In the medical field, stereoscopic applications are present in diagnosis, pre-operative planning, minimally invasive surgery, instruction, and training. The use of stereoscopic applications has afforded new ways to interact with patient data, such as immersive virtual environments. This increased usage of stereoscopic applications also raises many basic research questions on human perception and performance.

Current studies show mixed results on the benefits of stereoscopic applications with regards to general performance. The benefits depend on the specific task as well as the application domain. The work presented here attempts to answer the general question: How would adding the stereopsis depth cue affect the performance of visual spatial tasks in a medical context? Visual spatial tasks are needed in medicine to understand the relationships between shapes and organs for a variety of activities in patient diagnosis and treatment.

The general research question was decomposed into specific hypotheses and three studies were conducted to study them. These studies measured performance of a visual spatial computer task using medical imaging data. Participants assessed the relative positions of three different objects located inside a 3D volumetric representation of a patient's anatomy. The first study consisted of static views and recognition of the position of color objects. The second study consisted of static views using gray objects. The third study consisted of animated views of color objects. In all three studies the task was basically the same: To select which of two objects was closest to a reference object. In all three studies participants were first and second year medical students.

Thirty-four participants completed the first study. The results of this study showed some emerging patterns in which the stereoscopic display condition had a positive benefit on performance. The stereoscopic condition had a positive effect on performance for the most difficult cases but did not yield higher results under every case and condition.

The second study, completed by 44 participants, showed the stereoscopic condition had a positive benefit on performance in 20 out of the 40 tasks completed. These 40 tasks were divided into four cases, with varying degrees of difficulty, depending on the distances between the objects being judged (i.e. cylinders in this study). At distances between 5-15 mm, the stereoscopic condition yielded statistically significant higher performance. At other distance ranges, while stereopsis showed improvement it was not statistically significant.

Thirty-one participants completed the third study. These participants completed a visual spatial task with the addition of an animation to the volume. This allowed the representation to be viewed from multiple angles before the task was completed. Overall the stereoscopic condition had a benefit in performance over the monoscopic condition. As in the previous studies tasks that had the objects between 5 - 15 mm apart had higher performance in the stereoscopic condition. Females performance in the stereoscopic condition was higher and statistically significant than for the monoscopic condition. Participants over 25 years also had a statistically significant higher performance under the stereoscopic condition. It was also observed that the stereoscopic condition did not outperform the monoscopic one in every condition.

The results of these studies show that, in general, stereopsis has a positive benefit in performance for visual spatial tasks in medical contexts. This benefit certainly has a relationship with the difficulty of the task as well as age and gender. These initial insights are a step into further work to help generate design guidelines when developing stereoscopic applications.