Skip to main content
SpringerLink
Log in

Classification of Parkinson’s disease motor phenotype: a machine learning approach

  • Neurology and Preclinical Neurological Studies - Original Article
  • Published:
Journal of Neural Transmission Aims and scope Submit manuscript

Abstract

To assess the cortical activity in people with Parkinson’s disease (PwP) with different motor phenotype (tremor-dominant—TD and postural instability and gait difficulty—PIGD) and to compare with controls. Twenty-four PwP (during OFF and ON medication) and twelve age-/sex-/handedness-matched healthy controls underwent electrophysiological assessment of spectral ratio analysis through electroencephalography (EEG) at resting state and during the hand movement. We performed a machine learning method with 35 attributes extracted from EEG. To verify the efficiency of the proposed phenotype-based EEG classification the random forest and random tree were tested (performed 30 times, using a tenfolds cross validation in Weka environment). The analyses based on phenotypes indicated a slowing down of cortical activity during OFF medication state in PwP. PD with TD phenotype presented this characteristic at resting and the individuals with PIGD presented during the hand movement. During the ON state, there is no difference between phenotypes at resting nor during the hand movement. PD phenotypes may influence spectral activity measured by EEG. Random forest machine learning provides a slightly more accurate, sensible and specific approach to distinguish different PD phenotypes. The phenotype of PD might be a clinical characteristic that could influence cortical activity.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

Similar content being viewed by others

Abbreviations

ADL:

Activities of daily living

BMI:

Body mass index

CG:

Control group

CLH:

Contralateral hemisphere

DAR:

Delta-alpha ratio

EEG:

Electroencephalography

EHS:

Edinburgh Handedness scale

FFT:

Fast Fourier transform

fMRI:

Functional magnetic resonance image

GDS:

Geriatric depression scale

H&Y:

Hoehn & Yahr modified scale

LED:

Levodopa equivalent dosage

MoCA:

Montreal cognitive assessment

NA:

Not applied

PD:

Parkinson’s disease

PDSS:

Parkinson’s disease sleep scale

PIGD:

Postural instability with gait difficulty

PRI:

Power ratio index

PSD:

Power spectrum density

SD:

Standard deviation

SMA:

Supplementary motor area

TBR:

Theta-beta ratio

TD:

Tremor-dominant

TMS:

Transcranial magnetic stimulation

TUG:

Timed up and go test

UPDRS:

Unified Parkinson’s disease rating scale

References

  • Abos A, Baggio HC, Segura B et al (2019) Differentiation of multiple system atrophy from Parkinson’s disease by structural connectivity derived from probabilistic tractography. Sci Rep 9:16488

    Article  PubMed  PubMed Central  Google Scholar 

  • Ahmed Z, Mohamed K, Zeeshan S, Dong X (2020) Artificial intelligence with multi-functional machine learning platform development for better healthcare and precision medicine. Database. https://doi.org/10.1093/database/baaa010

    Article  PubMed  PubMed Central  Google Scholar 

  • Apóstolo J (2011) Adaptation into European Portuguese of the geriatric depression scale (GDS-15). Rev Referência 3

  • Awate SP, Yushkevich P, Licht D, Gee JC (2009) Gender differences in cerebral cortical folding: multivariate complexity-shape analysis with insights into handling brain-volume differences. Med Image Comput Comput Assist Interv 12:200–207

    PubMed  Google Scholar 

  • Bäumer T, Dammann E, Bock F et al (2007) Laterality of interhemispheric inhibition depends on handedness. Exp Brain Res 180:195–203

    Article  PubMed  Google Scholar 

  • Beudel M, Roosma E, Martinez Manzanera OE et al (2015) Parkinson bradykinesia correlates with EEG background frequency and perceptual forward projection. Parkinsonism Relat Disord 21:783–788

    Article  CAS  PubMed  Google Scholar 

  • Boon LI, Geraedts VJ, Hillebrand A et al (2019) A systematic review of MEG-based studies in Parkinson’s disease: the motor system and beyond. Hum Brain Mapp 40:2827–2848

    Article  PubMed  PubMed Central  Google Scholar 

  • Brazhnik E, Cruz AV, Avila I et al (2012) State-dependent spike and local field synchronization between motor cortex and substantia nigra in hemiparkinsonian rats. J Neurosci 32:7869–7880

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  • Cantillo-Negrete J, Carino-Escobar RI, Carrillo-Mora P et al (2016) Gender differences in quantitative electroencephalogram during a simple hand movement task in young adults. Rev Invest Clin 68:245–255

