Key points Sudden unloading of preloaded wrist muscles elicits motion to a new wrist position. Such motion is prevented if subjects unload muscles using the contralateral arm (self‐unloading). Corticospinal influences originated from the primary motor cortex maintain tonic influences on motoneurons of wrist muscles before sudden unloading but modify these influences prior to the onset and until the end of self‐unloading. Results are interpreted based on the previous finding that intentional actions are caused by central, particularly corticospinal, shifts in the spatial thresholds at which wrist motoneurons are activated, thus predetermining the attractor point at which the neuromuscular periphery achieves mechanical balance with environment forces. By maintaining or shifting the thresholds, descending systems let body segments go to the equilibrium position in the respective unloading tasks without the pre‐programming of kinematics or muscle activation patterns. The study advances the understanding of how motor actions in general, and anticipation in particular, are controlled. Abstract The role of corticospinal (CS) pathways in anticipatory motor actions was evaluated using transcranial magnetic stimulation (TMS) of the primary motor cortex projecting to motoneurons (MNs) of wrist muscles. Preloaded wrist flexors were suddenly unloaded by the experimenter or by the subject using the other hand (self‐unloading). After sudden unloading, the wrist joint involuntarily flexed to a new position. In contrast, during self‐unloading the wrist remained almost motionless, implying that an anticipatory postural adjustment occurred. In the self‐unloading task, anticipation was manifested by a decrease in descending facilitation of pre‐activated flexor MNs starting ∼72 ms before changes in the background EMG activity. Descending facilitation of extensor MNs began to increase ∼61 ms later. Conversely, these influences remained unchanged before sudden unloading, implying the absence of anticipation. We also tested TMS responses during EMG silent periods produced by brief muscle shortening, transiently resulting in similar EMG levels before the onset and after the end of self‐unloading. We found reduced descending facilitation of flexor MNs after self‐unloading. To explain why the wrist excursion was minimized in self‐unloading due to these changes in descending influences, we relied on previous demonstrations that descending systems pre‐set the threshold positions of body segments at which muscles begin to be activated, thus predetermining the equilibrium point to which the system is attracted. Based on this notion, a more consistent explanation of the kinematic, EMG and descending patterns in the two types of unloading is proposed compared to the alternative notion of direct pre‐programming of kinematic and/or EMG patterns.