Comparison of two neural network classifiers in the differential diagnosis of essential tremor and Parkinson's disease by 123I-FP-CIT brain SPECT

Barbara Palumbo, Mario Luca Fravolini, Susanna Nuvoli, Angela Spanu, Kai Stephan Paulus, Orazio Schillaci, Giuseppe Madeddu

Research output: Contribution to journalArticle

Abstract

Purpose: To contribute to the differentiation of Parkinson's disease (PD) and essential tremor (ET), we compared two different artificial neural network classifiers using 123I-FP-CIT SPECT data, a probabilistic neural network (PNN) and a classification tree (ClT). Methods: 123I-FP-CIT brain SPECT with semiquantitative analysis was performed in 216 patients: 89 with ET, 64 with PD with a Hoehn and Yahr (H&Y) score of ≥2 (early PD), and 63 with PD with a H&Y score of ≥2.5 (advanced PD). For each of the 1,000 experiments carried out, 108 patients were randomly selected as the PNN training set, while the remaining 108 validated the trained PNN, and the percentage of the validation data correctly classified in the three groups of patients was computed. The expected performance of an "average performance PNN" was evaluated. In analogy, for ClT 1,000 classification trees with similar structures were generated. Results: For PNN, the probability of correct classification in patients with early PD was 81.9±8.1% (mean±SD), in patients with advanced PD 78.9±8.1%, and in ET patients 96.6±2.6%. For ClT, the first decision rule gave a mean value for the putamen of 5.99, which resulted in a probability of correct classification of 93.5±3.4%. This means that patients with putamen values >5.99 were classified as having ET, while patients with putamen values

Original languageEnglish
Pages (from-to)2146-2153
Number of pages8
JournalEuropean Journal of Nuclear Medicine and Molecular Imaging
Volume37
Issue number11
DOIs
Publication statusPublished - Nov 2010

Fingerprint

Essential Tremor
Single-Photon Emission-Computed Tomography
Parkinson Disease
Differential Diagnosis
Brain
Putamen
2-carbomethoxy-8-(3-fluoropropyl)-3-(4-iodophenyl)tropane

Keywords

  • I-FP-CIT SPECT
  • Classification tree
  • Neural network classifier
  • Parkinson's disease
  • Probabilistic neural network

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this

Comparison of two neural network classifiers in the differential diagnosis of essential tremor and Parkinson's disease by 123I-FP-CIT brain SPECT. / Palumbo, Barbara; Fravolini, Mario Luca; Nuvoli, Susanna; Spanu, Angela; Paulus, Kai Stephan; Schillaci, Orazio; Madeddu, Giuseppe.

In: European Journal of Nuclear Medicine and Molecular Imaging, Vol. 37, No. 11, 11.2010, p. 2146-2153.

Research output: Contribution to journalArticle

Palumbo, Barbara ; Fravolini, Mario Luca ; Nuvoli, Susanna ; Spanu, Angela ; Paulus, Kai Stephan ; Schillaci, Orazio ; Madeddu, Giuseppe. / Comparison of two neural network classifiers in the differential diagnosis of essential tremor and Parkinson's disease by 123I-FP-CIT brain SPECT. In: European Journal of Nuclear Medicine and Molecular Imaging. 2010 ; Vol. 37, No. 11. pp. 2146-2153.
@article{38da8976ac7c46229cd9219b8407025f,
title = "Comparison of two neural network classifiers in the differential diagnosis of essential tremor and Parkinson's disease by 123I-FP-CIT brain SPECT",
abstract = "Purpose: To contribute to the differentiation of Parkinson's disease (PD) and essential tremor (ET), we compared two different artificial neural network classifiers using 123I-FP-CIT SPECT data, a probabilistic neural network (PNN) and a classification tree (ClT). Methods: 123I-FP-CIT brain SPECT with semiquantitative analysis was performed in 216 patients: 89 with ET, 64 with PD with a Hoehn and Yahr (H&Y) score of ≥2 (early PD), and 63 with PD with a H&Y score of ≥2.5 (advanced PD). For each of the 1,000 experiments carried out, 108 patients were randomly selected as the PNN training set, while the remaining 108 validated the trained PNN, and the percentage of the validation data correctly classified in the three groups of patients was computed. The expected performance of an {"}average performance PNN{"} was evaluated. In analogy, for ClT 1,000 classification trees with similar structures were generated. Results: For PNN, the probability of correct classification in patients with early PD was 81.9±8.1{\%} (mean±SD), in patients with advanced PD 78.9±8.1{\%}, and in ET patients 96.6±2.6{\%}. For ClT, the first decision rule gave a mean value for the putamen of 5.99, which resulted in a probability of correct classification of 93.5±3.4{\%}. This means that patients with putamen values >5.99 were classified as having ET, while patients with putamen values",
keywords = "I-FP-CIT SPECT, Classification tree, Neural network classifier, Parkinson's disease, Probabilistic neural network",
author = "Barbara Palumbo and Fravolini, {Mario Luca} and Susanna Nuvoli and Angela Spanu and Paulus, {Kai Stephan} and Orazio Schillaci and Giuseppe Madeddu",
year = "2010",
month = "11",
doi = "10.1007/s00259-010-1481-6",
language = "English",
volume = "37",
pages = "2146--2153",
journal = "European Journal of Pediatrics",
issn = "0340-6199",
publisher = "Springer Berlin Heidelberg",
number = "11",

