We tested 28 individuals with schizophrenia (SZ) and 16 healthy individuals on a test of logical reasoning and “cognitive gating,” defined as the ability to discriminate between relevant and irrelevant information in confirming or disconfirming a given belief. The Logical Reasoning and Cognitive Gating Task tests both processes under neutral and affect-laden conditions. This is done by presenting formally identical constructs using benign and emotionally arousing language. When separated by symptom profiles, we found statistically significant differences for performance and arousal response between patients with delusions, patients with formal thought disorder, and patients with neither delusions nor formal thought disorder, as well as between patients and healthy controls. When analyzed by error type, we found that nearly all errors by delusional patients were caused by overly restrictive information choice, a pattern that may be related to a delusional patient's tendency to “jump to conclusions” on Bayesian probabilistic tasks. This is in contrast to patients with formal thought disorder, whose low performance resulted also from overly extensive information choice. The tendencies towards restriction were exacerbated by arousal, which is consistent with studies on cognition and arousal in healthy individuals. After briefly examining research on emotional arousal and SZ, and the interaction between emotional arousal and restriction of perceptual cues in healthy individuals, we conclude by suggesting a model which accounts for the distinctive cognitive characteristics of delusional patients by their possessing distinct vulnerabilities to emotional arousal. Specifically, these results suggest the possibility that delusional patients process information in a manner that is essentially intact. However, delusional patients may possess an acute vulnerability to emotional arousal that might cause delusional individuals to behave cognitively as if they were healthy individuals under significantly more severe forms of stress.