Compute CCEP responses
How to compute the average response to a stimulation train extracted from a SEEG file - the cortico-cortical evoked potential used to probe how one brain region influences another.
Find the stimulation train in the recording, exclude the channels and segments the stimulation artifact makes unusable, apply a montage and filtering, then average the responses across pulses to get the evoked potential.
The methods and software used in this tutorial have not been certified for clinical practice, and should be considered for research purposes only.
What a CCEP tells you
Stimulating at one electrode and recording the response at others shows how the stimulated region connects to the rest of the implanted network. Because a single pulse produces a response buried in ongoing activity, the response is averaged across the pulses in a stimulation train - which is what makes it visible.
What you will need
Applications - Brainstorm, from the Applications catalogue.
Data - the CCEP tutorial dataset. Copy it into your own workspace before you start.
Prior work - electrode positions, so responses can be attributed to locations. See Place SEEG electrodes with Brainstorm or Localize electrodes with CiCLONE.
The workflow
Import the recording
Load the SEEG file containing the stimulation train.
Mark the channels
Identify the stimulating channels and mark bad channels so they are excluded from what follows.
Detect the stimulation artifact
Locate each pulse in the train from its artifact - this is what the averaging is aligned to.
Apply a montage
Re-reference the recording, typically to a bipolar montage, so responses are local rather than dominated by the reference.
Filter the signal
Apply the frequency filtering the analysis calls for, without removing the response itself.
Average across pulses
Average the epochs around each pulse to produce the evoked potential, and review it per channel.
Where to go next
Simulate seizures with The Virtual Brain moves from measured responses to modelling.