Working with sEEG at NIL

We do experiments and record our data with patients that are implanted with the electrodes for the treatment of epilepsy. The electrodes are implanted by a neurosurgeon (for example Pieter) and remain implanted for up to three weeks. During this period, patients wait until they had enough seizures to reliably identify the seizure origin and determine whether it is a candidate for resection. We piggyback on this procedure and conduct experiments with these patients during that waiting period. By making the experiments as fun as possible, we ensure VERY high compliance and participation rates.

Link on sEEG

• This article by Pieter and others describes the procedure.
• Read this great study about intracranial EEG in general.
• This article (almost a book) is a fantastic primer on intracranial EEG.
• In this article we describe why we believe sEEG is excellent for BCI.
• We also really like this article about the different signals measured with electric neural recordings.

Running experiments

We run our experiment on several locations, but mostly at Kempenhaeghe. Refer to Recoriding at Kempenheaghe for more information

Creating new experiments

Experiments are usually implemented in python, even though LSL supports a lot of different programming languages. For simple experiments, we often use tkinter (tutorial here). Check out https://github.com/c-herff/LSL-Experiments for some example experiments. This script connects to a marker stream and prints them out and is used for debugging. For more complex experiments, we have used Unity in the past.

To synchronize recordings between neural data and experiment design, we use LabStreamingLayer as our Backend. LSL also allows us to synchonize other modalities, such as microphone recordings, video recordings and even eyetrackers, as well.

Data analysis

To get started on the data processing and analyis see: Getting Started scripts.

We have explored re-referencing a lot, read Maxime's findings here: re-referencing

Anatomical labeling

To visualize electrodes, we co-register pre-surgical MR (T1-weighted) and post-surgical CT images. This is done using img_pipe.

img_pipe requires freesurfer. To get that running on Windows (via the Windows Subsystem for Linux) you can use this script https://github.com/neurotechcenter/Freesurfer4Windows. If you're unfamiliar with linux, this could help: https://maker.pro/linux/tutorial/basic-linux-commands-for-beginners. Jeremy (former lab-member) wrote a tutorial on how to get it running on google-cloud servers.

Currently, members of our lab are assigned to do the imaging for each patient. Therefore, you don't need to do this yourself nor install freesurfer.