Abstract
Purpose
Accurate detection of depression at an individual level using structural magnetic resonance imaging (sMRI) remains a challenge. Brain volumetric changes at a structural level appear to have importance in depression biomarkers studies. An automated algorithm is developed to select brain sMRI volumetric features for the detection of depression.
Methods
A feature selection (FS) algorithm called degree of contribution (DoC) is developed for selection of sMRI volumetric features. This algorithm uses an ensemble approach to determine the degree of contribution in detection of major depressive disorder. The DoC is the score of feature importance used for feature ranking. The algorithm involves four stages: feature ranking, subset generation, subset evaluation, and DoC analysis. The performance of DoC is evaluated on the Duke University Multi-site Imaging Research in the Analysis of Depression sMRI dataset. The dataset consists of 115 brain sMRI scans of 88 healthy controls and 27 depressed subjects. Forty-four sMRI volumetric features are used in the evaluation.
Results
The DoC score of forty-four features was determined as the accuracy threshold (Acc_Thresh) was varied. The DoC performance was compared with that of four existing FS algorithms. At all defined Acc_Threshs, DoC outperformed the four examined FS algorithms for the average classification score and the maximum classification score.
Conclusion
DoC has a good ability to generate reduced-size subsets of important features that could yield high classification accuracy. Based on the DoC score, the most discriminant volumetric features are those from the left-brain region.
Similar content being viewed by others
References
Kipli K, Kouzani A, Williams L (2013) Towards automated detection of depression from brain structural magnetic resonance images. Neuroradiology, pp. 1–18. doi:10.1007/s00234-013-1139-8
Kipli K, Kouzani AZ, Joordens M (2012) Computer-aided detection of depression from magnetic resonance images. In: International conference on complex medical engineering (CME), 2012 ICME, 1–4 July 2012. pp 500–505. doi:10.1109/iccme.2012.6275745
Guyon I, Elisseeff A (2003) An introduction to variable and feature selection. J Mach Learn Res 3:1157–1182
Daelemans W, Hoste V, Meulder F, Naudts B (2003) Combined optimization of feature selection and algorithm parameters in machine learning of language. In: Lavrač N, Gamberger D, Blockeel H, Todorovski L (eds) Machine learning: ECML 2003, vol 2837. Lecture notes in computer science. Springer, Berlin, pp 84–95. doi:10.1007/978-3-540-39857-8_10
Ruiz R, Riquelme J, Aguilar-Ruiz J (2005) Heuristic search over a ranking for feature selection. In: Cabestany J, Prieto A, Sandoval F (eds) Computational intelligence and bioinspired systems, vol 3512. Lecture Notes in Computer Science. Springer, Berlin, pp 742–749. doi:10.1007/11494669_91
Costafreda SG, Chu C, Ashburner J, Fu CHY (2009) Prognostic and diagnostic potential of the structural neuroanatomy of depression. PLoS ONE 4(7):e6353. doi:10.1371/journal.pone.0006353
Mwangi B, Ebmeier KP, Matthews K, Douglas Steele J (2012) Multi-centre diagnostic classification of individual structural neuroimaging scans from patients with major depressive disorder. Brain 135(5):1508–1521. doi:10.1093/brain/aws084
Fu CHY, Mourao-Miranda J, Costafreda SG, Khanna A, Marquand AF, Williams SCR, Brammer MJ (2008) Pattern classification of sad facial processing: toward the development of neurobiological markers in depression. Biol Psychiatry 63(7):656–662. doi:10.1016/j.biopsych.2007.08.020
Marquand AF, Mourão-Miranda J, Brammer MJ, Cleare AJ, Fu CHY (2008) Neuroanatomy of verbal working memory as a diagnostic biomarker for depression. NeuroReport 19(15):1507–1511
Chyzhyk D, Graña M, Savio A, Maiora J (2012) Hybrid dendritic computing with kernel-LICA applied to Alzheimer’s disease detection in MRI. Neurocomputing 75(1):72–77. doi:10.1016/j.neucom.2011.02.024
