Machine Learning EEG to Predict Cognitive Functioning and Processing Speed Over a 2-Year Period in Multiple Sclerosis Patients and Controls
- 605 Downloads
Event-related potentials (ERPs) show promise to be objective indicators of cognitive functioning. The aim of the study was to examine if ERPs recorded during an oddball task would predict cognitive functioning and information processing speed in Multiple Sclerosis (MS) patients and controls at the individual level. Seventy-eight participants (35 MS patients, 43 healthy age-matched controls) completed visual and auditory 2- and 3-stimulus oddball tasks with 128-channel EEG, and a neuropsychological battery, at baseline (month 0) and at Months 13 and 26. ERPs from 0 to 700 ms and across the whole scalp were transformed into 1728 individual spatio-temporal datapoints per participant. A machine learning method that included penalized linear regression used the entire spatio-temporal ERP to predict composite scores of both cognitive functioning and processing speed at baseline (month 0), and months 13 and 26. The results showed ERPs during the visual oddball tasks could predict cognitive functioning and information processing speed at baseline and a year later in a sample of MS patients and healthy controls. In contrast, ERPs during auditory tasks were not predictive of cognitive performance. These objective neurophysiological indicators of cognitive functioning and processing speed, and machine learning methods that can interrogate high-dimensional data, show promise in outcome prediction.
KeywordsCognitive function Multiple sclerosis Electroencephalography Oddball paradigm Machine learning Longitudinal
This study was partly funded by an Enterprise Ireland (eBiomed: eHealthCare based on Biomedical Signal Processing and ICT for Integrated Diagnosis and Treatment of Disease), a Science Foundation Ireland grant to R.B. Reilly (09/RFP/NE2382), an IRCSET grants to H. Kiiski and S. Ó. Donnchadha (http://www.ircset.ie), a Health Service Executive funding to M.C. O’Brien and a Science Foundation Ireland grant to R. Whelan (16/ERCD/3797). The study sponsors had no involvement in the collection, analysis and interpretation of data and in the writing of the manuscript.
Compliance with Ethical Standards
Conflict of interest
The authors declare that they have no conflict of interest.
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee (the Ethics and Medical Research Committee of the St. Vincent’s Healthcare Group) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Written informed consent was obtained from all individual participants included in the study on each testing occasion.
Supplementary Video 1 ERP activity over the scalp during visual 2-stimulus oddball task (0-700ms) that predicted cognitive functioning at Month 0. Higher beta choice frequency values denote better accuracy in predicting cognitive functioning score. (AVI 2023 KB)
Supplementary Video 2 ERP activity over the scalp during visual 2-stimulus oddball task (0-700ms) that predicted cognitive functioning at Month 13. Higher beta choice frequency values denote better accuracy in predicting cognitive functioning score. (AVI 2109 KB)
Supplementary Video 3 ERP activity over the scalp during visual 3-stimulus oddball task (0-700ms) that predicted cognitive functioning at Month 13. Higher beta choice frequency values denote better accuracy in predicting cognitive functioning score. (AVI 1947 KB)
Supplementary Video 4 ERP activity over the scalp during visual 2-stimulus oddball task (0-700ms) that predicted processing speed and working memory performance at Month 0. Higher beta choice frequency values denote better accuracy in predicting processing speed and working memory score. (AVI 1935 KB)
Supplementary Video 5 ERP activity over the scalp during visual 2-stimulus oddball task (0-700ms) that predicted processing speed and working memory performance at Month 13. Higher beta choice frequency values denote better accuracy in predicting processing speed and working memory score. (AVI 1990 KB)
Supplementary Video 6 ERP activity (µV) in multiple sclerosis and healthy control participants during visual 2-stimulus oddball task at Month 0. (AVI 1574 KB)
- Beck AT, Steer RA, Brown GK (1996) Beck depression inventory-II: manual. Psychological Corporation, San AntonioGoogle Scholar
- Benton AL, Hamsher K (1989) Multilingual aphasia examination. AJA Associates, Iowa CityGoogle Scholar
- Cawley GC, Talbot NLC (2010) On over-fitting in model selection and subsequent selection bias in performance evaluation. J Mach Learn Res 11:2079–2107Google Scholar
- Covey TJ, Shucard JL, Shucard DW (2016) Evaluation of cognitive control and distraction using event-related potentials in healthy individuals and patients with multiple sclerosis. In: International conference on augmented cognition. Springer International Publishing, pp 165–176Google Scholar
