Methods
Experiments were carried out on four adult male cats. The animals were trained to perform target-reaching to a tray with a morsel of food presented through a hole in the frontal wall. The conditioning stimulus was a tone (70dB, 100ms duration) from a loudspeaker behind the wall. The tray appeared 450ms after the conditioning stimulus, but was only available for 250 ms so the cat was required to start the movement before its appearance. Single unit activity was recorded in the forelimb representation of area 4G

Results
A total number of 350 cells were investigated. Three categories of neurones were found. The first type showed a response with two components. The first component was time-locked to the stimulus with latencies ranging from 20 to 100 msec in different cells - a high correlation of the response latency with the onset of the sensory cue (r>0.8) allows to consider them as "sensory responses". The second component was time-locked to the movement. It was fused with the first one but could be well identified by a much higher discharge frequency. The second type of neurones had single component discharges with latencies from 150 to 300 ms correlated both to the onset of the stimulus and to the movement (r>0.7). The third type showed responses with latencies more than 300 ms time-locked only to the movement.
The cells of the first type were investigated in a discrimination task in which the animals were required to initiate or withhold the movement depending on the frequency of the tone (Go/No-Go task). For stimuli implying withholding of the movement (No-Go signal) the responses were absent or substantially decreased. When indifferent acoustical stimuli were presented, there was no response, but it appeared successively if such stimuli were followed by presentation of the reward. These findings show that sensory responses are conditional in origin and are selective for the cue relevant for the movement. The response latencies of different cells suggest that the conditioning stimulus initiates a progressive recruitment of motor cortical neurones during preparation for the movement and that responses correlated to the sensory cue in cells activated early, are transformed into a movement-correlated response in cells activated later.

Conclusions
Neurones in the cat motor cortex are not only participating in movement execution but are also involved in the early stage of sensorimotor transformation related to identification of the sensory cue as a conditioning stimulus for the movement.