Neuronal activity in somatosensory cortex of monkeys using a precision grip. II. Responses To object texture and weights.

Three monkeys were trained to lift and hold a test object within a 12- to 25-mm position window for 1 s. The activity of single neurons was recorded during performance of the task in which both the weight and surface texture of the object were systematically varied. Whenever possible, each cell was tested with three weights (15, 65, and 115 g) and three textures (smooth metal, fine 200 grit sandpaper, and rough 60 grit sandpaper). Of 386 cells recorded in 3 monkeys, 45 cells had cutaneous receptive fields on the index or thumb or part of the thenar eminence and were held long enough to be tested in all 9 combinations of texture and weight. Recordings were made for the entire anterior-posterior extent of the thumb and index finger areas in somatosensory cortex including area 7b. However, the statistical analysis required a selection of only those cells for which nine complete recording conditions were available limiting the sample to cells in areas 2, 5, and 7b. Significant differences in the grip force accompanied 98% of the changes in texture and 78% of the changes in weight. Increasing the object weight also increased the force tangential to the skin surface as measured by the load or lifting force. The peak discharge during lifting was judged to be the most sensitive index of cell activity and was analyzed with a two-way analysis of variance (ANOVA). In addition, peak cell discharge was normalized to allow comparisons among different combinations of texture and weight as well as comparisons among different neurons. Overall, the peak firing frequency of 87% of the cells was significantly modulated by changes in object texture, but changes in object weight affected the peak activity of only 58% of the cells. Almost all (17/18, 94%) of the static cells were influenced by the object texture, and 81% of the dynamic cells that were active only briefly at grip and lift onset were modulated by texture. For some cells, surface texture had a significant effect on neuronal discharge that was independent of the object weight. In contrast, weight-related responses were never simple main effects of the weight alone and appeared instead as significant interactions between texture and weight. Four neurons either increased or decreased activity in a graded fashion with surface structure (roughness) regardless of the object weight (P < 0.05). Ten other neurons showed increases or decreases in response to one or two textures, which might represent either a graded response or a tuning preference for a specific texture. The firing frequency of the majority (31/45) of neurons reflected an interaction of both texture and weight. The cells with texture-related but weight-independent activities were thought to encode surface characteristics that are largely independent of the grip and lifting forces used to manipulate the object. Such constancies could be used to construct internal representations or mental models for planning and controlling object manipulation.