Antagonist muscle co-activation is thought to be partially regulated by cortical influences, but direct motor cortex involvement is not fully understood. Corticomuscular coherence (CMC) measures direct functional coupling of the motor cortex and muscles. As antagonist co-activation differs according to training status, comparison of CMC in agonist and antagonist muscles and in strength-trained and endurance-trained individuals may provide in-depth knowledge of cortical implication in antagonist muscle co-activation. Electroencephalographic and electromyographic signals were recorded, while 10 strength-trained and 11 endurance-trained participants performed isometric knee contractions in flexion and extension at various torque levels. CMC magnitude in 13–21 and 21–31 Hz frequency bands was quantified by CMC analysis between Cz electroencephalographic electrode activity and all electromyographic signals. CMC was significant in both 13–21 and 21–31 Hz frequency bands in flexor and extensor muscles regardless of participant group, torque level, and direction of contraction. CMC magnitude decreased more in antagonist than in agonist muscles as torque level increased. Finally, CMC magnitude was higher in strength-trained than in endurance-trained participants. These findings provide experimental evidence that the motor cortex directly regulates both agonist and antagonist muscles. Nevertheless, the mechanisms underlying muscle activation may be specific to their function. Between-group modulation of corticomuscular coherence may result from training-induced adaptation, re-emphasizing that corticomuscular coherence analysis may be efficient in characterizing corticospinal adaptations after long-term muscle specialization.