    PubMed  Google Scholar 

  • Cavanagh JF, Frank MJ (2014) Frontal theta as a mechanism for cognitive control. Trends Cogn Sci 18:414–421

    Article  PubMed  PubMed Central  Google Scholar 

  • Chaudhuri KR, Pal S, DiMarco A et al (2002) The Parkinson’s disease sleep scale: a new instrument for assessing sleep and nocturnal disability in Parkinson’s disease. J Neurol Neurosurg Psychiatry 73:629–635

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  • Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP (2002) SMOTE: synthetic minority over-sampling technique. J Artif Intell Res 16:321–357

    Article  Google Scholar 

  • Cheng Y, Lee P-L, Yang C-Y et al (2008) Gender differences in the mu rhythm of the human mirror-neuron system. PLoS ONE 3:e2113

    Article  PubMed  PubMed Central  Google Scholar 

  • Cozac VV, Gschwandtner U, Hatz F et al (2016) Quantitative EEG and cognitive decline in Parkinson’s disease. Parkinsons Dis 2016:9060649

    PubMed  PubMed Central  Google Scholar 

  • de Freitas Barbosa VA, Gomes JC, de Santana MA et al (2021) Heg.IA: an intelligent system to support diagnosis of Covid-19 based on blood tests. Res Biomed Eng. https://doi.org/10.1007/s42600-020-00112-5

    Article  Google Scholar 

  • de Oliveira APS, de Santana MA, Andrade MKS et al (2020) Early diagnosis of Parkinson’s disease using EEG, machine learning and partial directed coherence. Res Biomed Eng 36:311–331

    Article  Google Scholar 

  • de Sousa RL, de Medeiros JGM, de Moura ACL et al (2007) Validade e fidedignidade da Escala de Depressão Geriátrica na identificação de idosos deprimidos em um hospital geral. J Bras Psiquiatr 56:102–107

    Article  Google Scholar 

  • de Souza RG, dos Santos Lucas e Silva G, dos Santos WP et al (2021) Computer-aided diagnosis of Alzheimer’s disease by MRI analysis and evolutionary computing. Res Biomed Eng 37:455–483

    Article  Google Scholar 

  • Delorme A, Makeig S (2004) EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods 134:9–21

    Article  PubMed  Google Scholar 

  • Emek-Savaş DD, Özmüş G, Güntekin B et al (2017) Decrease of delta oscillatory responses in cognitively normal Parkinson’s disease. Clin EEG Neurosci 48:355–364

    Article  PubMed  Google Scholar 

  • Espinola CW, Gomes JC, Pereira JMS, dos Santos WP (2021) Vocal acoustic analysis and machine learning for the identification of schizophrenia. Res Biomed Eng 37:33–46

    Article  Google Scholar 

  • Fereshtehnejad S-M, Romenets SR, Anang JBM et al (2015) New clinical subtypes of Parkinson disease and their longitudinal progression: a prospective cohort comparison with other phenotypes. JAMA Neurol 72:863–873

    Article  PubMed  Google Scholar 

  • Gao C, Sun H, Wang T et al (2018) Model-based and model-free machine learning techniques for diagnostic prediction and classification of clinical outcomes in Parkinson’s disease. Sci Rep 8:1–21

    Google Scholar 

  • Geraedts VJ, Boon LI, Marinus J et al (2018) Clinical correlates of quantitative EEG in Parkinson disease: a systematic review. Neurology 91:871–883

    Article  PubMed  Google Scholar 

  • Gomes JC, Masood AI, de Silva LHS et al (2021) Covid-19 diagnosis by combining RT-PCR and pseudo-convolutional machines to characterize virus sequences. Sci Rep 11:11545

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  • Gu Q, Zhang H, Xuan M et al (2016a) Automatic classification on multi-modal MRI data for diagnosis of the postural instability and gait difficulty subtype of Parkinson’s disease. J Parkinsons Dis 6:545–556

    Article  PubMed  Google Scholar 

  • Gu Y, Chen J, Lu Y, Pan S (2016b) Integrative frequency power of EEG correlates with progression of mild cognitive impairment to dementia in Parkinson’s disease. Clin EEG Neurosci 47:113–117

    Article  PubMed  Google Scholar 

  • Hall SD, Prokic EJ, McAllister CJ et al (2014) GABA-mediated changes in inter-hemispheric beta frequency activity in early-stage Parkinson’s disease. Neuroscience 281:68–76

    Article  CAS  PubMed  Google Scholar 

  • He X, Zhang Y, Chen J et al (2017) Changes in theta activities in the left posterior temporal region, left occipital region and right frontal region related to mild cognitive impairment in Parkinson’s disease patients. Int J Neurosci 127:66–72

