The inspiratory muscles (IM) can be trained by having a subject breathe through inspiratory resistive loads or by use of unloaded hyperpnea. These disparate training protocols are characterized by high inspiratory pressure (force) or high inspiratory flow (velocity), respectively. We tested the hypothesis that the posttraining improvements in IM pressure or flow performance are specific to training protocols in a way that is similar to force-velocity specificity of skeletal muscle training. IM training was accomplished in 15 normal subjects by use of three protocols: high inspiratory pressure-no flow (group A, n = 5), low inspiratory pressure-high flow (group B, n = 5), and intermediate inspiratory pressure and flow (group C, n = 5). A control group (n = 4) did no training. Before and after training, we measured esophageal pressure (Pes) and inspiratory flow (VI) during single maximal inspiratory efforts against a range of external resistances including an occluded airway. Efforts originated below relaxation volume (Vrel), and peak Pes and VI were measured at Vrel. Isovolume maximal Pes-VI plots were constructed to assess maximal inspiratory pressure-flow performance. Group A (pressure training) performed 30 maximal static inspiratory maneuvers at Vrel daily, group B (flow training) performed 30 sets of three maximal inspiratory maneuvers with no added external resistance daily, and group C (intermediate training) performed 30 maximal inspiratory efforts on a midrange external resistance (7 mm ID) daily. Subjects trained 5 days/wk for 6 wk. Data analysis included comparison of posttraining Pes-VI slopes among training groups.(ABSTRACT TRUNCATED AT 250 WORDS)