De Martino F, Valente G, Staeren N, Ashburner J, Goebel R, Formisano E (2008) Combining multivariate voxel selection and support vector machines for mapping and classification of fMRI spatial patterns. NeuroImage 43(1):44–58. doi:10.1016/j.neuroimage.2008.06.037
Craddock RC, Holtzheimer PE, Hu XP, Mayberg HS (2009) Disease state prediction from resting state functional connectivity. Magn Reson Med 62(6):1619–1628. doi:10.1002/mrm.22159
Zeng L-L, Shen H, Liu L, Wang L, Li B, Fang P, Zhou Z, Li Y, Hu D (2012) Identifying major depression using whole-brain functional connectivity: a multivariate pattern analysis. Brain 135(5):1498–1507. doi:10.1093/brain/aws059
Nouretdinov I, Costafreda SG, Gammerman A, Chervonenkis A, Vovk V, Vapnik V, Fu CHY (2011) Machine learning classification with confidence: application of transductive conformal predictors to MRI-based diagnostic and prognostic markers in depression. NeuroImage 56(2):809–813. doi:10.1016/j.neuroimage.2010.05.023
Gong Q, Wu Q, Scarpazza C, Lui S, Jia Z, Marquand A, Huang X, McGuire P, Mechelli A (2011) Prognostic prediction of therapeutic response in depression using high-field MR imaging. NeuroImage 55(4):1497–1503. doi:10.1016/j.neuroimage.2010.11.079
Mwangi B, Matthews K, Steele JD (2012) Prediction of illness severity in patients with major depression using structural MR brain scans. J Magn Reson Imaging 35(1):64–71. doi:10.1002/jmri.22806
Bao F, Ghosh S, Giard J, Parsey R, Klein A (2011) Brain shape analysis for predicting treatment remission in major depressive disorder. Paper presented at the 41st annual meeting for the society for neuroscience
Kipli K, Kouzani AZ, Hamid IRA (2013) Investigating machine learning techniques for detection of depression using structural MRI volumetric features. Int J Biosci Biochem Bioinform 3(5):444–448
Kipli K, Kouzani AZ, Joordens M (2013) Evaluation of feature selection algorithms for detection of depression from brain sMRI scans. In: Complex medical engineering (CME), 2013 ICME international conference on, 25–28 May 2013. pp 64–69. doi:10.1109/ICCME.2013.6548213
Kipli K, Kouzani AZ (2013) An algorithm for determination of rank and degree of contribution of sMRI volumetric features in depression detection. Paper presented at the 35th annual international conference of the IEEE engineering in medicine and biology society (EMBC’13), Osaka, Japan
Saeys Y, Abeel T, de Peer Y (2008) Towards robust feature selection techniques. In: Proceedings of Benelearn, pp 45–46
Saeys Y, Inza I, Larrañaga P (2007) A review of feature selection techniques in bioinformatics. Bioinformatics 23(19):2507–2517
Kabir MM, Shahjahan M, Murase K (2012) A new hybrid ant colony optimization algorithm for feature selection. Expert Syst Appl 39(3):3747–3763. doi:10.1016/j.eswa.2011.09.073
Whitney AW (1971) A direct method of nonparametric measurement selection. IEEE Trans Comput C–20(9):1100–1103. doi:10.1109/T-C.1971.223410
Marill T, Green DM (1963) On the effectiveness of receptors in recognition systems. IEEE Trans Inf Theory 9(1):11–17
Stearns SD (1976) On selecting features for pattern classifiers. In: Proceedings of the 3rd international conference on pattern recognition (ICPR 1976), pp 71–75
Kittler J (1978) An introduction to feature extraction, pp 41–60
Pudil P, Novovičová J, Kittler J (1994) Floating search methods in feature selection. Pattern Recognit Lett 15(11):1119–1125. doi:10.1016/0167-8655(94)90127-9
Peng H, Fulmi L, Ding C (2005) Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancy. IEEE Trans Pattern Anal Mach Intell 27(8):1226–1238. doi:10.1109/TPAMI.2005.159
Guyon I, Weston J, Barnhill S, Vapnik V (2002) Gene selection for cancer classification using support vector machines. Mach Learn 46(1–3):389–422. doi:10.1023/a:1012487302797
Tibshirani R (2011) Regression shrinkage and selection via the lasso: a retrospective. J R Stat Soc Ser B (Statistical Methodology) 73(3):273–282
