Silent period (SP) is widely used in transcranial magnetic stimulation studies. Methodologically, SP is usually elicited at stimulus intensities corresponding to a certain percentage of corticomotor threshold. Because this approach might lead to factitious SP changes, the present study was designed to develop, in a stepwise manner, a method for investigating SP independently of corticomotor threshold. First, stimulus-response (S-R) curves of SP against stimulus intensity (SI) were constructed and quantitatively described in healthy volunteers. Second, various methodological issues such as the optimum model for describing the relationship between SP duration and SI and the importance of the type of stimulating coil were addressed. Finally, the proposed method and a commonly used method (eliciting SPs at 130% MT SI) were directly compared for a group of epileptic patients for whom administration of oxcarbazepine resulted in significant corticomotor threshold elevation. Twenty-one subjects (eleven females, median age, 38 years) were studied. SPs were obtained with a figure-of-eight coil using a standardized procedure (recording, FDI). Pilot experiments indicated that at least four trials were required, at each intensity level, to estimate the mean SP duration within 10% of the true mean. Therefore, SPs were determined from the average of four trials with 5% increments from 5 to 100% maximum SI. In a second set of experiments, SPs were obtained for fifteen subjects using a circular coil. In a third set of experiments, eight epileptic patients were studied before and after administration of oxcarbazepine (mean dose 1553 mg, range 900-1800 mg). The S-R curves were fitted to a Boltzman function and to first-order to fourth-order polynomial and sigmoid functions. The Boltzman function described the data accurately (R2=0.947-0.990). In addition, direct comparison of the six models with an F-test proved the superiority of the first. The best-fit parameters of the reference curve, i.e. the maximum and minimum values, the slope, and V50 (the SI at which SP duration is halfway between Min and Max) were 230.8+/-3.31 ms (x+/-SEM), -11.51+/-3.31 ms, 11.56+/-0.65%, and 49.82+/-0.65%, respectively. When the curves obtained with the circular coil were compared with those obtained with the figure-of-eight coil, there were differences between V50 (51.69+/-0.72 vs 47.95+/-0.82, P<0.001) and SP threshold (31.15 vs 24.77, P<0.01) whereas the other best-fit values did not differ significantly. Oxcarbazepine increased corticomotor threshold from 45.3+/-5.8% at baseline to 59.4+/-10.4% (P<0.001). According to the commonly used method, the drug significantly prolonged SP (from 117.6+/-42.4 ms to 143.5+/-46.5 ms, P<0.001) and, consequently, enhanced brain inhibition. In contrast, study of the SP curves led to the conclusion that oxcarbazepine does not affect the Max value and slope but significantly increases V50 and SP threshold (from 54.5+/-4.9% to 59.9+/-7.2% and from 29.1+/-6.4% to 34.6+/-6.8%, respectively, P<0.01). These findings imply that oxcarbazepine does not enhance brain inhibitory mechanisms. Thus, in situations characterized by significant changes in corticomotor threshold the proposed method provides results clearly different from a commonly used approach. It is concluded that S-R curves obtained with a figure-of-eight coil in 5% increments and fitted to a Boltzman function provide an accurate, comprehensive, and clinically applicable method for exploring SP.