In the previous paper regarding the somatosensory control of the human precision grip, we concluded that the elicited automatic grip force adjustments are graded by the amplitude of the imposed loads when restraining an 'active' object subjected to unpredictable pulling forces (Johansson et al. 1992a). Using the same subjects and apparatus, the present study examines the capacity to respond to imposed load forces applied at various rates. Grip and load forces (forces normal and tangential to the grip surfaces) and the position of the object in the pulling direction (distal) were recorded. Trapezoidal load force profiles with plateau amplitudes of 2 N were delivered at the following rates of loading and unloading in an unpredictable sequence: 2 N/s, 4 N/s or 8 N/s. In addition, trials with higher load rate (32 N/s) at a low amplitude (0.7 N) were intermingled. The latencies between the start of the loading and the onset of the grip force response increased with decreasing load force rate. They were 80 +/- 9 ms, 108 +/- 13 ms, 138 +/- 27 ms and 174 +/- 39 ms for the 32, 8, 4 and 2 N/s rates, respectively. These data suggested that the grip response was elicited after a given minimum latency once a load amplitude threshold was exceeded. The amplitude of the initial rapid increase of grip force (i.e., the 'catch-up' response) was scaled by the rate of the load force, whereas its time course was similar for all load rates. This response was thus elicited as a unit, but its amplitude was graded by afferent information about the load rate arising very early during the loading. The scaling of the catch-up response was purposeful since it facilitated a rapid reconciliation of the ratio between the grip and load force to prevent slips. In that sense it apparently also compensated for the varying delays between the loading phase and the resultant grip force responses. However, modification of the catch-up response may occur during its course when the loading rate is altered prior to the grip force response or very early during the catch-up response itself Hence, afferent information may be utilized continuously in updating the response although its motor expression may be confined to certain time contingencies. Moreover, this updating may take place after an extremely short latency (45-50 ms). Our findings support the idea that the initiation as well as ongoing regulation of the motor responses is dependent on supraspinal control, but afferent signals directly processed through fast segmental networks may also contribute in the regulation. The grip force responses to the unloading phases, they were also graded by the load force rate and the response latencies increased with decreasing load force rate. However, the latencies were longer and more variable, and no catch-up responses were observed. Rather, the grip force decline was programmed for the inter-trial grip force level.