- Crawford JR (1992) Current and premorbid intelligence measures in neuropsychological assessment. In: Crawford JR, McKinlay WW (eds) A handbook of neuropsychological assessment. Erlbaum, London, pp 21–49Google Scholar
- Delis DC, Kramer JH, Kaplan E, Ober BA (2000) California verbal learning test: second edition (CVLT-II). The Psychological Corporation, San AntonioGoogle 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(1):9–21. https://doi.org/10.1016/j.jneumeth.2003.10.009S0165027003003479 CrossRefPubMedGoogle Scholar
- Hoffmann S, Tittgemeyer M, von Cramon DY (2007) Cognitive impairment in multiple sclerosis. Curr Opin Neurol 20(3):275–280. https://doi.org/10.1097/WCO.0b013e32810c8e8700019052-200706000-00006 CrossRefPubMedGoogle Scholar
- Holdnack HA (2001) Wechsler test of adult reading: WTAR. The Psychological Corporation, San AntonioGoogle Scholar
- Jollans L, Whelan R (2017) Neuromarkers for mental disorders: Harnessing population neuroscience. In: Werdecker A (ed) Biomarkers for demographic research. Springer (In press)Google Scholar
- Kiiski H, Reilly RB, Lonergan R, Kelly S, O’Brien M, Kinsella K et al (2011) Change in PASAT performance correlates with change in P3 ERP amplitude over a 12-month period in multiple sclerosis patients. J Neurol Sci 305(1–2):45–52. https://doi.org/10.1016/j.jns.2011.03.018 CrossRefPubMedGoogle Scholar
- Kiiski H, Reilly RB, Lonergan R, Kelly S, O’Brien MC, Kinsella K et al (2012) Only low frequency event-related EEG activity is compromised in multiple sclerosis: insights from an independent component clustering analysis. PLoS ONE 7(9):e45536. https://doi.org/10.1371/journal.pone.0045536 CrossRefPubMedPubMedCentralGoogle Scholar
- Kimiskidis VK, Papaliagkas V, Sotirakoglou K, Kouvatsou ZK, Kapina VK, Papadaki E et al (2016) Cognitive event-related potentials in multiple sclerosis: Correlation with MRI and neuropsychological findings. Mult Scler Relat Disord 10:192–197. https://doi.org/10.1016/j.msard.2016.10.006 CrossRefPubMedGoogle Scholar
- Lopez-Gongora M, Escartin A, Martinez-Horta S, Fernandez-Bobadilla R, Querol L, Romero S et al (2015) Neurophysiological evidence of compensatory brain mechanisms in early-stage multiple sclerosis. PLoS ONE 10(8):e0136786. https://doi.org/10.1371/journal.pone.0136786 CrossRefPubMedPubMedCentralGoogle Scholar
- Polman C, Barkhof F, Sandberg-Wollheim M, Linde A, Nordle O, Nederman T (2005) Treatment with laquinimod reduces development of active MRI lesions in relapsing MS. Neurology 64(6):987–991. https://doi.org/10.1212/01.WNL.0000154520.48391.69 CrossRefPubMedGoogle Scholar
- Smith A (1982) Symbol digit modalities test: manual. Western Psychological Services, Los AngelesGoogle Scholar
- Sundgren M, Wahlin A, Maurex L, Brismar T (2015b) Event related potential and response time give evidence for a physiological reserve in cognitive functioning in relapsing-remitting multiple sclerosis. J Neurol Sci 356(1–2):107–112. https://doi.org/10.1016/j.jns.2015.06.025 CrossRefPubMedGoogle Scholar
- Titlic M, Mihalj M, Petrovic AB, Suljic E (2016) P300 as an auxiliary method in clinical practice: a review of literature. J Health Sci 6(3):143–148Google Scholar
- Trenerry MR, Crossan B, DeBoe J, Leber WR (1989) Stroop neuropsychological screening test: manual. Psychological Assessment Resources, FloridaGoogle Scholar
- Vazquez-Marrufo M, Gonzalez-Rosa JJ, Galvao-Carmona A, Hidalgo-Munoz A, Borges M, Pena JL et al (2013) Retest reliability of individual p3 topography assessed by high density electroencephalography. PLoS ONE 8(5):e62523. https://doi.org/10.1371/journal.pone.0062523 CrossRefPubMedPubMedCentralGoogle Scholar
- Vazquez-Marrufo M, Galvao-Carmona A, Gonzalez-Rosa JJ, Hidalgo-Munoz AR, Borges M, Ruiz-Pena JL et al (2014) Neural correlates of alerting and orienting impairment in multiple sclerosis patients. PLoS ONE 9(5):e97226. https://doi.org/10.1371/journal.pone.0097226 CrossRefPubMedPubMedCentralGoogle Scholar
- Whelan R, Lonergan R, Kiiski H, Nolan H, Kinsella K, Bramham J et al (2010a) A high-density ERP study reveals latency, amplitude, and topographical differences in multiple sclerosis patients versus controls. Clin Neurophysiol 121(9):1420–1426. https://doi.org/10.1016/j.clinph.2010.03.019 CrossRefPubMedGoogle Scholar
- Whelan R, Lonergan R, Kiiski H, Nolan H, Kinsella K, Hutchinson M et al (2010b) Impaired information processing speed and attention allocation in multiple sclerosis patients versus controls: a high-density EEG study. J Neurol Sci 293(1):45–50. https://doi.org/10.1016/j.jns.2010.03.010 CrossRefPubMedGoogle Scholar
- Yarkoni T, Westfall J (2016) Choosing prediction over explanation in psychology: Lessons from machine learning. Unpublished manuscript. Retrieved from http://jakewestfall.org/publications/Yarkoni_Westfall_choosing_prediction.pdf