In amyotrophic lateral sclerosis (ALS) patients, respiratory insufficiency is a major burden. Diaphragm conditioning by electrical stimulation could interfere with lung function decline by promoting the development of type 1 muscle fibres. We describe an ancillary study to a prospective, non-randomized trial (NCT00420719) assessing the effects of diaphragm pacing on forced vital capacity (FVC). Sleep-related disturbances being early clues to diaphragmatic dysfunction, we postulated that they would provide a sensitive marker. Stimulators were implanted laparoscopically in the diaphragm close to the phrenic motor point in 18 ALS patients for daily conditioning. ALS functioning score (ALSFRS), FVC, sniff nasal inspiratory pressure (SNIP), and polysomnographic recordings (PSG, performed with the stimulator turned off) were assessed before implantation and after four months of conditioning (n = 14). Sleep efficiency improved (69 ± 15% to 75 ± 11%, p = 0.0394) with fewer arousals and micro-arousals. This occurred against a background of deterioration as ALSFRS-R, FVC, and SNIP declined. There was, however, no change in NIV status or the ALSFRS respiratory subscore, and the FVC decline was mostly due to impaired expiration. Supporting a better diaphragm function, apnoeas and hypopnoeas during REM sleep decreased. In conclusion, in these severe patients not expected to experience spontaneous improvements, diaphragm conditioning improved sleep and there were hints at diaphragm function changes.