Zou H, Hastie T (2005) Regularization and variable selection via the elastic net. J R Stat Soc Ser B (Statistical Methodology) 67(2):301–320
McIntosh AR, Lobaugh NJ (2004) Partial least squares analysis of neuroimaging data: applications and advances. NeuroImage 23:S250–S263
Mwangi B, Tian T, Soares J (2013) A review of feature reduction techniques in neuroimaging. Neuroinform, pp 1–16. doi:10.1007/s12021-013-9204-3
Belanche L, González F (2011) Review and evaluation of feature selection algorithms in synthetic problems. arXiv preprint arXiv:11012320
Huan L, Lei Y (2005) Toward integrating feature selection algorithms for classification and clustering. IEEE Trans Knowl Data Eng 17(4):491–502. doi:10.1109/tkde.2005.66
Santana LA, Canuto AP (2012) Bi-objective genetic algorithm for feature selection in ensemble systems. In: Villa AP, Duch W, Érdi P, Masulli F, Palm G (eds) Artificial neural networks and machine learning—ICANN 2012, vol 7552. Lecture notes in computer Science. Springer, Berlin, pp 701–709. doi:10.1007/978-3-642-33269-2_88
Abeel T, Helleputte T, Van de Peer Y, Dupont P, Saeys Y (2010) Robust biomarker identification for cancer diagnosis with ensemble feature selection methods. Bioinformatics 26(3):392–398. doi:10.1093/bioinformatics/btp630
Saeys Y, Abeel T, Peer Y (2008) Robust feature selection using ensemble feature selection techniques. In: Daelemans W, Goethals B, Morik K (eds) Machine learning and knowledge discovery in databases, vol 5212. Lecture notes in computer science. Springer, Berlin, pp 313–325. doi:10.1007/978-3-540-87481-2_21
Ben Brahim A, Limam M (2013) Robust ensemble feature selection for high dimensional data sets. In: High performance computing and simulation (HPCS), 2013 international conference on. IEEE, pp 151–157
Yang P, Liu W, Zhou BB, Chawla S, Zomaya AY (2013) Ensemble-based wrapper methods for feature selection and class imbalance learning. In: Advances in knowledge discovery and data mining. Springer, Berlin, pp 544–555
Das R, Turkoglu I, Sengur A (2009) Effective diagnosis of heart disease through neural networks ensembles. Expert Syst Appl 36(4):7675–7680. doi:10.1016/j.eswa.2008.09.013
Netzer M, Millonig G, Osl M, Pfeifer B, Praun S, Villinger J, Vogel W, Baumgartner C (2009) A new ensemble-based algorithm for identifying breath gas marker candidates in liver disease using ion molecule reaction mass spectrometry. Bioinformatics 25(7):941–947. doi:10.1093/bioinformatics/btp093
Tsymbal A, Puuronen S, Patterson DW (2003) Ensemble feature selection with the simple Bayesian classification. Inf Fusion 4(2):87–100. doi:10.1016/S1566-2535(03)00004-6
Xu J, Sun L, Gao Y, Xu T (2013) An ensemble feature selection technique for cancer recognition. Biomed Mater Eng 23:S1053–S1060
Tsai C-F, Hsiao Y-C (2010) Combining multiple feature selection methods for stock prediction: Union, intersection, and multi-intersection approaches. Decision Support Syst 50(1):258–269. doi:10.1016/j.dss.2010.08.028
Brendel M, Zaccarelli R, Devillers L (2010) A quick sequential forward floating feature selection algorithm for emotion detection from speech. In: INTERSPEECH, pp 1157–1160
NIRL Imaging Database http://nirlarc.duhs.duke.edu/nirle/
Zhao Z, Taylor WD, Styner M, Steffens DC, Krishnan KRR, MacFall JR (2008) Hippocampus shape analysis and late-life depression. PLoS ONE 3(3):e1837. doi:10.1371/journal.pone.0001837
Greenberg DL, Payne ME, MacFall JR, Steffens DC, Krishnan RR (2008) Hippocampal volumes and depression subtypes. Psychiatry Res Neuroimaging 163(2):126–132. doi:10.1016/j.pscychresns.2007.12.009
Taylor WD, MacFall JR, Steffens DC, Payne ME, Provenzale JM, Krishnan KRR (2003) Localization of age-associated white matter hyperintensities in late-life depression. Prog Neuro-Psychopharmacol Biol Psychiatry 27(3):539–544. doi:10.1016/s0278-5846(02)00358-5
Steffens DC, Tupler LA, Krishnan KRR (1998) Magnetic resonance imaging signal hypointensity and iron content of putamen nuclei in elderly depressed patients. Psychiatry Res Neuroimaging 83(2):95–103. doi:10.1016/s0925-4927(98)00032-8
