The brain is a complex organ that is able to perform equally complex tasks. To protect it well it is shielded by a thick skull. This is why it has been nearly impossible to study the living brain for a long time.
Luckily we are now able to look inside the skull using various scanning techniques. Some techniques are great for seeing anatomy while other methods are able to show brain activity.
Before the operation where the DBS lead is implanted a brainscan is made to locate the target structure in the patients brain. Although every brain looks more or less the same, just like we all have our face organized in roughly the same manner, there still is a large variation across all subjects. Since – for DBS especially – every millimeter counts, the pre-operative images are crucial for the neurosurgeon to plan a succesful therapy.
A post-operative scan is made to see where exactly the DBS implant is placed and if the target was hit. Although this brainscan is not required for the treatment it can be very helpful for the neurologist; especially if some context is provided. Because the implanted lead has multiple electrical contacts, there are still many options to finetune the therapy after the surgery.
As you can see on the image to the right, hardly any anatomical information is visible on the post-operative image: a CT only shows hard materials, like bone and metal.
In order for these images to be useful they should be blended together, a proces that’s called fusion or registration.
Afer registering pre-op and post-op images also anatomical atlases can be registered and electrical stimulation can be simulated. This way a stack of 2D images is converted into a 3D scene.
Although an anatomical atlas can be very helpful, it doesn’t tell you where exactly the orange VTA (Volume of tissue activation) will result in the best therapy. Or where, for instance, side effects may be induced.
Therefore, it’s the focus of the visualDBSlab, to use this 3D landscape as our starting point and bring more relevant information into the scene.
As stated before:
- brain anatomy shows great variability
- millimeters matter for DBS
So in order to do studies involving more than one patient, it’s important to normalize the stimulation locations to a normative space. The MNIbrain is such a space. It has been developed at the Montreal Neurological Institute and basically is an average of many brains.
The problem we have to solve: How do get our data into the MNI brain?
A frequently used method is to stretch and squeeze a patients brain scan to make it fit the model. In our case we aren’t interested in a good normalization across the full brain. We want to have a perfect registration at a very local level. That’s why we use a local atlas registration to bridge to the MNIbrain. This way we can be sure that the position of the electrode in relation to the relevant anatomy does not get lost during the transfer.