Assessment of the capacity of human subjects and S-I neurons to distinguish opposing directions of stimulus motion across the skin.

The ability of human subjects and the capacities of single S-I neurons of macaque monkeys to distinguish opposing directions of movement over the skin were investigated by employing experimental paradigms and data analyses based on sensory decision theory (SDT). It is shown that these techniques can be utilized to provide behavioral and neurophysiological indices of directional sensitivity which have the same metric, and are amenable to statistical tests for significance. The influences of 3 different paradigms and modes of relative operating characteristic (ROC) curve construction on SDT indices of human cutaneous directional sensitivity were investigated. Response latency (RL) was used as an objective indication of certainty in all 3 paradigms; in one of the 3 paradigms the subject also rated the certainty of each report. The SDT indices of cutaneous directional sensitivity and response bias were shown to be independent of the paradigm and mode of ROC curve construction investigated, and the SDT 'Gaussian-equal variance' hypothesis was concluded to be consistent with the data provided by all 3 paradigms. A considerable amount of inter-subject as well as intra-subject variability in human cutaneous directional sensitivity is demonstrated for all subjects tested. This variability appears to be an attribute of the processes underlying the sensing of stimulus direction since it is present even when stimulus conditions are maintained constant. Experimental designs were developed which account for this variability, thus allowing detection and quantitation of the influence of variations in stimulus conditions on human directional sensitivity. It is demonstrated that for S-I neurons, an ROC curve can be generated from the responses to multiple replications of opposing directions of movement across the receptive field. The large number of stimulus presentations required to estimate directional sensitivity from ROC curves involves a prolonged period of single neuron recording that is difficult to achieve even under ideal experimental conditions. It is shown that one can obtain a reliable estimate of single neuron directional sensitivity (i.e. delta'e) using relatively few stimulus replications when mean firing rate is assumed to represent that aspect of the neural response carrying information about stimulus direction. These indices allow assessment of the selectivity of single S-I neurons for direction as stimulus parameters are varied. Examples are provided which show (utilizing delta'e) that those stimulus conditions evoking maximal firing rates from S-I neurons are often not optimal for signalling direction of movement across the skin.