Objective: To classify tremor syndromes from freely hand drawn pen-on-paper Archimedes spirals using classical machine learning techniques.

Background: Spiral drawings (Archimedean) are commonly used in clinics to identify, quantify and differentiate tremor syndromes. Considering the complex variations in tremor and spirals in different conditions, machine learning techniques maybe leveraged to accurately detect tremor from spirals. We applied various classical machine learning approaches to analyse and classify the pen-on-paper Archimedes spirals from patients with tremor.

Method: We scanned and processed 130 spirals [50 Dystonic Tremor (DT), 50 Essential Tremor (ET), 30 Parkinson’s Disease (PD)] from patients and 50 from healthy volunteers using an automated algorithm to compute mean deviation (MD) and tremor variability (TV). We applied classical machine learning classifiers such as Support Vector Classifier (SVC), Random Forest (RF) and K-Nearest Neighbour (KNN) to the derived spiral parameters to predict classification into normal and patient spirals.

Results: KNN classifier was better than RF in distinguishing patients from controls using spiral data (KNN: F1 score=0.88 and accuracy=0.82, RF: F1 score=0.76 and accuracy=0.71).

Conclusion: Classical machine learning techniques applied to spiral datasets can distinguish between tremor subjects and healthy volunteers. With a bigger sample size, these techniques may similarly be used to differentiate among distinct tremor syndromes.

References: [1] Darnall ND, Donovan CK, Aktar S, Tseng HY, Barthelmess P, Cohen PR, et al. Application of machine learning and numerical analysis to classify tremor in patients affected with essential tremor or Parkinson’s disease. Gerontechnology. 2012 Jun 7;10(4):208–19.
[2] Butt AH, Rovini E, Dolciotti C, De Petris G, Bongioanni P, Carboncini MC, et al. Objective and automatic classification of Parkinson disease with Leap Motion controller. Biomed Eng Online. 2018 Nov 12;17(1):168.
[3] Ishii N, Mochizuki Y, Shiomi K, Nakazato M, Mochizuki H. Spiral drawing: Quantitative analysis and artificial-intelligence-based diagnosis using a smartphone. J Neurol Sci. 2020 Apr 15;411:116723.

« Back to 2022 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/machine-learning-algorithms-for-classification-and-analysis-of-pen-on-paper-archimedes-spirals/