconsin Card Sorting Test). In fact, working memory deficits may be a core component of the illness.¹⁴ Working memory is important for comparing current and past experience and for integrating this information to guide current and future behavior. Impairment in this domain might lead to problems in accurate interpretation of symptoms. The inability to hold symptom information in working memory while comparing it with past experiences may make it difficult to successfully categorize the current symptom as aberrant, which might be manifested as unawareness of illness.

This is a preliminary study with a relatively small sample size. Future studies to replicate and extend these findings and hypotheses are needed. Our sample included a relatively small number of females, although the male-to-female ratio (∼3:1) is not inconsistent with other samples. Further, our design did not permit us to examine the influence of medication on frontal subregion volumes.

Despite these limitations, our data suggest that deficits in awareness of illness in individuals with schizophrenia may have a neural substrate involving subregions of the frontal lobe. Direct assessment of working memory as a function of unawareness might lend further support to this hypothesis. Measurement of the parietal lobes may further understanding of their contribution to manifested unawareness of illness in patients with schizophrenia.

ACKNOWLEDGMENTS

The authors thank Robert Vidaver, John Crampton, Cheryl Brown, Molly Sparling, Megan Kazlauskas, Denise Bono, and Chad Moritz for their assistance with data collection and analysis. This research was supported by a Young Investigator Award from the National Alliance for Research on Schizophrenia and Depression. Additional support was provided by New Hampshire Hospital and the Ira DeCamp Foundation.

TABLE 1. Demographic characteristics of the patient sample (N=15)

Enlarge table