    Article  PubMed  Google Scholar 

  • Hoops S, Nazem S, Siderowf AD et al (2009) Validity of the MoCA and MMSE in the detection of MCI and dementia in Parkinson disease. Neurology 73:1738–1745

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  • Ishii R, Canuet L, Aoki Y et al (2017) Healthy and pathological brain aging: from the perspective of oscillations, functional connectivity, and signal complexity. Neuropsychobiology 75:151–161

    Article  PubMed  Google Scholar 

  • Jávor-Duray BN, Vinck M, van der Roest M et al (2017) Alterations in functional cortical hierarchy in Hemiparkinsonian rats. J Neurosci 37:7669–7681

    Article  PubMed  PubMed Central  Google Scholar 

  • Khedr EM, Al-Fawal B, Abdel Wraith A et al (2019) The effect of 20 Hz versus 1 Hz repetitive transcranial magnetic stimulation on motor dysfunction in Parkinson’s disease: which is more beneficial? J Parkinsons Dis 9:379–387

    Article  PubMed  Google Scholar 

  • Khedr EM, Lefaucheur J-P, Hasan AM, Osama K (2021) Are there differences in cortical excitability between akinetic-rigid and tremor-dominant subtypes of Parkinson’s disease? Neurophysiol Clin 51:443–453

    Article  PubMed  Google Scholar 

  • Klem GH (1999) The ten-twenty electrode system of the international federation. The international federation of clinical nenrophysiology. Electroencephalogr Clin Neurophysiol Suppl 52:3–6

    CAS  PubMed  Google Scholar 

  • Kolmancic K, Perellón-Alfonso R, Pirtosek Z et al (2019) Sex differences in Parkinson’s disease: a transcranial magnetic stimulation study. Mov Disord 34:1873–1881

    Article  PubMed  Google Scholar 

  • Lang AE, Eberly S, Goetz CG et al (2013) Movement disorder society unified Parkinson disease rating scale experiences in daily living: longitudinal changes and correlation with other assessments. Mov Disord 28:1980–1986

    Article  PubMed  Google Scholar 

  • Lichter DG, Benedict RHB, Hershey LA (2021) Freezing of gait in Parkinson’s disease: risk factors, their interactions, and associated nonmotor symptoms. Parkinsons Dis 2021:8857204

    PubMed  PubMed Central  Google Scholar 

  • Luccas FJ, Anghinah R, Braga NI et al (1999) Guidelines for recording/analyzing quantitative EEG and evoked potentials. Part II: clinical aspects. Arq Neuropsiquiatr 57:132–146

    Article  CAS  PubMed  Google Scholar 

  • Luders E, Narr KL, Thompson PM et al (2004) Gender differences in cortical complexity. Nat Neurosci 7:799–800

    Article  CAS  PubMed  Google Scholar 

  • Mestre TA, Fereshtehnejad S-M, Berg D et al (2021) Parkinson’s disease subtypes: critical appraisal and recommendations. J Parkinsons Dis 11:395–404

    Article  PubMed  PubMed Central  Google Scholar 

  • Morita A, Kamei S, Serizawa K, Mizutani T (2009) The relationship between slowing EEGs and the progression of Parkinson’s disease. J Clin Neurophysiol 26:426–429

    Article  PubMed  Google Scholar 

  • Neuper C, Pfurtscheller G (2001) Evidence for distinct beta resonance frequencies in human EEG related to specific sensorimotor cortical areas. Clin Neurophysiol 112:2084–2097

    Article  CAS  PubMed  Google Scholar 

  • Niedermeyer E, da Silva FHL (2005) Electroencephalography: basic principles, clinical applications, and related fields. Lippincott Williams & Wilkins

    Google Scholar 

  • Niethammer M, Feigin A, Eidelberg D (2012) Functional neuroimaging in Parkinson’s disease. Cold Spring Harb Perspect Med 2:a009274

    Article  PubMed  PubMed Central  Google Scholar 

  • Oldfield RC (1971) The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9:97–113

    Article  CAS  PubMed  Google Scholar 

  • Pan P, Zhang Y, Liu Y et al (2017) Abnormalities of regional brain function in Parkinson’s disease: a meta-analysis of resting state functional magnetic resonance imaging studies. Sci Rep. https://doi.org/10.1038/srep40469

    Article  PubMed  PubMed Central  Google Scholar 

  • Pandis N (2014) Cross-sectional studies. Am J Orthod Dentofacial Orthop 146:127–129

