We have previously shown that perceptual judgements of motion direction are based in part on the activity of direction selective neurons in extrastriate visual area MT (Salzman et al., 1990, 1992). In those experiments, we applied low-amplitude microstimulation pulses (10 microA, 200 Hz) to clusters of MT neurons whose preferred directions were similar. The effect of microstimulation was to bias the monkeys' choices on a direction discrimination task toward the preferred direction of neurons at the stimulation site. The results suggest that microstimulation generated a directionally specific cortical signal by activating selectively neurons near the electrode tip. To test this notion more directly, we have now examined the behavioral effects of varying current amplitude, current frequency, and electrode position. In the majority of experiments, the directional bias in the monkeys' choices was reduced or eliminated as current amplitude increased to 80 microA. In addition, 80 microA stimulating pulses frequently impaired overall performance as measured by the percentage of correct responses. This decrement in performance indicated that 80 microA pulses introduced “noise” into the neural circuitry encoding motion direction, presumably by increasing current spread to activate a larger population of neurons representing all directions of motion. In contrast, increasing current frequency to 500 Hz (10 microA pulses) preserved the directional specificity of microstimulation effects. The precise position of the stimulating electrode also influenced the magnitude of microstimulation effects; in some cases, differences in position on the order of 100 microns determined whether an experiment yielded a very large effect or no effect at all. Thus, directionally specific activation of cortical circuitry within MT can be disrupted by increases in current spread or by small changes in electrode position. These observations suggest that the effects of low-amplitude microstimulation depend upon direct activation of a well-localized population of neurons.