Linear predictive coding distinguishes spectral EEG features of Parkinson's disease

Md Fahim Anjum; Soura Dasgupta; Raghuraman Mudumbai; Arun Singh; James F Cavanagh; Nandakumar S Narayanan

doi:10.1016/j.parkreldis.2020.08.001

Back

Linear predictive coding distinguishes spectral EEG features of Parkinson's disease

Journal article

Peer reviewed

Linear predictive coding distinguishes spectral EEG features of Parkinson's disease

Md Fahim Anjum, Soura Dasgupta, Raghuraman Mudumbai, Arun Singh, James F Cavanagh and Nandakumar S Narayanan

Parkinsonism & related disorders, Vol.79, pp.79-85

10/2020

DOI: 10.1016/j.parkreldis.2020.08.001

PMID: 32891924

Files and links (1)

url

https://www.ncbi.nlm.nih.gov/pmc/articles/7900258View

Open Access

Abstract

We have developed and validated a novel EEG-based signal processing approach to distinguish PD and control patients: Linear-predictive-coding EEG Algorithm for PD (LEAPD). This method efficiently encodes EEG time series into features that can detect PD in a computationally fast manner amenable to real time applications. We included a total of 41 PD patients and 41 demographically-matched controls from New Mexico and Iowa. Data for all participants from New Mexico (27 PD patients and 27 controls) were used to evaluate in-sample LEAPD performance, with extensive cross-validation. Participants from Iowa (14 PD patients and 14 controls) were used for out-of-sample tests. Our method utilized data from six EEG leads which were as little as 2 min long. For the in-sample dataset, LEAPD differentiated PD patients from controls with 85.3 ± 0.1% diagnostic accuracy, 93.3 ± 0.5% area under the receiver operating characteristics curve (AUC), 87.9 ± 0.9% sensitivity, and 82.7 ± 1.1% specificity, with multiple cross-validations. After head-to-head comparison with state-of-the-art methods using our dataset, LEAPD showed a 13% increase in accuracy and a 15.5% increase in AUC. When the trained classifier was applied to a distinct out-of-sample dataset, LEAPD showed reliable performance with 85.7% diagnostic accuracy, 85.2% AUC, 85.7% sensitivity, and 85.7% specificity. No statistically significant effect of levodopa-ON and levodopa-OFF sessions were found. We describe LEAPD, an efficient algorithm that is suitable for real time application and captures spectral EEG features using few parameters and reliably differentiates PD patients from demographically-matched controls. •A novel machine-learning approach to diagnose Parkinson's disease with EEG.•Cross-validation and out-of-sample tests yield more than 85% accuracy.•Outperforms other state-of-the-art EEG methods.•Computationally efficient and amenable to real-time implementation.

Parkinson's disease

Diagnosis

Classifier

EEG

Details

Title: Subtitle: Linear predictive coding distinguishes spectral EEG features of Parkinson's disease
Creators: Md Fahim Anjum - Department of Electrical and Computer Engineering, The University of Iowa, Iowa, USA
Soura Dasgupta - Department of Electrical and Computer Engineering, The University of Iowa, Iowa, USA
Raghuraman Mudumbai - Department of Electrical and Computer Engineering, The University of Iowa, Iowa, USA
Arun Singh - Division of Basic Biomedical Sciences, Sanford School of Medicine, University of South Dakota, South Dakota, USA
James F Cavanagh - Department of Psychology, University of New Mexico, New Mexico, USA
Nandakumar S Narayanan - Department of Neurology, The University of Iowa, Iowa, USA
Resource Type: Journal article
Publication Details: Parkinsonism & related disorders, Vol.79, pp.79-85
DOI: 10.1016/j.parkreldis.2020.08.001
PMID: 32891924
NLM abbreviation: Parkinsonism Relat Disord
ISSN: 1353-8020
eISSN: 1873-5126
Publisher: Elsevier Ltd
Grant note: DOI: 10.13039/100000002, name: National Institutes of Health, award: NIH R01NS100849-01A1
Language: English
Date published: 10/2020
Academic Unit: Neurology; Electrical and Computer Engineering; Iowa Neuroscience Institute
Record Identifier: 9984070831802771

Metrics

99 Record Views

106 Times Cited - Web of Science