Fast, quantitative myelin maps: Macromolecular pool fraction (MPF) using an optimized protocol

Poster No:


Submission Type:

Abstract Submission 


Kimberly Desmond1, Tobias Wood2, Sofia Chavez1


1Centre for Addiction and Mental Health (CAMH), Toronto, Ontario, Canada, 2King's College, London, United Kingdom

First Author:

Kimberly Desmond, PhD  
Centre for Addiction and Mental Health (CAMH)
Toronto, Ontario, Canada


Tobias Wood, PhD  
King's College
London, United Kingdom
Sofia Chavez  
Centre for Addiction and Mental Health (CAMH)
Toronto, Ontario, Canada


The macromolecular pool fraction (MPF) has been shown to correlate with myelin (Khodanovich 2017). Whole brain MPF maps can be generated efficiently from T1, B1 and B0 maps and the collection of two additional volumes: a reference with no MT pulse (MT0) and an MT-weighted volume (MTΔ) acquired at an optimal frequency offset, Δ (Yarnykh 2012). Further time efficiency was proposed by using the T1 and B1 maps to create a synthetic MT0, synMT0 (Yarnykh 2016). In the original work, T1 maps were generated using a variable-flip-angle method from two images that were acquired with the same sequence (spoiled-gradient echo, SPGR) and resolution as MTΔ. Restrictions on TR for MT-prepared scans makes matching scans across MT and T1 mapping acquisitions not optimal. Recently, we showed that synMT can be created using a T1 mapping method relying on fastSPGR (FSPGR) scans if a calibration procedure is used to account for systematic differences across sequences (Chavez 2019). The calibration was found to be subject-independent and thus the scaling factor, f, was determined a priori and applied on subsequent synMT. However, discrepancies in bandwidth across sequences led to inaccuracies in synMPF0 in regions of large susceptibility gradients (Δχ). In this work, we optimize T1 and B1 mapping acquisitions to reduce scan time for synMT generation and we match sequences to reduce inconsistencies in regions of large Δχ.


Simulations were used to study the effect of T1 and B1 inaccuracies on MPF maps computed using the acquired MT (acqMPF) and calibrated synMT (synMPF). The errors in B1 and T1 maps are expected to be compounded for synMPF since the same values are used to compute synMT and then, in turn, MPF.

Scan time for synMPF mapping is reduced by acquiring lower resolution T1 maps and an efficient 2D-B1 mapping implementation of the double-angle-method (DAM). The modified FSPGR sequence with in-plane acceleration allows for a large reduction in scan time for MTΔ.

Four volunteers (37±10yrs) were scanned on a 3T scanner (GE Healthcare). MT0 and MTΔ were acquired using the modified FSPGR sequence; calibrated B1 and T1 maps were obtained using calibrated DAM (Chavez 2018a) and VFA (Chavez 2018b) methods, respectively. Image processing was done using FSL (FMRIB software library) and MATLAB. The scaling factor for synMT0 calibration was found as per Chavez 2019 for each subject, then averaged to yield the scaling factor applied to all synMT0. MPF maps were computed using acquired and calibrated synthetic MT0, yielding acqMPF and synMPF, respectively.
Supporting Image: Table1.png
   ·Table 1. Scanning Parameters and Timing for all Acquisitions


Simulations predict that B1 errors lead to more severe discrepancies in synMPF and acqMPF errors than T1 errors. Small T1 errors (5%) can lead to large synMPF errors (15%) regardless of true (B1,T1) while acqMPF is much less affected. In all cases, errors are more pronounced for GM values than WM values. The figure shows acqMPF and synMPF and values in WM and GM, for two subjects. The synMPF is well-matched to acqMPF in WM (<1.5% difference) which is our primary focus. In GM, synMPF is biased towards an overestimation relative to acqMPF (18%-32%). This is due to (i) the smaller GM MPF values and (ii) larger acqMPF and synMPF error discrepancies for GM T1 and MPF values (predicted by simulations)
Supporting Image: Figure1.png
   ·Figure 1. Assessing the Accuracy of synMPF Maps


We showed that the proposed optimized MPF acquisitions yield fast synMPF estimates (scan time saving of 32% relative to acqMPF). Results for synMPF are very well matched to acqMPF in WM and slightly overestimated in GM. Matching sequences for MTΔ and synMT0 improved synMPF estimates, especially in regions of large Δχ. By using faster B1 mapping and MT sequences, we were able to reduce the time to acquire the full set of images required for MPF from 15 min 46s to 10 min 38s.

Modeling and Analysis Methods:

Methods Development

Neuroanatomy, Physiology, Metabolism and Neurotransmission:

Cortical Anatomy and Brain Mapping 2

Novel Imaging Acquisition Methods:

Anatomical MRI
Imaging Methods Other 1


White Matter
Other - Magnetization Transfer

1|2Indicates the priority used for review

My abstract is being submitted as a Software Demonstration.


Please indicate below if your study was a "resting state" or "task-activation” study.


Healthy subjects only or patients (note that patient studies may also involve healthy subjects):

Healthy subjects

Was any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.


Was any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.

Not applicable

Please indicate which methods were used in your research:

Structural MRI
Computational modeling

For human MRI, what field strength scanner do you use?


Which processing packages did you use for your study?


Provide references using author date format

Chavez, S. E. (2018a). 'Assessing B1 map errors in vivo: measuring stability and absolute accuracy despite the lack of gold standard'. Proceedings of the 26th Annual Meeting of ISMRM, Paris, France.
Chavez, S. E. (2018b). 'Calibrating variable flip angle (VFA)based T1 maps: when and why a simple scaling factor is justified'. Proceedings of the 26th Annual Meeting of ISMRM, Paris, France.
Chavez, S. E. (2019). 'Producing a Synthetic Magnetization Transfer Reference Volume with T1 Maps Acquired using a Different Pulse Sequence: Practical Considerations'. Proceedings of the 27th Annual Meeting of ISMRM, Montreal, Quebec, Canada.
Khodanovich, M. Y. (2017). 'Histological validation of fast macromolecular proton fraction mapping as a quantitative myelin imaging method in the cuprizone demyelination model.' Scientific reports, vol. 7, pp. 46686.
Yarnykh, V. L. (2012). 'Fast macromolecular proton fraction mapping from a single off‐resonance magnetization transfer measurement.' Magnetic Resonance in Medicine, vol. 68, no. 1, pp. 166-178.
Yarnykh, V. L. (2016). 'Time‐efficient, high‐resolution, whole brain three‐dimensional macromolecular proton fraction mapping.' Magnetic Resonance in Medicine, vol. 75, no. 5, pp. 2100-2106.