Different multivariate techniques for automated classification of MRI data in Alzheimer's disease and mild cognitive impairment

Carlos Aguilar; Eric Westman; J-Sebastian Muehlboeck; Patrizia Mecocci; Bruno Vellas; Magda Tsolaki; Iwona Kloszewska; Hilkka Soininen; Simon Lovestone; Christian Spenger; Andrew Simmons; Lars-Olof Wahlund

doi:10.1016/j.pscychresns.2012.11.005

Different multivariate techniques for automated classification of MRI data in Alzheimer's disease and mild cognitive impairment

Carlos Aguilar, Eric Westman, J-Sebastian Muehlboeck, Patrizia Mecocci, Bruno Vellas, Magda Tsolaki, Iwona Kloszewska, Hilkka Soininen, Simon Lovestone, Christian Spenger, Andrew Simmons, Lars-Olof Wahlund

Research output: Contribution to journal › Article › peer-review

98 Citations (Scopus)

Abstract

Automated structural magnetic resonance imaging (MRI) processing pipelines and different multivariate techniques are gaining popularity for Alzheimer's disease (AD) research. We used four supervised learning methods to classify AD patients and controls (CTL) and to prospectively predict the conversion of mild cognitive impairment (MCI) to AD from baseline MRI data. A total of 345 participants from the AddNeuroMed cohort were included in this study; 116 AD patients, 119 MCI patients and 110 CTL individuals. High resolution sagittal 3D MP-RAGE datasets were acquired and MRI data were processed using FreeSurfer. We explored the classification ability of orthogonal projections to latent structures (OPLS), decision trees (Trees), artificial neural networks (ANN) and support vector machines (SVM). Applying 10-fold cross-validation demonstrated that SVM and OPLS were slightly superior to Trees and ANN, although not statistically significant for distinguishing between AD and CTL. The classification experiments resulted in up to 83% sensitivity and 87% specificity for the best techniques. For the prediction of conversion of MCI patients at baseline to AD at 1-year follow-up, we obtained an accuracy of up to 86%. The value of the multivariate models derived from the classification of AD vs. CTL was shown to be robust and efficient in the identification of MCI converters.

Original language	English
Article number	N/A
Pages (from-to)	89-98
Number of pages	10
Journal	Psychiatry Research. Neuroimaging
Volume	212
Issue number	2
DOIs	https://doi.org/10.1016/j.pscychresns.2012.11.005
Publication status	Published - 30 May 2013

Keywords

Acknowledged-BRC
Acknowledged-BRU
Acknowledged-BRU-13/14

Access to Document

10.1016/j.pscychresns.2012.11.005

Cite this

Aguilar, C., Westman, E., Muehlboeck, J.-S., Mecocci, P., Vellas, B., Tsolaki, M., Kloszewska, I., Soininen, H., Lovestone, S., Spenger, C., Simmons, A., & Wahlund, L.-O. (2013). Different multivariate techniques for automated classification of MRI data in Alzheimer's disease and mild cognitive impairment. Psychiatry Research. Neuroimaging, 212(2), 89-98. Article N/A. https://doi.org/10.1016/j.pscychresns.2012.11.005

@article{20bfbda5ea404b079e73f67647779972,

title = "Different multivariate techniques for automated classification of MRI data in Alzheimer's disease and mild cognitive impairment",

abstract = "Automated structural magnetic resonance imaging (MRI) processing pipelines and different multivariate techniques are gaining popularity for Alzheimer's disease (AD) research. We used four supervised learning methods to classify AD patients and controls (CTL) and to prospectively predict the conversion of mild cognitive impairment (MCI) to AD from baseline MRI data. A total of 345 participants from the AddNeuroMed cohort were included in this study; 116 AD patients, 119 MCI patients and 110 CTL individuals. High resolution sagittal 3D MP-RAGE datasets were acquired and MRI data were processed using FreeSurfer. We explored the classification ability of orthogonal projections to latent structures (OPLS), decision trees (Trees), artificial neural networks (ANN) and support vector machines (SVM). Applying 10-fold cross-validation demonstrated that SVM and OPLS were slightly superior to Trees and ANN, although not statistically significant for distinguishing between AD and CTL. The classification experiments resulted in up to 83% sensitivity and 87% specificity for the best techniques. For the prediction of conversion of MCI patients at baseline to AD at 1-year follow-up, we obtained an accuracy of up to 86%. The value of the multivariate models derived from the classification of AD vs. CTL was shown to be robust and efficient in the identification of MCI converters.",

