Whole-Brain Exploratory Analysis of Functional Task Response Following Erythropoietin Treatment in Mood Disorders: A Supervised Machine Learning Approach

Søren Føns Vind Nielsen, Kristoffer Hougaard Madsen, Maj Vinberg, Lars V Kessing, Hartwig R Siebner, Kamilla W Miskowiak

Research output: Contribution to journalJournal articleResearchpeer-review

14 Downloads (Pure)

Abstract

A core symptom of mood disorders is cognitive impairment in attention, memory and executive functions. Erythropoietin (EPO) is a candidate treatment for cognitive impairment in unipolar and bipolar disorders (UD and BD) and modulates cognition-related neural activity across a fronto-temporo-parietal network. This report investigates predicting the pharmacological treatment from functional magnetic resonance imaging (fMRI) data using a supervised machine learning approach. A total of 84 patients with UD or BD were included in a randomized double-blind parallel-group study in which they received eight weekly infusions of either EPO (40 000 IU) or saline. Task fMRI data were collected before EPO/saline infusions started (baseline) and 6 weeks after last infusion (follow-up). During the scanning sessions, participants were given an n-back working memory and a picture encoding task. Linear classification models with different regularization techniques were used to predict treatment status from both cross-sectional data (at follow-up) and longitudinal data (difference between baseline and follow-up). For the n-back and picture encoding tasks, data were available and analyzed for 52 (EPO; n = 28, Saline; n = 24) and 59 patients (EPO; n = 31, Saline; n = 28), respectively. We found limited evidence that the classifiers used could predict treatment status at a reliable level of performance (≤60% accuracy) when tested using repeated cross-validation. There was no difference in using cross-sectional versus longitudinal data. Whole-brain multivariate decoding applied to pharmaco-fMRI in small to moderate samples seems to be suboptimal for exploring data driven neuronal treatment mechanisms.
Original languageEnglish
Article number1246
JournalFrontiers in Neuroscience
Volume13
Number of pages11
ISSN1662-4548
DOIs
Publication statusPublished - 2019

Keywords

  • Erythropoietin
  • Functional magnetic resonance imaging
  • Machine learning
  • Mood disorders
  • Cognitive dysfunction

Cite this

@article{665491159de943188eeeb4a9e9a4e595,
title = "Whole-Brain Exploratory Analysis of Functional Task Response Following Erythropoietin Treatment in Mood Disorders: A Supervised Machine Learning Approach",
abstract = "A core symptom of mood disorders is cognitive impairment in attention, memory and executive functions. Erythropoietin (EPO) is a candidate treatment for cognitive impairment in unipolar and bipolar disorders (UD and BD) and modulates cognition-related neural activity across a fronto-temporo-parietal network. This report investigates predicting the pharmacological treatment from functional magnetic resonance imaging (fMRI) data using a supervised machine learning approach. A total of 84 patients with UD or BD were included in a randomized double-blind parallel-group study in which they received eight weekly infusions of either EPO (40 000 IU) or saline. Task fMRI data were collected before EPO/saline infusions started (baseline) and 6 weeks after last infusion (follow-up). During the scanning sessions, participants were given an n-back working memory and a picture encoding task. Linear classification models with different regularization techniques were used to predict treatment status from both cross-sectional data (at follow-up) and longitudinal data (difference between baseline and follow-up). For the n-back and picture encoding tasks, data were available and analyzed for 52 (EPO; n = 28, Saline; n = 24) and 59 patients (EPO; n = 31, Saline; n = 28), respectively. We found limited evidence that the classifiers used could predict treatment status at a reliable level of performance (≤60{\%} accuracy) when tested using repeated cross-validation. There was no difference in using cross-sectional versus longitudinal data. Whole-brain multivariate decoding applied to pharmaco-fMRI in small to moderate samples seems to be suboptimal for exploring data driven neuronal treatment mechanisms.",
keywords = "Erythropoietin, Functional magnetic resonance imaging, Machine learning, Mood disorders, Cognitive dysfunction",
author = "Nielsen, {S{\o}ren F{\o}ns Vind} and Madsen, {Kristoffer Hougaard} and Maj Vinberg and Kessing, {Lars V} and Siebner, {Hartwig R} and Miskowiak, {Kamilla W}",
year = "2019",
doi = "10.3389/fnins.2019.01246",
language = "English",
volume = "13",
journal = "Frontiers in Neuroscience",
issn = "1662-4548",
publisher = "Frontiers Research Foundation",

}

Whole-Brain Exploratory Analysis of Functional Task Response Following Erythropoietin Treatment in Mood Disorders: A Supervised Machine Learning Approach. / Nielsen, Søren Føns Vind; Madsen, Kristoffer Hougaard; Vinberg, Maj; Kessing, Lars V; Siebner, Hartwig R; Miskowiak, Kamilla W.

