PL EN


Preferencje help
Widoczny [Schowaj] Abstrakt
Liczba wyników
2017 | 77 | Suppl.1 |

Tytuł artykułu

Performance of a matching pursuit based classifier on vep: a potential parameter for assessment of the state of patients with disorders of consciousness

Warianty tytułu

Języki publikacji

EN

Abstrakty

EN
INTRODUCTION: P300 event‑related potential reflects the brain response to external stimuli. Attention paid to one of many repeatedly presented stimuli can be detected from the relative strengths of the responses; if the subject actively counts the occurrences of one of the stimuli (target), specific waveform responsе, namely the P300 potential, may be observed, and it will be absent during the unattended (non‑target) stimuli. This difference can be used as an indicator of conscious information processing in unresponsive patients suffering from disorders of consciousness (DoC). However, P300 waveforms recorded from those patients may significantly differ from the classical shape known from healthy subjects. AIM(S): We test the possibility of replacing the classical indicators used for assessing the difference between responses to target and non-target stimuli by cross-validation of a classifier detecting responses, based upon multivariate matching pursuit (MMP) parameterization. METHOD(S): Visual P300 potentials were recorded in a standard paradigm from a group of healthy subjects and patients in different states of disorder of consciousness. MMP algorithm was used as parametrization, and based upon a subset of recorded data a classifier was trained to distinguish responses to target and non-target stimuli. RESULTS: Cross‑validation performance of the classifier measured as the area under corresponding ROC curves discriminates the healthy group from DoC patients, and in some cases correlates with the severity of DoC. CONCLUSIONS: Replacing estimation of the statistical significance of the average P300 amplitudes by the performance of MMP‑based classification assessed by cross‑validation allows for nonparametric detection of conscious responses in DoC patients, whose ERPs do not always exhibit the classical components. FINANCIAL SUPPORT: This research was supported by the Polish National Science Centre grant 2015/17/B/ ST7/03784.

Słowa kluczowe

Wydawca

-

Rocznik

Tom

77

Numer

Opis fizyczny

p.64-65

Twórcy

autor
  • University of Warsaw, Faculty of Physics, Warsaw, Poland
autor
  • University of Warsaw, Faculty of Physics, Warsaw, Poland
autor
  • University of Warsaw, Faculty of Physics, Warsaw, Poland
autor
  • University of Warsaw, Faculty of Physics, Warsaw, Poland
autor
  • University of Warsaw, Faculty of Physics, Warsaw, Poland
  • University of Warsaw, Faculty of Physics, Warsaw, Poland
autor
  • University of Warsaw, Faculty of Physics, Warsaw, Poland

Bibliografia

Typ dokumentu

Bibliografia

Identyfikatory

Identyfikator YADDA

bwmeta1.element.agro-754aadaf-400c-49ad-9c98-4e6d9f95a0b6
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.