Unilateral diaphragm dysfunction (UDD) is an underdiagnosed cause of dyspnea. Inspiratory muscle training (IMT) is the only conservative treatment for UDD, but the mechanisms of improvement are unknown. We characterised the effects of IMT on dyspnea, exercise tolerance, and respiratory muscle function in people with UDD.Fifteen people with UDD (73% male, 61±8 years) were randomised to 6-months of IMT (50% maximal inspiratory mouth pressure (PI,max), n=10) or sham training (10% PI,max, n=5) (30 breaths 2x·day−1). UDD was confirmed by phrenic nerve stimulation and persisted throughout the training period. Symptoms were assessed by the Transitional Dyspnea Index (TDI) and exercise tolerance by constant-load cycle tests performed pre- and post-training. Esophageal (Pes) and gastric (Pga) pressures were measured with a dual-balloon catheter. Electromyography (EMG) and oxygenation (near-infrared spectroscopy) of respiratory muscles were assessed continuously during exercise.The IMT group (from 45±6%PI,maxto 62±23%PI,max) and sham group (no progression) completed 92% and 86% of prescribed sessions, respectively. PI,max, TDI scores, and cycle endurance time improved significantly more after IMTversussham (mean between-group differences: 28[95%CI:13–28]cmH2O, 3.0[95%CI:0.9–5.1]points, and 6.0[95%CI:0.4–11.5]minutes, respectively). During exercise at iso-time, Pes, Pga, and EMG of the scalene muscles were reduced and oxygen saturation index of the scalene and abdominal muscles higher post-trainingversuspre- only in the IMT group (allp<0.05).The effects of IMT on dyspnea and exercise tolerance in UDD were not mediated by improvement in isolated diaphragm function, but may reflect improvements in strength, coordination, and/or oxygenation of the extra-diaphragmatic respiratory muscles.