Fig. 4: The “decision tree” ranking method.

This figure shows an example of the “selection decision tree” online ranking technique which uses the two or three recording channels with the highest beta feature value to select the two most likely best contact-levels. A For this method the clinician performs a BrainSense™ Survey for the contact-levels using the clinician programmer. These contact-level recordings can thereafter be used on-screen to visually select the channel (contact pair) with the highest beta peak (i.e. channel “0 to 2” in this example). B This information is used to choose the appropriate “selection decision tree”, i.e. the one which starts with the selected channel. Hereafter, the channel with the second highest beta peak (here channel “1 to 2”) is selected in the first branch of the decision tree. Finally, the channel with the third highest beta peak (here channel “0 to 3”) is selected in the second and final branch. The final block highlights the most promising contact-levels, identified here as levels 0 and 1. The large power between contact pairs 0–2 and 1–2 (highest and second-highest) could be due to either high power on contact 2 (and low on 0 and 1), or high power on contacts 0 and 1 (and low on 2). Because the third highest power is on contact pairs 0–3, this indicates that the large power is on 0 (and thus not on 2), only leaving out 0 and 1. These conclusions are derived using the “selection decision tree” technique only.