Important Steps in the Development of Cognitive-Enhancing Drugs in Schizophrenia

The relationship between impaired cognitive functioning and poor social and occupational functioning is a major factor behind recent efforts to stimulate the development of cognitive-enhancing agents (1, 2) . The FDA has required that any such agent not only have a significant effect on some aspect of cognition but also have a beneficial effect on role function (3) . The requirement that a potential cognitive-enhancing agent also have an effect on functional outcome has raised several methodological issues, including how best to assess functional outcome in the context of a clinical trial. In particular, the likelihood that the employment status or social network of a subject would significantly change during the course of an 8–12 week clinical trial is minimal, especially given the role of factors beyond the control of clinical trials, such as psychosocial supports. This has led to the suggestion that proxy measures of functional outcome may be a more appropriate approach for assessing drug effects on function (3) . The proxy measure would be designed to assess the capacity of an individual to perform various social and occupational role functions. The use of a proxy measure assumes that change in cognitive function will improve functional capacity and that improvement in functional capacity is ultimately associated with improved functional outcome. Although several studies have demonstrated the cross-sectional association between functional capacity and outcome (4, 5), there have been little data to support the assumption of the relationship between change in cognitive function and capacity.

The study by Harvey and colleagues in this issue of the Journal addresses this topic. In an 8-week, double-blind comparison of quetiapine and risperidone, the authors examined the relationship among a performance-based measure of social competence (the Social Skills Performance Assessment [6] ), neuropsychological test performance, and a measure of social cognition (the Penn Emotional Acuity Test [7] ). The Social Skills Performance Assessment comprises two 3-minute vignettes that require the subject to role play meeting a new neighbor and complaining to a landlord about an unrepaired leak. The neuropsychological tests included measures of attention, verbal memory, executive function, and verbal fluency. The Penn Emotional Acuity Test requires subjects to accurately identify facial emotions. There were significant improvements on several neuropsychological measures and on the Social Skills Performance Assessment over the course of the study. Executive and verbal memory measure change scores contributed significantly and explained 20% of the variance in Social Skills Performance Assessment change scores. There were no significant changes in the Penn Emotional Acuity Test with treatment, nor was change in this measure associated with change in the Social Skills Performance Assessment.

The most important findings from the study are the demonstration that a simple measure of functional capacity can change during the course of a short clinical trial, i.e., 8 weeks, and that this change is associated with change in neuropsychological test performance. These results help address the critical issue of what type of measure should be used as a coprimary measure in clinical trials of cognitive-enhancing drugs. However, several important questions remain. First, the functional capacity measure used in the study only included vignettes to assess social competence. The extent to which measures of work competence will show similar ability to change within the context of a short clinical trial and whether this change will be related to change in neuropsychological test performance needs to be established. Second, the magnitude of the correlations between the individual neuropsychological tests and the Social Skills Performance Assessment was small. The size of the correlations may reflect 1) the relatively small magnitude of change in the neuropsychological test scores; 2) the limited number of neuropsychological measures used (for example, there were no measures of processing speed or working memory, which have been shown to be related to role function [ 1, 8 ]); 3) the short duration of the clinical trial; or 4) the actual magnitude of the association between neuropsychological test performance and measures of social competence. If the current estimate of the magnitude of the association between these measures is valid, then the ability to detect such associations will require relatively large patient groups. Finally, long-term follow-up studies will be required to demonstrate that change in functional capacity measures ultimately leads to improvement in functional outcome. As Harvey and colleagues note, this latter task is complicated by the myriad of largely uncontrollable environmental factors that have an impact on the ability of an individual to improve her/his social and occupational situation.

A related issue in the evaluation of new drugs is the development of valid biomarkers. The article by Lee and colleagues addresses this problem. They examined the longitudinal relationship between recovery from an acute psychotic episode and change in blood flow, as assessed with fMRI, during the performance of a social cognition task. Fourteen patients with schizophrenia and 14 normal subjects participated in the study. The patients were scanned during their acute episode and 10–11 weeks later, presumably during recovery from their episode. Patients were treated with essentially the same medications at baseline and follow-up. The normal subjects also received two fMRI scans over the same time span. The scans were conducted while the subjects were asked to make either an empathic or forgivability judgment. Subjects were presented two possible answers and, in a forced choice paradigm, the subjects were asked to select one of the answers as their response. At the time of each scan, the patients received positive and negative symptom, insight, and social function ratings. Their illness insight was assessed with the Schedule for the Assessment of Insight (9), and the Life Skills Profile (10) was used to assess social functioning.

The patients with schizophrenia showed increased blood flow in multiple brain regions during the second scan. In particular, there was increased activation of the left medial prefrontal cortex during empathic and forgivability judgments. The right lingual, cuneus, and left ventral anterior cingulate gyri also showed increased activation with recovery during the performance of the social cognition paradigm. The increased medial prefrontal cortex activation was significantly associated with changes in insight and social functioning scores. The right lingual gyral activation was also significantly related to improved insight scores. The results are important for several reasons. First, they implicate these regions in the functional neuroanatomy of social cognition and function. Second, they suggest that the level of activity of these regions is not a fixed trait characteristic but rather can change with either clinical state or treatment. Finally, left prefrontal medial activation during the performance of a social cognition task may be able to serve as a biomarker for treatment development. The delineation of the underlying mechanism of this change in activation may provide interesting leads for drug development. Alternatively, the paradigm may be used to evaluate CNS activity and drug effects.

In summary, the development of new drugs for the treatment of cognitive impairments and associated functional disabilities is going to require a multidimensional approach that utilizes and integrates information from preclinical pharmacology, cognitive neuroscience, and clinical neuropharmacology. The Harvey et al. and Lee et al. studies in the current issue of the Journal are instructive examples of how these approaches can be used to facilitate new drug development.