Pushing the limits of BCI accuracy: Winning solution of the Grasp & Lift EEG challenge - Archive ouverte HAL Access content directly
Conference Papers Year : 2016

Pushing the limits of BCI accuracy: Winning solution of the Grasp & Lift EEG challenge

Abstract

To better understand the relationship between EEG signals and hand movements the WAY Consortium has organized the Grasp-and-Lift EEG Detection challenge. It was held in 2015 from 29th June to 31th August on Kaggle, a platform for competitive predictive modeling, and attracted 379 contesting teams. The goal of the challenge was to detect 6 different events related to hand movement during a task of grasping and lifting an object, using only EEG signal. The 6 events were representing different stages of a sequence of hand movements (hand starts moving, starts lifting the object, etc.). True labels were extracted from EMG signal, and provided as a +/-150ms frame centered on the occurrence of the event. Contestants were asked to provide probabilities of detection for the 6 events and for every time sample. The evaluation metric for this challenge was the Area Under the ROC Curve (AUC) averaged over the 6 event types. Finally, the model must be causal, i.e. only the data from the past can be used to predict the events. This abstract presents the winning solution of this challenge.
Fichier principal
Vignette du fichier
BCI_Meeting_barachant_Cycon.pdf (288.16 Ko) Télécharger le fichier
poster_kaggle_barachant.pdf (3.01 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Origin : Files produced by the author(s)

Dates and versions

hal-01349562 , version 1 (27-07-2016)

Identifiers

  • HAL Id : hal-01349562 , version 1

Cite

Alexandre Barachant, Rafał Cycon. Pushing the limits of BCI accuracy: Winning solution of the Grasp & Lift EEG challenge. 6th International Brain-Computer Interface Meeting, May 2016, Monterey, United States. ⟨hal-01349562⟩
291 View
619 Download

Share

Gmail Facebook X LinkedIn More