To examine the importance of blood-borne vs. neural mechanisms for hormonal responses and substrate mobilization during exercise, six spinal cord-injured tetraplegic (C5-T1) males (mean age: 35 yr, range: 24-55 yr) were recruited to perform involuntary, electrically induced cycling [functional electrical stimulation (FES)] to fatigue for 24.6 +/- 2.3 min (mean and SE), and heart rate rose from 67 +/- 7 (rest) to 107 +/- 5 (exercise) beats/min. Voluntary arm cranking in tetraplegics (ARM) and voluntary leg cycling in six matched, long-term immobilized (2-12 mo) males (Vol) served as control experiments. In FES, peripheral glucose uptake increased [12.4 +/- 1.1 (rest) to 19.5 +/- 4.3 (exercise) mumol.min-1.kg-1; P < 0.05], whereas hepatic glucose production did not change from basal values [12.4 +/- 1.4 (rest) vs. 13.0 +/- 3.4 (exercise) mumol.min-1.kg-1]. Accordingly, plasma glucose decreased [from 5.4 +/- 0.3 (rest) to 4.7 +/- 0.3 (exercise) mmol/l; P < 0.05]. Plasma glucose did not change in response to ARM or Vol. Plasma free fatty acids and beta-hydroxybutyrate decreased only in FES experiments (P < 0.05). During FES, increases in growth hormone (GH) and epinephrine and decreases in insulin concentrations were abolished. Although subnormal throughout the exercise period, norepinephrine concentrations increased during FES, and responses of heart rate, adrenocorticotropic hormone, beta-endorphin, renin, lactate, and potassium were marked. In conclusion, during exercise, activity in motor centers and afferent muscle nerves is important for normal responses of GH, catecholamines, insulin, glucose production, and lipolysis. Humoral feedback and spinal or simple autonomic nervous reflex mechanisms are not sufficient. However, such mechanisms are involved in redundant control of heart rate and neuroendocrine activity in exercise.