Neuro-Surgical Navigation System MIT Artificial Intelligence Lab and Brigham and Women's Hospital Overview We designed and implemented a surgical navigation system that has been used to guide the chief of neurosurgery at Brigham and Women's Hospital through over 200 intra-operative procedures. The system allows the surgeon to more safely and more accurately navigate through the patient's head with reference to a preoperative scan. Patient Specific Models A preoperative MR scan of the patient is acquired, consisting of many 2D slices that make up a full 3D volume (left). Internal anatomical structures are labeled or segmented semi-automatically and 3D surface models of the structures are generated (right). Registration In the operating room, a laser triangulation system extracts 3D positions from the patient's skin (left). Those 3D points are then registered or aligned to the MR skin model of the patient (right). This registration relates the patient's position in the OR to the coordinate system of the MR images. Enhanced Reality Visualization and Surgical Navigation The internal anatomy of the patient can then be superimposed on live video of the patient, giving the surgeon a type of "X-Ray Vision" to see inside the patient's head (left). The surgeon also uses trackable probes and instruments to navigate through the anatomy, while the system displays the real-time position of the tip in the 3D model and orthogonal MR slices (right). Pediatric Epilepsy In cases of pediatric epilepsy, an electrode grid is placed on the cortex to identify the location of the lesion (left). The positions of the grid points are recorded using the probe and visualized in the 3D models (right). The responses from the grid are color-coded in the 3D models providing the surgeon with a map both of the lesion to be removed and of critical areas such as the motor and sensory corticies to be avoided.