The aim of this study was to analyze the changes in neuronal activity of the motor cortex associated with preparation for movement. Monkeys were trained to perform a between-hands choice-reaction time (RT) task. They pressed upon two levers with both hands and, after a preparatory period of 1-s duration initiated by a warning tone, they must point at either a left- or right-located target, when illuminated as a response signal, with either the left or right hand. The level of preparation for performing either sided movement was changed by manipulating, across block of trials, the relative probabilities for the left and right hands to point at the corresponding target. When considering behavioral data averaged over daily sessions, it was found that RT significantly decreased as response probability increased, whereas movement time (MT) was not affected. However, large changes in this rule were observed when data collected within a single session were considered. Response probability manipulation was acting on RT alone in 27% of sessions, on MT alone in 19% of sessions, and on both RT and MT in 17% of sessions. Statistically significant trial-by-trial correlations between RT and MT were found; of positive sign for 34% of sessions and of negative sign for 12% of sessions. Electromyographic recordings from the biceps and triceps brachii showed that muscle activity remained stable during the preparatory period, both muscles of the performing arm began to be coactivated about 70-130 ms before the lever release, biceps activation being progressively larger than that of the triceps, and the rising slope of biceps activation was steeper for short than for long RTs. Single-cell recording techniques were used to study the neuronal activity of primary motor cortex during the performance of the task. One hundred and fifteen units in area 4, whose activity was recorded during a whole experimental session, were selected for further analysis. Disregarding 4% of units which did not exhibit any movement-related change in activity, it was found that twenty-one percent of units were related to contralateral movements only, 2% to ipsilateral movements only, and 73% to both sided movements. Among this last subset, 27% of units exhibited reciprocal movement-related changes in activity, the most often an excitation for contralateral movements and an inhibition for ipsilateral movements, and 50% of units exhibited a movement-related change in activity in the same direction for both movements, the most often an excitation.(ABSTRACT TRUNCATED AT 400 WORDS)