Abstract
Schizophrenia (SCZ)-related verbal memory impairment is hypothesized to be mediated, in part, by frontal lobe (FTL) dysfunction. However, little research has contrasted the performance of SCZ patients with that of patients exhibiting circumscribed frontal lesions. The current study compared verbal episodic memory in patients with SCZ and focal FTL lesions (left frontal, LF; right frontal, RF; and bi-frontal, BF) on a four-trial list learning task consisting of three lists of varying semantic organizational structure. Each dependent variable was examined at two levels: scores collapsed across all four trials and learning scores (i.e., trial 4–trial 1). Performance deficits were observed in each patient group across most dependent measures at both levels. Regarding patient group differences, SCZ patients outperformed LF/BF patients (i.e., either learning scores or scores collapsed across trial) on free recall, primacy, primary memory, secondary memory, and subjective organization, whereas they only outperformed RF patients on the semantically blocked list on recency and primary memory. Collectively, these results indicate that the pattern of memory performance is largely similar between patients with SCZ and those with RF lesions. These data support tentative arguments that verbal episodic memory deficits in SCZ may be mediated by frontal dysfunction in the right hemisphere.
ACKNOWLEDGEMENT
The authors gratefully acknowledge Rae Dolman, Jenny Chao, and Dan LaPorte for their assistance with subject recruitment, data collection and data management. This research was supported by an operating grant to RBZ from the Medical Research Council of Canada. The authors have no conflict of interest (financial or otherwise) to report with regards to the research reported in this manuscript.
Notes
1. Barnette and MacLean (Citation1999) have utilized Monte Carlo simulations in order to model how standard effect size measures vary systematically as a function of sample size (n) and number of groups (K). From these, they have provided reliable estimations of mean effect size values that could be expected on the basis of chance alone given a particular combination of n and K. In our study, comparisons were always made between two groups (i.e., K = 2), and the lowest sample sizes were in the frontal lesion groups (RF: n = 7, LF: n = 10, BF: n = 12). Therefore, using conservative parameter estimates of K = 2 and n = 10, it can be seen from the data provided by Barnette and MacLean (Citation1999) that a mean effect size value of 0.38 (SD = 0.30) can be expected by chance alone. With this data in mind, we reasoned that a mean effect size that is greater than 1.5 SDs above this chance cut-off could be considered statistically meaningful. Accordingly, we set our meaningful difference cut-off at Cohen’s d > 0.83. As a result, we interpreted group differences that exceeded thresholds of either p < .05 or d > 0.83 as being statistically meaningful. This allowed us to account for the small sample size in the frontal lobe lesion groups and, by extension, minimize the possibility of Type II error.
2. The issue of ceiling effects applies to learning scores for each dependent measure. As such, we examined each dependent variable to determine if the trial 1 group means at any level of list were within 20% of ceiling. We found that only the RF group’s trial 1 recency recall on the blocked list was within 20% of ceiling (14.3%). Importantly, the HC group means were not within 20% of ceiling on trial 1 for any dependent measure at any level of list. However, some individual HC participants were occasionally at ceiling on trial 1 for some measures, particularly on the blocked list. For these participants, there was no way to achieve a learning score above zero. This would have artificially deflated their learning scores on these measures and, by extension, reduced the magnitude of HC-patient group differences.