Scout scans BIDS format & fmriprep usage

,
1

Hi,

I have some scout scans which I’m not too sure how to move forward with. I’m wondering whether I went about BIDS format correctly (eg. was it correct to name the scans as epi?). I would also appreciate some clarification on how the scans are used in preprocessing (on fmriprep).

These scans were recorded after an initial series of localizer scans and a series of MPRAGE scans. The scout scans are followed by b0 fieldmap scans (which I can use for distortion correction through the Phase-difference B0 mapping method in fmriprep) and the remaining rest and task functional scans.

Here’s my current BIDS format:

sub-01/ses-01/
    anat/
        sub-01_ses-1_acq-MPRAGE_T1w.json
        sub-01_ses-1_acq-MPRAGE_T1w.nii.gz
    fmap/
        sub-01_ses-1_epi.json
        sub-01_ses-1_epi.nii.gz
        sub-01_ses-1_magnitude1.json
        sub-01_ses-1_magnitude1.nii.gz
        sub-01_ses-1_magnitude2.json
        sub-01_ses-1_magnitude2.nii.gz
        sub-01_ses-1_phasediff.json
        sub-01_ses-1_phasediff.nii.gz
    func/
        sub-01_ses-1_task-rest_bold.json
        sub-01_ses-1_task-rest_bold.nii.gz
        sub-01_ses-1_task-learn1_bold.json
        sub-01_ses-1_task-learn1_bold.nii.gz 
        etc.

Here’s an example of a scout scan json file from my current BIDS format conversion:

{
“Modality”: “MR”,
“MagneticFieldStrength”: 3,
“ImagingFrequency”: 123.204,
“Manufacturer”: “Siemens”,
“ManufacturersModelName”: “TrioTim”,
“InstitutionName”: “MR6”,
“InstitutionAddress”: “Spruce_Street_3400_Philadelphia_3438CB_Pennsylvania_US_19104”,
“DeviceSerialNumber”: “35069”,
“StationName”: “hup6”,
“PatientPosition”: “HFS”,
“ProcedureStepDescription”: “MR_HEAD_WO_IV_CONTRAST”,
“SoftwareVersions”: “syngo_MR_B17”,
“MRAcquisitionType”: “2D”,
“SeriesDescription”: “BOLD_1.5mm_TR2000_mb4_P1_SCOUT”,
“ProtocolName”: “BOLD_1.5mm_TR2000_mb4_P1_SCOUT”,
“ScanningSequence”: “EP”,
“SequenceVariant”: “SK_SS”,
“ScanOptions”: “PFP_FS”,
“SequenceName”: “epfid2d1_128”,
“ImageType”: [
“ORIGINAL”,
“PRIMARY”,
“M”,
“MB”,
“ND”,
“MOSAIC”
],
“SeriesNumber”: 3,
“AcquisitionTime”: “21:03:23.905000”,
“AcquisitionNumber”: 1,
“ImageComments”: “Unaliased_MB4_PE3_(remote)”,
“SliceThickness”: 1.5,
“SpacingBetweenSlices”: 1.5,
“SAR”: 0.0736174,
“EchoTime”: 0.025,
“RepetitionTime”: 2,
“FlipAngle”: 60,
“PartialFourier”: 0.625,
“BaseResolution”: 128,
“ShimSetting”: [
4042,
-6608,
-3892,
494,
58,
-30,
17,
-79
],
“TxRefAmp”: 320.939,
“PhaseResolution”: 1,
“ReceiveCoilName”: “32Ch_Head”,
“PulseSequenceDetails”: “%CustomerSeq%_cmrr_mbep2d_bold”,
“MultibandAccelerationFactor”: 4,
“PercentPhaseFOV”: 100,
“PercentSampling”: 100,
“PhaseEncodingSteps”: 80,
“AcquisitionMatrixPE”: 128,
“ReconMatrixPE”: 128,
“BandwidthPerPixelPhaseEncode”: 8.878,
“EffectiveEchoSpacing”: 0.000879984,
“DerivedVendorReportedEchoSpacing”: 0.000879984,
“TotalReadoutTime”: 0.111758,
“PixelBandwidth”: 1776,
“PhaseEncodingDirection”: “j-”,
“SliceTiming”: [
0,
0.905,
1.8075,
0.7025,
1.6075,
0.5025,
1.405,
0.3025,
1.205,
0.1,
1.005,
1.9075,
0.8025,
1.7075,
0.6025,
1.5075,
0.4025,
1.305,
0.2025,
1.105,
0,
0.905,
1.8075,
0.7025,
1.6075,
0.5025,
1.405,
0.3025,
1.205,
0.1,
1.005,
1.9075,
0.8025,
1.7075,
0.6025,
1.5075,
0.4025,
1.305,
0.2025,
1.105,
0,
0.905,
1.8075,
0.7025,
1.6075,
0.5025,
1.405,
0.3025,
1.205,
0.1,
1.005,
1.9075,
0.8025,
1.7075,
0.6025,
1.5075,
0.4025,
1.305,
0.2025,
1.105,
0,
0.905,
1.8075,
0.7025,
1.6075,
0.5025,
1.405,
0.3025,
1.205,
0.1,
1.005,
1.9075,
0.8025,
1.7075,
0.6025,
1.5075,
0.4025,
1.305,
0.2025,
1.105
],
“ImageOrientationPatientDICOM”: [
1,
0,
0,
0,
1,
0
],
“InPlanePhaseEncodingDirectionDICOM”: “COL”,
“ConversionSoftware”: “dcm2niix”,
“ConversionSoftwareVersion”: “v1.0.20200331”,
“Dcm2bidsVersion”: “2.1.4”,
“IntendedFor”: []
}

Thank you!

2

How are the scout scans different from the localizer scans? I see that the scouts use the CMRR EPI sequence and aren’t single-band reference images, but beyond that I’m not sure what they’re for. Do they use the opposite phase-encoding direction of the BOLD scans?

3

Historically, the word ‘scout’ was used by Siemens as a synonym for anatomical localizer. This is a fMRI sequence, not a traditional localizer, but my assumption is that the intention was to acquire a very brief series to check the anatomical coverage and artifacts prior to acquiring a subsequent long fMRI series.

@tsalo, as you note, this is not a single band reference. In addition, it is not a reversed phase-encoding polarity series. Specifically, for TOPUP one must spin-echo sequences, whereas this is a gradient-echo sequence.

Unless this is a long run for a fMRI task used as a functional localizer (a functional scout for subsequent runs), I think these scans should be discarded.

4

Thank you for looking into this! As Chris pointed out, I confused anatomical with functional localizers. I believe the functional localizers aren’t specifically being used for recording task performance so I will be discarding these scans.