Insight and Neurocognitive Function in Schizophrenia

SIR: Although impaired awareness of illness (poor insight) was already recognized as the single most common symptom in acute schizophrenia three decades ago, only recently have vigorous research efforts been made to uncover mechanisms associated with impaired insight. Neuropsychological accounts have focused on poor insight as a result of neurocognitive deficits secondary to the cerebral disease process in schizophrenia, specifically those associated with frontal lobe dysfunction. Consistent with these accounts, several studies have provided evidence of a relationship between poor insight and poor performance on neuropsychological tests sensitive to frontal lobe functioning,¹ although relationships with generalized cognitive deficits have also been described.² However, results of other studies regarding the relation between insight and neurocognitive function are equivocal.^3,4

We investigated the relationship between insight and measures of short-term/working memory and source memory in 38 patients with schizophrenia who were also participants in a larger study on the neurocognitive basis of hallucinations. Insight was measured with item 12 of the general psychopathology subscale of the Positive and Negative Syndrome Scale (PANSS). We included measures of passive and active working memory: the Digit Span forward and backward (from the Wechsler Adult Intelligence Scale) and the Visual Elevator task (from the Test of Everday Attention), respectively. In addition, we included a measure of source memory;⁵ this type of memory has been related to frontal cortex activation and to schizophrenia symptoms.

Multiple regression analysis showed that the four neurocognitive tests did not explain any significant portion of observed variance in insight ratings (F_model=1.88, df=37, P=0.14). Individual partial correlations between the tests and insight ratings were neither significant (all r<0.27, all P>0.10). However, insight ratings did correlate significantly with the negative symptoms subscale of the PANSS (r=0.37, n=38, P=0.02), in accordance with a very recent report by Buckley et al.⁶ Negative symptoms might thus be a confounding factor for which previous studies have failed to control.

In sum, our findings provide additional evidence for the recent conclusion by Carroll et al.⁴ that poor insight in schizophrenia may result directly from the disease process itself rather than being secondary to neurocognitive deficits.