The feasibility of producing a visual prosthesis for the blind using intracortical microstimulation (ICMS) of the visual cortex was studied in a 42-year-old woman who had been totally blind for 22 years secondary to glaucoma. Thirty-eight microelectrodes were implanted in the right visual cortex, near the occipital pole, for a period of 4 months. Percepts reported as small spots of light, called phosphenes, were produced with 34 of the 38 implanted microelectrodes. Threshold currents for phosphene generation with trains of biphasic pulses were as low as 1.9 microA, and most of the microelectrodes had thresholds below 25 microA. Phosphene brightness could be modified with stimulus amplitude, frequency and pulse duration. Repeated stimulation over a period of minutes produced a gradual decrease in phosphene brightness. Phosphenes did not flicker. The apparent size of phosphenes ranged from a "pin-point' to a "nickel' (20 mm diameter coin) held at arm's length. Phosphene size usually decreased as stimulation current was increased but increased slightly as the train length (TL) was increased. At levels of stimulation near threshold, the phosphenes were often reported to have colours. As the stimulation level was increased, the phosphenes generally became white, greyish or yellowish. Individual phosphenes appeared at different distances from the subject. When two phosphenes were simultaneously generated, the apparent distances of the individual phosphenes sometimes changed to make them appear to be at about the same distance. When three or more phosphenes were simultaneously generated, they became coplanar. Except for rare occasions, phosphenes extinguished rapidly at the termination of the stimulation train. When stimulation TLs were increased beyond 1 s, phosphenes usually disappeared before the end of the train. The duration of phosphene perception could be increased by interrupting a long stimulation train with brief pauses in stimulation. Intracortical microelectrodes spaced 500 microns apart generated separate phosphenes, but microelectrodes spaced 250 microns typically did not. This two-point resolution was about five times closer than has typically been achieved with surface stimulation. With some individual microelectrodes, a second closely spaced phosphene was sometimes produced by increasing the stimulation current. Phosphenes moved with eye movements. When up to six phosphenes were simultaneously elicited, they all moved with the same relative orientation during eye movements. All phosphenes were located in the left hemi-field with the majority above the horizontal meridian. There was a clustering of most of the phosphenes within a relatively small area of visual space. The potentially greater microelectrode density and lower power requirements of ICMS compared with surface stimulation appears encouraging for a visual prosthesis. However, further studies with blind subjects are required to optimize stimulation parameters and test complex image recognition before the feasibility of a visual prosthesis based on ICMS can be established.