Validation of structural brain connectivity networks

The impact of scanning parameters

Karen S. Ambrosen, Simon F. Eskildsen, Max Hinne, Kristine Krug, Henrik Lundell, Mikkel N. Schmidt, Marcel A.J. van Gerven, Morten Mørup, Tim B. Dyrby*

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

Evaluation of the structural connectivity (SC) of the brain based on tractography has mainly focused on the choice of diffusion model, tractography algorithm, and their respective parameter settings. Here, we systematically validate SC derived from a post mortem monkey brain, while varying key acquisition parameters such as the b-value, gradient angular resolution and image resolution. As gold standard we use the connectivity matrix obtained invasively with histological tracers by Markov et al. (2014). As performance metric, we use cross entropy as a measure that enables comparison of the relative tracer labeled neuron counts to the streamline counts from tractography. We find that high angular resolution and high signal-to-noise ratio are important to estimate SC, and that SC derived from low image resolution (1.03 mm3) are in better agreement with the tracer network, than those derived from high image resolution (0.53 mm3) or at an even lower image resolution (2.03 mm3). In contradiction, sensitivity and specificity analyses suggest that if the angular resolution is sufficient, the balanced compromise in which sensitivity and specificity are identical remains 60–64% regardless of the other scanning parameters. Interestingly, the tracer graph is assumed to be the gold standard but by thresholding, the balanced compromise increases to 70–75%. Hence, by using performance metrics based on binarized tracer graphs, one risks losing important information, changing the performance of SC graphs derived by tractography and their dependence of different scanning parameters.

Original languageEnglish
Article number116207
JournalNeuroImage
Volume204
Number of pages13
ISSN1053-8119
DOIs
Publication statusPublished - 1 Jan 2020

Cite this

Ambrosen, Karen S. ; Eskildsen, Simon F. ; Hinne, Max ; Krug, Kristine ; Lundell, Henrik ; Schmidt, Mikkel N. ; van Gerven, Marcel A.J. ; Mørup, Morten ; Dyrby, Tim B. / Validation of structural brain connectivity networks : The impact of scanning parameters. In: NeuroImage. 2020 ; Vol. 204.
@article{6ad91fe0da634661a9ae332d0a7fdb99,
title = "Validation of structural brain connectivity networks: The impact of scanning parameters",
abstract = "Evaluation of the structural connectivity (SC) of the brain based on tractography has mainly focused on the choice of diffusion model, tractography algorithm, and their respective parameter settings. Here, we systematically validate SC derived from a post mortem monkey brain, while varying key acquisition parameters such as the b-value, gradient angular resolution and image resolution. As gold standard we use the connectivity matrix obtained invasively with histological tracers by Markov et al. (2014). As performance metric, we use cross entropy as a measure that enables comparison of the relative tracer labeled neuron counts to the streamline counts from tractography. We find that high angular resolution and high signal-to-noise ratio are important to estimate SC, and that SC derived from low image resolution (1.03 mm3) are in better agreement with the tracer network, than those derived from high image resolution (0.53 mm3) or at an even lower image resolution (2.03 mm3). In contradiction, sensitivity and specificity analyses suggest that if the angular resolution is sufficient, the balanced compromise in which sensitivity and specificity are identical remains 60–64{\%} regardless of the other scanning parameters. Interestingly, the tracer graph is assumed to be the gold standard but by thresholding, the balanced compromise increases to 70–75{\%}. Hence, by using performance metrics based on binarized tracer graphs, one risks losing important information, changing the performance of SC graphs derived by tractography and their dependence of different scanning parameters.",
author = "Ambrosen, {Karen S.} and Eskildsen, {Simon F.} and Max Hinne and Kristine Krug and Henrik Lundell and Schmidt, {Mikkel N.} and {van Gerven}, {Marcel A.J.} and Morten M{\o}rup and Dyrby, {Tim B.}",
year = "2020",
month = "1",
day = "1",
doi = "10.1016/j.neuroimage.2019.116207",
language = "English",
volume = "204",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Elsevier",

}

Validation of structural brain connectivity networks : The impact of scanning parameters. / Ambrosen, Karen S.; Eskildsen, Simon F.; Hinne, Max; Krug, Kristine; Lundell, Henrik; Schmidt, Mikkel N.; van Gerven, Marcel A.J.; Mørup, Morten; Dyrby, Tim B.

In: NeuroImage, Vol. 204, 116207, 01.01.2020.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Validation of structural brain connectivity networks

T2 - The impact of scanning parameters

AU - Ambrosen, Karen S.

AU - Eskildsen, Simon F.

AU - Hinne, Max

AU - Krug, Kristine

AU - Lundell, Henrik

AU - Schmidt, Mikkel N.

AU - van Gerven, Marcel A.J.

AU - Mørup, Morten

AU - Dyrby, Tim B.

PY - 2020/1/1

Y1 - 2020/1/1

N2 - Evaluation of the structural connectivity (SC) of the brain based on tractography has mainly focused on the choice of diffusion model, tractography algorithm, and their respective parameter settings. Here, we systematically validate SC derived from a post mortem monkey brain, while varying key acquisition parameters such as the b-value, gradient angular resolution and image resolution. As gold standard we use the connectivity matrix obtained invasively with histological tracers by Markov et al. (2014). As performance metric, we use cross entropy as a measure that enables comparison of the relative tracer labeled neuron counts to the streamline counts from tractography. We find that high angular resolution and high signal-to-noise ratio are important to estimate SC, and that SC derived from low image resolution (1.03 mm3) are in better agreement with the tracer network, than those derived from high image resolution (0.53 mm3) or at an even lower image resolution (2.03 mm3). In contradiction, sensitivity and specificity analyses suggest that if the angular resolution is sufficient, the balanced compromise in which sensitivity and specificity are identical remains 60–64% regardless of the other scanning parameters. Interestingly, the tracer graph is assumed to be the gold standard but by thresholding, the balanced compromise increases to 70–75%. Hence, by using performance metrics based on binarized tracer graphs, one risks losing important information, changing the performance of SC graphs derived by tractography and their dependence of different scanning parameters.

AB - Evaluation of the structural connectivity (SC) of the brain based on tractography has mainly focused on the choice of diffusion model, tractography algorithm, and their respective parameter settings. Here, we systematically validate SC derived from a post mortem monkey brain, while varying key acquisition parameters such as the b-value, gradient angular resolution and image resolution. As gold standard we use the connectivity matrix obtained invasively with histological tracers by Markov et al. (2014). As performance metric, we use cross entropy as a measure that enables comparison of the relative tracer labeled neuron counts to the streamline counts from tractography. We find that high angular resolution and high signal-to-noise ratio are important to estimate SC, and that SC derived from low image resolution (1.03 mm3) are in better agreement with the tracer network, than those derived from high image resolution (0.53 mm3) or at an even lower image resolution (2.03 mm3). In contradiction, sensitivity and specificity analyses suggest that if the angular resolution is sufficient, the balanced compromise in which sensitivity and specificity are identical remains 60–64% regardless of the other scanning parameters. Interestingly, the tracer graph is assumed to be the gold standard but by thresholding, the balanced compromise increases to 70–75%. Hence, by using performance metrics based on binarized tracer graphs, one risks losing important information, changing the performance of SC graphs derived by tractography and their dependence of different scanning parameters.

U2 - 10.1016/j.neuroimage.2019.116207

DO - 10.1016/j.neuroimage.2019.116207

M3 - Journal article

VL - 204

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

M1 - 116207

ER -