Lai T-J, Payne ME, Byrum CE, Steffens DC, Krishnan KRR (2000) Reduction of orbital frontal cortex volume in geriatric depression. Biol Psychiatry 48(10):971–975. doi:10.1016/s0006-3223(00)01042-8
Steffens DC, Byrum CE, McQuoid DR, Greenberg DL, Payne ME, Blitchington TF, MacFall JR, Krishnan KRR (2000) Hippocampal volume in geriatric depression. Biol Psychiatry 48(4):301–309. doi:10.1016/s0006-3223(00)00829-5
Tupler LA, Krishnan KRR, McDonald WM, Dombeck CB, D’Souza S, Steffens DC (2002) Anatomic location and laterality of MRI signal hyperintensities in late-life depression. J Psychosom Res 53(2):665–676. doi:10.1016/s0022-3999(02)00425-7
Lee S-H, Payne ME, Steffens DC, McQuoid DR, Lai T-J, Provenzale JM, Krishnan KRR (2003) Subcortical lesion severity and orbitofrontal cortex volume in geriatric depression. Biol Psychiatry 54(5):529–533. doi:10.1016/s0006-3223(03)00063-5
Steffens DC, Trost WT, Payne ME, Hybels CF, MacFall JR (2003) Apolipoprotein E genotype and subcortical vascular lesions in older depressed patients and control subjects. Biol Psychiatry 54(7):674–681. doi:10.1016/s0006-3223(02)01782-1
Steffens DC, McQuoid DR, Welsh-Bohmer KA, Krishnan KRR (2003) Left orbital frontal cortex volume and performance on the benton visual retention test in older depressives and controls. Neuropsychopharmacology 28(12):2179–2183
Taylor WD, Steffens DC, McQuoid DR, Payne ME, Lee S-H, Lai T-J, Krishnan KRR (2003) Smaller orbital frontal cortex volumes associated with functional disability in depressed elders. Biol Psychiatry 53(2):144–149. doi:10.1016/s0006-3223(02)01490-7
Taylor WD, MacFall JR, Payne ME, McQuoid DR, Steffens DC, Provenzale JM, Krishnan RR (2005) Greater MRI lesion volumes in elderly depressed subjects than in control subjects. Psychiatry Res Neuroimaging 139(1):1–7. doi:10.1016/j.pscychresns.2004.08.004
Taylor WD, Steffens DC, Payne ME, MacFall JR, Marchuk DA, Svenson IK, Krishnan KRR (2005) Influence of serotonin transporter promoter region polymorphisms on hippocampal volumes in late-life depression. Arch Gen Psychiatry 62(5):537–544. doi:10.1001/archpsyc.62.5.537
Bae JN, MacFall JR, Krishnan KRR, Payne ME, Steffens DC, Taylor WD (2006) Dorsolateral prefrontal cortex and anterior cingulate cortex white matter alterations in late-life depression. Biol Psychiatry 60(12):1356–1363. doi:10.1016/j.biopsych.2006.03.052
Chen PS, McQuoid DR, Payne ME, Steffens DC (2006) White matter and subcortical gray matter lesion volume changes and late-life depression outcome: a 4-year magnetic resonance imaging study. Int Psychogeriatr 18(3):445–456. doi:10.1017/s1041610205002796
MacFall JR, Taylor WD, Rex DE, Pieper S, Payne ME, McQuoid DR, Steffens DC, Kikinis R, Toga AW, Krishnan KRR (2005) Lobar distribution of lesion volumes in late-life depression: the biomedical informatics research network (BIRN). Neuropsychopharmacology 31(7):1500–1507
Potter GG, Blackwell AD, McQuoid DR, Payne ME, Steffens DC, Sahakian BJ, Welsh-Bohmer KA, Krishnan KRR (2007) Prefrontal white matter lesions and prefrontal task impersistence in depressed and nondepressed elders. Neuropsychopharmacology 32(10):2135–2142
Taylor WD, MacFall JR, Payne ME, McQuoid DR, Steffans DC, Provenzale JM, Krishnan KRR (2007) Orbitofrontal cortex volume in late life depression: influence of hyperintense lesions and genetic polymorphisms. Psychol Med 37(12):1763–1773. doi:10.1017/S0033291707000128
Steffens DC, Taylor WD, McQuoid DR, Krishnan KRR (2008) Short/long heterozygotes at 5HTTLPR and white matter lesions in geriatric depression. Int J Geriatr Psychiatry 23(3):244–248. doi:10.1002/gps.1869
Taylor WD, Zchner S, McQuoid DR, Payne ME, MacFall JR, Steffens DC, Speer MC, Krishnan KRR (2008) The brain-derived neurotrophic factor val66met polymorphism and cerebral white matter hyperintensities in late-life depression. Am J Geriatr Psychiatry 16(4):263–271. doi:10.1097/JGP.0b013e3181591c30
