BACKGROUND: One of the hypothesised mechanisms by which foot orthoses obtain their clinical effect is by influencing muscle activity, however previous studies have reported highly variable findings. The aim of this study was to determine whether orthoses change muscle activity in people with flat-arched feet towards a pattern observed in people with normal-arched feet. METHODS: Thirty young asymptomatic adults with flat-arched feet were recruited. Foot posture was classified using two clinical measurements and four skeletal alignment measurements from weight-bearing foot x-rays. Electromyographic activity was recorded while walking from tibialis posterior and peroneus longus via in-dwelling wire electrodes, and from tibialis anterior and medial gastrocnemius via surface electrodes. Four experimental conditions were assessed: (i) barefoot, (ii) shoe only, (iii) a heat-moulded (modified) prefabricated foot orthosis, and (iv) a 20-degree inverted-style customised foot orthosis. FINDINGS: During the contact phase of gait, tibialis posterior electromyographic amplitude decreased significantly with the prefabricated orthosis (peak amplitude - 19% decrease, P=0.007; RMS amplitude--22% decrease, P=0.002) and the customised orthosis (peak amplitude--12% decrease, P<0.001, RMS amplitude--13% decrease, P=0.001), compared with the shoe-only condition. During the midstance/propulsive phase, peroneus longus electromyographic amplitude increased significantly with the prefabricated orthosis, compared with the shoe-only (peak amplitude--21% increase, P=0.024; RMS amplitude--24% increase, P=0.019) and customised orthosis conditions (peak amplitude--16% increase, P=0.028). INTERPRETATION: The foot orthoses significantly altered tibialis posterior and peroneus longus electromyographic amplitude. However, only the modified prefabricated orthosis changed peroneus longus electromyographic amplitude towards a pattern observed with normal-arched feet. Otherwise, few differences were found between the modified prefabricated and customised orthoses. Further research is required to determine whether these changes in muscle function are associated with clinical outcomes.