event-related synthetic aperture magnetometry (samerf)
DESCRIPTION
Event-related Synthetic Aperture Magnetometry (SAMerf). Outline. Review of traditional SAM Introduction to SAMerf Cheyne et al. motor experiment Sliding window SAMerf. Review of traditional SAM. Estimates equivalent current dipole source power within specified frequency bands - PowerPoint PPT PresentationTRANSCRIPT
Event-related Synthetic Aperture Magnetometry
(SAMerf)
Outline
• Review of traditional SAM
• Introduction to SAMerf
• Cheyne et al. motor experiment
• Sliding window SAMerf
Review of traditional SAM
• Estimates equivalent current dipole source power within specified frequency bands
• Based on sensor covariance in time windows
• Uses optimal spatial filters to estimate source power on a grid of voxels
SAM analysis of an n-back working memory task
How do we increase temporal resolution?
• Sliding window SAM– Calculate SAM images for small overlapping
time windows
• Virtual channels– Use the SAM spatial filters to estimate time
series
Virtual Channels
How can we increase signal-to-noise in a virtual channel?
• Average - either in the temporal or frequency domain
• Averaging in time will produce an evoked response (and ignore induced activity)
Introduction to SAMerf
• Traditional SAM is performed on time windows and frequency bands of interest
• Virtual channels are created for each voxel
• The virtual channels are averaged to generate an event-related response (ERF)
• Amplitudes of the ERFs at small time windows are used to produce 3D maps
Cheyne et al. motor experimentSpatiotemporal mapping of cortical activity accompanying voluntary movements using an event-related beamforming approach
Douglas Cheyne, Leyla Bakhtazad, William Gaetz
Neuromagnetic Imaging Laboratory, Hospital for Sick Children Research Institute, Toronto, Ontario, Canada
Hum Brain Mapp. 2006 Mar;27(3):213-29.
SAMerf processing stream
Movement-related fields
Single subject
Mean of 8 subjects
Sliding window SAMerf
• Activity is averaged over a small time window
• The averaging window is slid to observe temporal changes
• Good for increasing signal-to-noise and characterizing high frequency bursts
Low frequency component High frequency component
5 Clicks – 250ms ISI 4-6 seconds between click trains
Stimuli
Five Click Auditory Experiment
Time-Frequency analysis using the Stockwell Transform
Sliding window SAM with 50ms windows and 25ms steps Gamma-band: 25-50Hz
left
right
Right IFG
Summary
• Use traditional SAM to find power changes in frequency bands
• Use SAMerf to localize evoked fields and phase-locked activity
Thanks!