460 AM J OPTOM & PHYSIOL OPTICS

Vol. 65, No. 6

Fig. 5. Convergence performance of two representative patients who had failed traditional convergence training. AP had a divergence excess type of intermittent exotropia, whereas PM had a convergence insufficiency. Both patients experienced an immediate improvement in convergence ranges with exposure to automated convergence training and a concurrent elimination of their deviation. (Reprinted from Am J Optom Physiol Opt 1980;57:209.)

funduscopy were unremarkable. Distance cover testing indicated orthophoria, whereas near cover testing revealed a 10^ intermittent exotropia (exotropia occurring about 80% of the time). Nearpoint of convergence (NPC) was 25/46 cm (10/18 in), with diplopia occurring during deviation. Stereopsis was 20 sec arc on the Randot test (old version). Phorometric findings were not possible due to communication problems.

Topper vectogram using localization as a response cue for binocularity revealed base-out (BO) X/3/1^ and base-in (B!) X/7/3^. Computer Orthoptics testing of subjective angle measured

10^ of exophoria; a unilateral cover test done during subjective alignment indicated align ment, i.e., normal retinal correspondence. Ver gence measurement with RDS revealed BO ranges of X/4/1 and BI ranges of X/12/6. Worth 4 dot testing indicated crossed diplopia with no evidence of suppression. Monocular accommodative facility testing with the Computer Orthopter using +2.00/-2.00 flippers was 11 cpm.

Therapy began with vectograms, pencil push ups, Keystone stereograms, and RDS vergence programs of the Computer Orthopter. She could not maintain attention nor respond accurately to Vectograms, Keystone stereograms, or syn optophore targets. Therefore, RDS were used exclusively to increase both BO and BI vergence ranges. The RDS were presented with a small stereoscopic square in one of four positions, i.e., top, bottom, left, or right. Correct responses, using a joystick to indicate location of the stereoscopic square, resulted in a beep for positive reinforcement. This was associated with a concurrent increase in vergence demand. As long as correct responses were made, indicating binocular fusion, vergence ranges increased by 1^ per response. However, loss of fusion produced inability to perceive the RDS as indicated by an inappropriate joystick response. Incorrect responses were recorded by an inappropriate joystick response or lack of response within a 6-s time period (time error). Whenever an error occurred auditory feedback was delivered (boop sound) and the vergence demand decreased by 2^. This self-motivating system was effective in maintaining attention, and in improving convergence and divergence ability. Fusional range improved to BO = 55^ and B! = 2O^ in 5 30-min sessions. An additional three sessions were given to ensure retention of this new skill.

After this a RDS jump duction program was used. The initial setting was BO = 5^ and BI = 0^. The patient first made a response to the BO vergence demand for 5^, then the B! target was presented at 0^. The next stimulus presentation was at 6^ BO followed by 1^ B!. As long as the response was correct the jump duction task was automatically incremented by 1^; i.e., 7 BO, 2 B!, 8 BO, 3 B!, 9 BO, 4 B!, etc. If an error was made, the vergence demand was decreased by 2^, thus ensuring correct binocular responding. A decreasing vergence demand only occurred on the side on which error was made. This method resulted in an incremental jump duction. The patient was able to obtain large jump ductions on the order of BO = 45^ and B! = l5^.

These two RDS procedures produced large sustained (ramp) and large step vergence ability. The last phase of training utilized rapid vergence changes. In one such task, a flat fusion