    Article  PubMed  Google Scholar 

  • Pang H, Yu Z, Yu H et al (2021) Use of machine learning method on automatic classification of motor subtype of Parkinson’s disease based on multilevel indices of rs-fMRI. Parkinsonism Relat Disord 90:65–72

    Article  PubMed  Google Scholar 

  • Pfurtscheller G, Lopes da Silva FH (1999) Event-related EEG/MEG synchronization and desynchronization: basic principles. Clin Neurophysiol 110:1842–1857

    Article  CAS  PubMed  Google Scholar 

  • Poewe W, Seppi K, Tanner CM et al (2017) Parkinson disease. Nat Rev Dis Primers 3:17013

    Article  PubMed  Google Scholar 

  • Pollok B, Krause V, Martsch W et al (2012) Motor-cortical oscillations in early stages of Parkinson’s disease. J Physiol 590:3203–3212

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  • Possti D, Fahoum F, Sosnik R et al (2021) Changes in the EEG spectral power during dual-task walking with aging and Parkinson’s disease: initial findings using Event-Related Spectral Perturbation analysis. J Neurol 268:161–168

    Article  PubMed  Google Scholar 

  • Schrag A, Barone P, Brown RG et al (2007) Depression rating scales in Parkinson’s disease: critique and recommendations. Mov Disord 22:1077–1092

    Article  PubMed  PubMed Central  Google Scholar 

  • Serizawa K, Kamei S, Morita A et al (2008) Comparison of quantitative EEGs between Parkinson disease and age-adjusted normal controls. J Clin Neurophysiol 25:361–366

    Article  PubMed  Google Scholar 

  • Shirahige L, Berenguer-Rocha M, Mendonça S et al (2020) Quantitative electroencephalography characteristics for Parkinson’s disease: a systematic review. J Parkinsons Dis 10:455–470

    Article  PubMed  PubMed Central  Google Scholar 

  • Shukla S, Thirugnanasambandam N (2021) Deriving mechanistic insights from machine learning and its possible implications in non-invasive brain stimulation research. Brain Stimul 14:1035–1037

    Article  PubMed  Google Scholar 

  • Simuni T, Caspell-Garcia C, Coffey C et al (2016) How stable are Parkinson’s disease subtypes in de novo patients: analysis of the PPMI cohort? Parkinsonism Relat Disord 28:62–67

    Article  PubMed  Google Scholar 

  • Singh A, Richardson SP, Narayanan N, Cavanagh JF (2018) Mid-frontal theta activity is diminished during cognitive control in Parkinson’s disease. Neuropsychologia 117:113–122

    Article  PubMed  PubMed Central  Google Scholar 

  • Singh A, Cole RC, Espinoza AI et al (2020) Frontal theta and beta oscillations during lower-limb movement in Parkinson’s disease. Clin Neurophysiol 131:694–702

    Article  PubMed  PubMed Central  Google Scholar 

  • Sowell ER, Peterson BS, Kan E et al (2007) Sex differences in cortical thickness mapped in 176 healthy individuals between 7 and 87 years of age. Cereb Cortex 17:1550–1560

    Article  PubMed  Google Scholar 

  • Soysal A, Sobe I, Atay T et al (2008) Effect of therapy on motor cortical excitability in Parkinson’s disease. Can J Neurol Sci 35:166–172

    Article  PubMed  Google Scholar 

  • Spagnolo F, Coppi E, Chieffo R et al (2013) Interhemispheric balance in Parkinson’s disease: a transcranial magnetic stimulation study. Brain Stimul 6:892–897

    Article  PubMed  Google Scholar 

  • Stebbins GT, Goetz CG, Burn DJ et al (2013) How to identify tremor dominant and postural instability/gait difficulty groups with the movement disorder society unified Parkinson’s disease rating scale: comparison with the unified Parkinson’s disease rating scale. Mov Disord 28:668–670

    Article  PubMed  Google Scholar 

  • Stoffers D, Bosboom JLW, Deijen JB et al (2008) Increased cortico-cortical functional connectivity in early-stage Parkinson’s disease: an MEG study. Neuroimage 41:212–222

    Article  CAS  PubMed  Google Scholar 

  • Sun D et al (2021) Differentiating Parkinson’s disease motor subtypes: a radiomics analysis based on deep gray nuclear lesion and white matter. Neurosci Lett 760:136083

    Article  CAS  PubMed  Google Scholar 

  • Tadel F, Baillet S, Mosher JC et al (2011) Brainstorm: a user-friendly application for MEG/EEG analysis. Comput Intell Neurosci 2011:879716

    Article  PubMed  PubMed Central  Google Scholar 

  • Tropini G, Chiang J, Wang ZJ et al (2011) Altered directional connectivity in Parkinson’s disease during performance of a visually guided task. Neuroimage 56:2144–2156

