Neural inspiratory time (TI) is a measurement of fundamental importance in studies of patient-ventilator interaction. The measurement is usually based on recordings of flow, esophageal pressure (Pes), and transdiaphragmatic pressure (Pdi), but the concordance of such estimates of neural TI with a more direct measurement of neural activity has not been systematically evaluated. To address this issue, we studied nine ventilator-supported patients in whom we employed esophageal electrode recordings of the diaphragmatic electromyogram (EMG) as the reference measurement of neural TI. Comparison of the indirect estimates of neural TI duration, based on flow, Pes, and Pdi against the reference measurement, revealed a mean difference (bias) ranging from -54 to 612 ms during spontaneous breathing and from -52 to 714 ms during mechanical ventilation; the respective precisions (standard deviations of the differences) ranged from 79 to 175 ms and from 74 to 221 ms. Because an indirect estimate of neural TI duration could be identical to that of the reference measurement and yet be displaced in time, this lag or lead was quantified as the phase angle of neural TI onset. Flow-based estimates of the onset of neural TI displayed a systematic lag, which may be explained at least in part by concurrent intrinsic positive end-expiratory pressure. In conclusion, the indirect estimates of the onset and duration of neural TI in ventilator-dependent patients displayed poor agreement with the diaphragmatic EMG measurement of neural TI.