The effects of speaking and listening on N1 amplitude in response to speech sounds differed between the patients and comparison subjects (group-by-condition-by-site interaction: F=5.69, df=2, 24, p<0.02, two-tailed); the group-by-condition interaction was significant for the Cz site (F=4.83, df=1, 12, p<0.05, two-tailed) but not Fza or Fz. This interaction was due to the fact that the N1 elicited by vowels was smaller when the vowels were spoken (mean=–0.96 μV, SD=1.09) than when they were played back (mean=–3.08 μV, SD=2.54) in the comparison subjects (paired t=–2.04, df=6, p=0.04, one-tailed, effect size=0.77) but not in the patients (paired t=0.84, df=6, p=0.22, one-tailed, effect size=0.32 ). In the patients, the N1 during speech (mean=–2.53 μV, SD=2.07) was not smaller than during playback (mean=–2.09 μV, SD=1.64) (F1). Hemispheric effects at the central coronal sites were assessed at C3, Cz, and C4. Although the group-by-condition interaction was significant (F=4.72, df=1, 12, p=0.05), neither the group-by-condition-by-hemisphere interaction (F=0.61, df=1, 12, p=0.52) nor the condition-by-hemisphere interaction (F=2.51, df=1, 12, p=0.12) was significant.