}

TY - JOUR

T1 - Comparison of two neural network classifiers in the differential diagnosis of essential tremor and Parkinson's disease by 123I-FP-CIT brain SPECT

AU - Palumbo, Barbara

AU - Fravolini, Mario Luca

AU - Nuvoli, Susanna

AU - Spanu, Angela

AU - Paulus, Kai Stephan

AU - Schillaci, Orazio

AU - Madeddu, Giuseppe

PY - 2010/11

Y1 - 2010/11

N2 - Purpose: To contribute to the differentiation of Parkinson's disease (PD) and essential tremor (ET), we compared two different artificial neural network classifiers using 123I-FP-CIT SPECT data, a probabilistic neural network (PNN) and a classification tree (ClT). Methods: 123I-FP-CIT brain SPECT with semiquantitative analysis was performed in 216 patients: 89 with ET, 64 with PD with a Hoehn and Yahr (H&Y) score of ≥2 (early PD), and 63 with PD with a H&Y score of ≥2.5 (advanced PD). For each of the 1,000 experiments carried out, 108 patients were randomly selected as the PNN training set, while the remaining 108 validated the trained PNN, and the percentage of the validation data correctly classified in the three groups of patients was computed. The expected performance of an "average performance PNN" was evaluated. In analogy, for ClT 1,000 classification trees with similar structures were generated. Results: For PNN, the probability of correct classification in patients with early PD was 81.9±8.1% (mean±SD), in patients with advanced PD 78.9±8.1%, and in ET patients 96.6±2.6%. For ClT, the first decision rule gave a mean value for the putamen of 5.99, which resulted in a probability of correct classification of 93.5±3.4%. This means that patients with putamen values >5.99 were classified as having ET, while patients with putamen values

AB - Purpose: To contribute to the differentiation of Parkinson's disease (PD) and essential tremor (ET), we compared two different artificial neural network classifiers using 123I-FP-CIT SPECT data, a probabilistic neural network (PNN) and a classification tree (ClT). Methods: 123I-FP-CIT brain SPECT with semiquantitative analysis was performed in 216 patients: 89 with ET, 64 with PD with a Hoehn and Yahr (H&Y) score of ≥2 (early PD), and 63 with PD with a H&Y score of ≥2.5 (advanced PD). For each of the 1,000 experiments carried out, 108 patients were randomly selected as the PNN training set, while the remaining 108 validated the trained PNN, and the percentage of the validation data correctly classified in the three groups of patients was computed. The expected performance of an "average performance PNN" was evaluated. In analogy, for ClT 1,000 classification trees with similar structures were generated. Results: For PNN, the probability of correct classification in patients with early PD was 81.9±8.1% (mean±SD), in patients with advanced PD 78.9±8.1%, and in ET patients 96.6±2.6%. For ClT, the first decision rule gave a mean value for the putamen of 5.99, which resulted in a probability of correct classification of 93.5±3.4%. This means that patients with putamen values >5.99 were classified as having ET, while patients with putamen values

KW - I-FP-CIT SPECT

KW - Classification tree

KW - Neural network classifier

KW - Parkinson's disease

KW - Probabilistic neural network

UR - http://www.scopus.com/inward/record.url?scp=77958494385&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77958494385&partnerID=8YFLogxK

U2 - 10.1007/s00259-010-1481-6

DO - 10.1007/s00259-010-1481-6

M3 - Article

C2 - 20567820

AN - SCOPUS:77958494385

VL - 37

SP - 2146

EP - 2153

JO - European Journal of Pediatrics

JF - European Journal of Pediatrics

SN - 0340-6199

IS - 11

ER -