Taylor WD, Zhao Z, Ashley-Koch A, Payne ME, Steffens DC, Krishnan RR, Hauser E, Macfall JR (2011) Fiber tract-specific white matter lesion severity findings in late-life depression and by AGTR1 A1166C genotype. Hum Brain Mapp. doi:10.1002/hbm.21445
Pan C-C, McQuoid DR, Taylor WD, Payne ME, Ashley-Koch A, Steffens DC (2009) Association analysis of the COMT/MTHFR genes and geriatric depression: an MRI study of the putamen. Int J Geriatr Psychiatry 24(8):847–855. doi:10.1002/gps.2206
Qiu A, Taylor WD, Zhao Z, MacFall JR, Miller MI, Key CR, Payne ME, Steffens DC, Krishnan KRR (2009) APOE related hippocampal shape alteration in geriatric depression. NeuroImage 44(3):620–626. doi:10.1016/j.neuroimage.2008.10.010
Taylor WD, Steffens DC, Ashley-Koch A, Payne ME, MacFall JR, Potocky CF, Krishnan KRR (2010) Angiotensin receptor gene polymorphisms and 2-year change in hyperintense lesion volume in men. Mol Psychiatry 15(8):816–822. http://www.nature.com/mp/journal/v15/n8/suppinfo/mp200926s1.html
Steffens DC, Krishnan KRR (1998) Structural neuroimaging and mood disorders: recent findings, implications for classification, and future directions. Biol Psychiatry 43(10):705–712. doi:10.1016/s0006-3223(98)00084-5
Payne ME, Fetzer DL, MacFall JR, Provenzale JM, Byrum CE, Krishnan KRR (2002) Development of a semi-automated method for quantification of MRI gray and white matter lesions in geriatric subjects. Psychiatry Res Neuroimaging 115(1–2):63–77. doi:10.1016/s0925-4927(02)00009-4
Frank E, Hall M, Trigg L, Holmes G, Witten IH (2004) Data mining in bioinformatics using Weka. Bioinformatics 20(15):2479–2481. doi:10.1093/bioinformatics/bth261
Somol P, Novovičová J (2008) Evaluating the stability of feature selectors that optimize feature subset cardinality. In: Vitoria Lobo N, Kasparis T, Roli F et al (eds) Structural, syntactic, and statistical pattern Recognition, vol 5342. Lecture notes in computer science. Springer, Berlin, pp 956–966. doi:10.1007/978-3-540-89689-0_99
Han Y, Yu L (2012) A variance reduction framework for stable feature selection. Stat Anal Data Min 5(5):428–445. doi:10.1002/sam.11152
Yu L, Liu H (2004) Efficient feature selection via analysis of relevance and redundancy. J Mach Learn Res 5:1205–1224
Dernoncourt D, Hanczar B, Zucker J-D (2014) Analysis of feature selection stability on high dimension and small sample data. Comput Stat Data Anal 71:681–693. doi:10.1016/j.csda.2013.07.012
Kalousis A, Prados J, Hilario M (2007) Stability of feature selection algorithms: a study on high-dimensional spaces. Knowl Inf Syst 12(1):95–116
Bermejo P, Gámez JA, Puerta JM (2014) Speeding up incremental wrapper feature subset selection with Naive Bayes classifier. Knowl-Based Syst 55:140–147. doi:10.1016/j.knosys.2013.10.016
Bermejo P, Gámez JA, Puerta JM (2011) A GRASP algorithm for fast hybrid (filter-wrapper) feature subset selection in high-dimensional datasets. Pattern Recognit Lett 32(5):701–711. doi:10.1016/j.patrec.2010.12.016
Bermejo P, de la Ossa L, Gámez JA, Puerta JM (2012) Fast wrapper feature subset selection in high-dimensional datasets by means of filter re-ranking. Knowl-Based Syst 25(1):35–44. doi:10.1016/j.knosys.2011.01.015
Acknowledgments
We thank the Duke University Neuropsychiatric Imaging Research Laboratory for making the Multisite Imaging Research in the Analysis of Depression (MIRIAD) MRI data available. The first author would like to acknowledge the funding from Ministry of Education Malaysia (MoE) and Universiti Malaysia Sarawak (UNIMAS).
Conflict of interest
There is no conflict of interest.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
See Table 5.
Rights and permissions
About this article
Cite this article
Kipli, K., Kouzani, A.Z. Degree of contribution (DoC) feature selection algorithm for structural brain MRI volumetric features in depression detection. Int J CARS 10, 1003–1016 (2015). https://doi.org/10.1007/s11548-014-1130-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11548-014-1130-9