Spinal cord injured (SCI) individuals most often contract their injury at a young age and are deemed to a life of more or less physical inactivity. In addition to the primary implications of the SCI, severe SCI individuals are stigmatized by conditions related to their physically inactive lifestyle. It is unknown if these inactivity related conditions are potentially reversible and the aim of the present study was, therefore, to examine the effect of exercise on SCI individuals. Ten such individuals (six with tetraplegia and four with paraplegia; age 27-45 years; time since injury 3-23 years) were exercise trained for 1 year using an electrically induced computerized feedback controlled cycle ergometer. They trained for up to three times a week (mean 2.3 times), 30 min on each occasion. The gluteal, hamstring and quadriceps muscles were stimulated via electrodes placed on the skin over their motor points. During the first training bouts, a substantial variation in performance was seen between the subjects. A majority of them were capable of performing 30 min of exercise in the first bout; however, two individuals were only able to perform a few minutes of exercise. After training for 1 year all of the subjects were able to perform 30 min of continuous training and the work output had increased from 4 +/- 1 (mean +/- SE) to 17 +/- 2 Kilo Joules per training bout (P < 0.05). The maximal oxygen uptake during electrically induced exercise increased from 1.20 +/- 0.08 litres per minute measured after a few weeks habituation to the exercise to 1.43 +/- 0.09 litres per minute after training for 1 year (P < 0.05). Magnetic resonance cross sectional images of the thigh were performed to estimate muscle mass and an increase of 12% (mean, P < 0.05) was seen in response to 1 year of training. In biopsies taken before exercise various degrees of atrophy were observed in the individual muscle fibres, a phenomenon that was partially normalized in all subjects after training. The fibre type distribution in skeletal muscles is known to shift towards type IIB fibres (fast twitch, fast fatiguable, glycolytic fibres) within the first 2 years after the spinal cord injury. The muscle in the present investigation contained of 63% myosin heavy chain (MHC) isoform IIB, 33% MHC isoform IIA (fast twitch, fatigue resistant) and less than 5% MHC isoform I (slow twitch) before training. A shift towards more fatigue resistant contractile proteins was found after 1 year of training. The percentage of MHC isoform IIA increased to 61% of all contractile protein and a corresponding decrease to 32% was seen in the fast fatiguable MHC isoform IIB, whereas MHC isoform I only comprised 7% of the total amount of MHC. This shift was accompanied by a doubling of the enzymatic activity of citrate synthase, as an indicator of mitochondrial oxidative capacity. It is concluded that inactivity-associated changes in exercise performance capacity and skeletal muscle occurring in SCI individuals after injury are reversible, even up to over 20 years after the injury. It follows that electrically induced exercise training of the paralysed limbs is an effective rehabilitation tool that should be offered to SCI individuals in the future.