fMRI: Biological Basis and Experiment Design Lecture 19: Data formats and data path
Post on 14-Jan-2016
30 Views
Preview:
DESCRIPTION
Transcript
fMRI: Biological Basis and Experiment DesignLecture 19: Data formats and data path
• DICOM• DICOM servers• Conversion between
formats• Slice-time correction and
trend removal
1 light year = 5,913,000,000,000 miles?
DICOM – it's not just a file format
Typical directory after a scan
Mosaicstime
MR-ST001-SE005-0007.dcmMR-ST001-SE005-0006.dcm
MR-ST001-SE005-0005.dcmMR-ST001-SE005-0004.dcm
MR-ST001-SE005-0003.dcmMR-ST001-SE005-0002.dcm
MR-ST001-SE005-0001.dcm
Getting the data home
• Plan A: push data to a convenient DICOM server in your department
• Plan B: push data to Drei (PC in console room); zip scanning session together and ftp it home.
• Plan C: burn CDs with data (directly from scanner, or from Drei)
Useful file formats and analysis packages
• Analyze (.hdr/.img pair)
• Nifti (.hdr/.img pair)
• ???
• .fmr; .amr; .vmr
• ???
• ???
• SPM
• FSL
• AFNI
• BrainVoyager
• Freesurfer
• AIR, NIS
Converting between formats
• Some packages will read DICOM and convert to their preferred format
• Some command line tools exist (e.g. dinifti from NYU)
• Some tools exist (e.g. MRIcro)
Typical pre-processing path
• Slice-time correction• Motion compensation• Normalization by mean• Trend removal
– Linear?– High-pass filter?
Slice-time correctionTR
1 3 5 2 4Slice order: 1 3 5 2 4 1 3 5 2 4 ...
Uncorrected data: erroneously assign acquired signal to start of TR
Corrected data: interpolate to estimate signal at start of TR
True signal
Interpolate to estimate signal at start of TR
Trend removal - frequency domain
Sglpixel
FFTStimulus response affected because cut-off was too close to stim frequency
Before After
top related