keywords = "Acknowledged-BRC , Acknowledged-BRU, Acknowledged-BRU-13/14",

author = "Carlos Aguilar and Eric Westman and J-Sebastian Muehlboeck and Patrizia Mecocci and Bruno Vellas and Magda Tsolaki and Iwona Kloszewska and Hilkka Soininen and Simon Lovestone and Christian Spenger and Andrew Simmons and Lars-Olof Wahlund",

year = "2013",

month = may,

day = "30",

doi = "10.1016/j.pscychresns.2012.11.005",

language = "English",

volume = "212",

pages = "89--98",

journal = "Psychiatry Research. Neuroimaging",

publisher = "Elsevier",

number = "2",

}

Aguilar, C, Westman, E, Muehlboeck, J-S, Mecocci, P, Vellas, B, Tsolaki, M, Kloszewska, I, Soininen, H, Lovestone, S, Spenger, C, Simmons, A & Wahlund, L-O 2013, 'Different multivariate techniques for automated classification of MRI data in Alzheimer's disease and mild cognitive impairment', Psychiatry Research. Neuroimaging, vol. 212, no. 2, N/A, pp. 89-98. https://doi.org/10.1016/j.pscychresns.2012.11.005

TY - JOUR

T1 - Different multivariate techniques for automated classification of MRI data in Alzheimer's disease and mild cognitive impairment

AU - Aguilar, Carlos

AU - Westman, Eric

AU - Muehlboeck, J-Sebastian

AU - Mecocci, Patrizia

AU - Vellas, Bruno

AU - Tsolaki, Magda

AU - Kloszewska, Iwona

AU - Soininen, Hilkka

AU - Lovestone, Simon

AU - Spenger, Christian

AU - Simmons, Andrew

AU - Wahlund, Lars-Olof

PY - 2013/5/30

Y1 - 2013/5/30

N2 - Automated structural magnetic resonance imaging (MRI) processing pipelines and different multivariate techniques are gaining popularity for Alzheimer's disease (AD) research. We used four supervised learning methods to classify AD patients and controls (CTL) and to prospectively predict the conversion of mild cognitive impairment (MCI) to AD from baseline MRI data. A total of 345 participants from the AddNeuroMed cohort were included in this study; 116 AD patients, 119 MCI patients and 110 CTL individuals. High resolution sagittal 3D MP-RAGE datasets were acquired and MRI data were processed using FreeSurfer. We explored the classification ability of orthogonal projections to latent structures (OPLS), decision trees (Trees), artificial neural networks (ANN) and support vector machines (SVM). Applying 10-fold cross-validation demonstrated that SVM and OPLS were slightly superior to Trees and ANN, although not statistically significant for distinguishing between AD and CTL. The classification experiments resulted in up to 83% sensitivity and 87% specificity for the best techniques. For the prediction of conversion of MCI patients at baseline to AD at 1-year follow-up, we obtained an accuracy of up to 86%. The value of the multivariate models derived from the classification of AD vs. CTL was shown to be robust and efficient in the identification of MCI converters.

AB - Automated structural magnetic resonance imaging (MRI) processing pipelines and different multivariate techniques are gaining popularity for Alzheimer's disease (AD) research. We used four supervised learning methods to classify AD patients and controls (CTL) and to prospectively predict the conversion of mild cognitive impairment (MCI) to AD from baseline MRI data. A total of 345 participants from the AddNeuroMed cohort were included in this study; 116 AD patients, 119 MCI patients and 110 CTL individuals. High resolution sagittal 3D MP-RAGE datasets were acquired and MRI data were processed using FreeSurfer. We explored the classification ability of orthogonal projections to latent structures (OPLS), decision trees (Trees), artificial neural networks (ANN) and support vector machines (SVM). Applying 10-fold cross-validation demonstrated that SVM and OPLS were slightly superior to Trees and ANN, although not statistically significant for distinguishing between AD and CTL. The classification experiments resulted in up to 83% sensitivity and 87% specificity for the best techniques. For the prediction of conversion of MCI patients at baseline to AD at 1-year follow-up, we obtained an accuracy of up to 86%. The value of the multivariate models derived from the classification of AD vs. CTL was shown to be robust and efficient in the identification of MCI converters.

KW - Acknowledged-BRC

KW - Acknowledged-BRU

KW - Acknowledged-BRU-13/14

U2 - 10.1016/j.pscychresns.2012.11.005

DO - 10.1016/j.pscychresns.2012.11.005

M3 - Article

C2 - 23541334

VL - 212

SP - 89

EP - 98

JO - Psychiatry Research. Neuroimaging

JF - Psychiatry Research. Neuroimaging

IS - 2

M1 - N/A

ER -

Different multivariate techniques for automated classification of MRI data in Alzheimer's disease and mild cognitive impairment

Abstract

Keywords

Access to Document

Fingerprint

Cite this