Ultrasound has been used to evaluate diaphragm thickness in the zone of apposition of the diaphragm to the rib cage. The purpose of this study was to determine if ultrasonography could distinguish between a paralyzed and normally functioning diaphragm. We predicted that a paralyzed diaphragm would be atrophic and not shorten, therefore it would be thin and not thicken during inspiration. Thirty subjects (five with bilateral diaphragm paralysis, seven with unilateral diaphragm paralysis, three with inspiratory weakness but normally functioning diaphragms, and 15 healthy control subjects) had diaphragm ultrasound performed with a 7.5 to 10.0 MHz transducer placed over the lower rib cage in the mid-axillary line. The thickness of the diaphragm (tdi) was measured to the nearest 0.1 mm at FRC (t(di)FRC) and TLC (t(di)TLC). Diaphragm thickening during inspiration (delta t(di)) was calculated as (t(di)TLC - t(di)FRC)/t(di)FRC. In patients with unilateral paralysis, t(di) and delta t(di) for the paralyzed hemidiaphragm were significantly less than those values for the normally functioning hemidiaphragm (1.7 +/- 0.2 mm versus 2.7 +/- 0.5 mm [mean + SD] p < 0.01 for t(di), and -8.5 +/- 13% versus 65 +/- 26% [p < 0.001] for delta t(di)). The t(di) and delta t(di) for patients with bilateral diaphragm paralysis were significantly less than those values for the healthy volunteers (1.8 +/- 0.2 versus 2.8 +/- 0.4 and -1 +/- 15% versus 37 +/- 9% for t(di) and delta t(di), respectively) (p < 0.001). We conclude that ultrasound measurements of t(di) and delta t(di) can be used to determine if a diaphragm is paralyzed and confirm our predictions that a chronically paralyzed diaphragm is atrophic and does not thicken during inspiration.