    Article  PubMed  Google Scholar 

  • Udupa K, Chen R (2013) Motor cortical plasticity in Parkinson’s disease. Front Neurol 4:128

    Article  PubMed  PubMed Central  Google Scholar 

  • Williams JR, Hirsch ES, Anderson K et al (2012) A comparison of nine scales to detect depression in Parkinson disease: which scale to use? Neurology 78:998–1006

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  • Winkler I, Haufe S, Tangermann M (2011) Automatic classification of artifactual ICA-components for artifact removal in EEG signals. Behav Brain Funct 7:30

    Article  PubMed  PubMed Central  Google Scholar 

  • Zarkowski P, Shin CJ, Dang T et al (2006) EEG and the variance of motor evoked potential amplitude. Clin EEG Neurosci 37:247–251

    Article  CAS  PubMed  Google Scholar 

  • Goetz CG (2012) Unified Parkinson’s Disease Rating Scale (UPDRS) and Movement Disorder Society Revision of the UPDRS (MDS-UPDRS). In: Rating Scales in Parkinson’s Disease, pp 62–83

  • Meneses MS (2003) Doença de Parkinson. Guanabara Koogan

  • Wang F, Pan Y, Zhang M, Hu K (2021) Predicting the onset of freezing of gait in de novo Parkinson’s disease. bioRxiv

  • Winkler I, Haufe S, Mueller K-R (2015) Removal of muscular artefacts for the analysis of brain oscillations: comparison between ICA and SSD. In: ICML workshop on statistics, machine learning and neuroscience (Stamlins 2015)

  • Witten IH, Frank E, Hall MA et al (2005) Practical machine learning tools and techniques. In: Data mining. p 4

Download references

Acknowledgements

We would like to thank Adriana Costa-Ribeiro, Clynton Correa and Érika Rodrigues for their thoughtful suggestions for the article.

Funding

Shirahige L was supported by the Fundação de Amparo à Ciência e Tecnologia de Pernambuco (FACEPE), Brazil (IBPG-1548-4.01/16). Monte-Silva K receives a grant (308291/2015-8) from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Brazil. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author information

Author notes

  1. Kátia Monte-Silva

    Present address: Applied Neuroscience Laboratory, Department of Physical Therapy, Universidade Federal de Pernambuco, w/n Jornalista Aníbal Fernandes Avenue, Recife, PE, 50740-560, Brazil

Authors and Affiliations

  1. Applied Neuroscience Laboratory, Department of Physical Therapy, Universidade Federal de Pernambuco, w/n Jornalista Aníbal Fernandes Avenue, Recife, PE, 50740-560, Brazil

    Lívia Shirahige, Brenda Leimig, Adriana Baltar & Amanda Bezerra

  2. Post-graduation Program of Neuropsychiatry and Behavioral Sciences, Universidade Federal de Pernambuco, Recife, PE, Brazil

    Lívia Shirahige & Adriana Baltar

  3. Neurodynamics Laboratory, Department of Physiology, Universidade Federal de Pernambuco, Recife, PE, Brazil

    Caio Vinícius Ferreira de Brito, Yasmin Samara Oliveira do Nascimento & Marcelo Cairrão

  4. Department of Biomedical Engineering, Universidade Federal de Pernambuco, Recife, PE, Brazil

    Juliana Carneiro Gomes & Wellignton Pinheiro dos Santos

  5. Physical Education and Sports Science Academic Group, National Institute of Education, Nanyang Technological University, Singapore, Singapore

    Wei-Peng Teo

  6. Center of Mathematics, Computation and Cognition, Universidade Federal do ABC, São Paulo, São Paulo, Brazil

    André Fonseca

Authors
  1. Lívia Shirahige
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Brenda Leimig
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Adriana Baltar
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Amanda Bezerra
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Caio Vinícius Ferreira de Brito
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Yasmin Samara Oliveira do Nascimento
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Juliana Carneiro Gomes
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Wei-Peng Teo
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Wellignton Pinheiro dos Santos
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Marcelo Cairrão
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. André Fonseca
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Kátia Monte-Silva
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Kátia Monte-Silva.

Ethics declarations

Conflict of interest

The authors report no conflicts of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 651 KB)

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Shirahige, L., Leimig, B., Baltar, A. et al. Classification of Parkinson’s disease motor phenotype: a machine learning approach. J Neural Transm 129, 1447–1461 (2022). https://doi.org/10.1007/s00702-022-02552-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00702-022-02552-y

Keywords

Search

Navigation