In: Frontiers in Neuroscience, Vol. 13, 1246, 2019.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Whole-Brain Exploratory Analysis of Functional Task Response Following Erythropoietin Treatment in Mood Disorders: A Supervised Machine Learning Approach

AU - Nielsen, Søren Føns Vind

AU - Madsen, Kristoffer Hougaard

AU - Vinberg, Maj

AU - Kessing, Lars V

AU - Siebner, Hartwig R

AU - Miskowiak, Kamilla W

PY - 2019

Y1 - 2019

N2 - A core symptom of mood disorders is cognitive impairment in attention, memory and executive functions. Erythropoietin (EPO) is a candidate treatment for cognitive impairment in unipolar and bipolar disorders (UD and BD) and modulates cognition-related neural activity across a fronto-temporo-parietal network. This report investigates predicting the pharmacological treatment from functional magnetic resonance imaging (fMRI) data using a supervised machine learning approach. A total of 84 patients with UD or BD were included in a randomized double-blind parallel-group study in which they received eight weekly infusions of either EPO (40 000 IU) or saline. Task fMRI data were collected before EPO/saline infusions started (baseline) and 6 weeks after last infusion (follow-up). During the scanning sessions, participants were given an n-back working memory and a picture encoding task. Linear classification models with different regularization techniques were used to predict treatment status from both cross-sectional data (at follow-up) and longitudinal data (difference between baseline and follow-up). For the n-back and picture encoding tasks, data were available and analyzed for 52 (EPO; n = 28, Saline; n = 24) and 59 patients (EPO; n = 31, Saline; n = 28), respectively. We found limited evidence that the classifiers used could predict treatment status at a reliable level of performance (≤60% accuracy) when tested using repeated cross-validation. There was no difference in using cross-sectional versus longitudinal data. Whole-brain multivariate decoding applied to pharmaco-fMRI in small to moderate samples seems to be suboptimal for exploring data driven neuronal treatment mechanisms.

AB - A core symptom of mood disorders is cognitive impairment in attention, memory and executive functions. Erythropoietin (EPO) is a candidate treatment for cognitive impairment in unipolar and bipolar disorders (UD and BD) and modulates cognition-related neural activity across a fronto-temporo-parietal network. This report investigates predicting the pharmacological treatment from functional magnetic resonance imaging (fMRI) data using a supervised machine learning approach. A total of 84 patients with UD or BD were included in a randomized double-blind parallel-group study in which they received eight weekly infusions of either EPO (40 000 IU) or saline. Task fMRI data were collected before EPO/saline infusions started (baseline) and 6 weeks after last infusion (follow-up). During the scanning sessions, participants were given an n-back working memory and a picture encoding task. Linear classification models with different regularization techniques were used to predict treatment status from both cross-sectional data (at follow-up) and longitudinal data (difference between baseline and follow-up). For the n-back and picture encoding tasks, data were available and analyzed for 52 (EPO; n = 28, Saline; n = 24) and 59 patients (EPO; n = 31, Saline; n = 28), respectively. We found limited evidence that the classifiers used could predict treatment status at a reliable level of performance (≤60% accuracy) when tested using repeated cross-validation. There was no difference in using cross-sectional versus longitudinal data. Whole-brain multivariate decoding applied to pharmaco-fMRI in small to moderate samples seems to be suboptimal for exploring data driven neuronal treatment mechanisms.

KW - Erythropoietin

KW - Functional magnetic resonance imaging

KW - Machine learning

KW - Mood disorders

KW - Cognitive dysfunction

U2 - 10.3389/fnins.2019.01246

DO - 10.3389/fnins.2019.01246

M3 - Journal article

VL - 13

JO - Frontiers in Neuroscience

JF - Frontiers in Neuroscience

SN - 1662-4548

M1 - 1246

ER -