Simulation of motion on the skin. II. Cutaneous mechanoreceptor coding of the width and texture of bar patterns displaced across the OPTACON.

1. These experiments assay the functional significance of receptive-field architecture for information processing. Rapidly adapting (RA) afferents have been previously shown to converge information from clusters of 14-25 Meissner's corpuscles, whereas afferents innervating Pacinian corpuscles (PCs) have only a single, large receptor terminal. We tested two opposing hypotheses of functional architecture: 1) summation models, in which an afferent integrates signals from all of its terminals, showing monotonic increases in activity as a function of contact area, and 2) winner-take-all models, in which the most strongly activated receptor in the cluster dominates axonal output by cancellation of signals from other branches. 2. Bar and stripe patterns have been swept across the finger or palm of the monkey's hand at speeds of 30-120 mm/s with the use of a computer-controlled grid of sequentially activated miniature probes (OPTACON stimulator). The dense packing of OPTACON probes permits placement of up to five groups of stimulators within an individual receptive field, allowing us to activate one or more clusters of Meissner's corpuscles simultaneously and to stimulate the bulbar corpuscle of PC afferents at different orientations through the skin. Integration of information from moving bar patterns has been tested with two protocols. In the variable width protocol, the total number of activated rows in the pattern is varied from one to five, with a constant spacing of 1.2 mm between pulsed rows. In the variable density protocol, the length of skin stimulated is held constant at 5 mm and the spacing of stimuli varied. 3. RA afferents show no evidence of summation of inputs within their receptive fields. Motion of wide bars across the field increases the duration of firing but not the total spikes evoked by each pulse. Responses to the leading edge of wide bars were found to be identical to those evoked by a single-row bar. Simultaneous activation of two to five rows evokes the same or fewer spikes per pulse than the most effective individual row tested alone. When broad-bar patterns are centered over the field, contacting the maximum number of receptors, RAs follow activity in the dominant branch or terminus, suppressing additional inputs. Lack of summation is observed at all pulse frequencies tested (25-100 Hz). 4. Moving bar patterns evoke responses as long as at least one row stimulates the receptive field; broader patterns evoke longer spike trains whose total number of impulses is proportional to bar width.(ABSTRACT TRUNCATED AT 400 WORDS)