Article

Technical Note Enhanced detection of artifacts in EEG data using higher-order statistics and independent component analysis

Computational Neurobiology Laboratory, Salk Institute for Biological Studies, 10010 N. Torrey Pines Road, La Jolla, CA 92107, USA.
NeuroImage (Impact Factor: 6.36). 03/2007; 34(4):1443-9. DOI: 10.1016/j.neuroimage.2006.11.004
Source: PubMed

ABSTRACT Detecting artifacts produced in EEG data by muscle activity, eye blinks and electrical noise is a common and important problem in EEG research. It is now widely accepted that independent component analysis (ICA) may be a useful tool for isolating artifacts and/or cortical processes from electroencephalographic (EEG) data. We present results of simulations demonstrating that ICA decomposition, here tested using three popular ICA algorithms, Infomax, SOBI, and FastICA, can allow more sensitive automated detection of small non-brain artifacts than applying the same detection methods directly to the scalp channel data. We tested the upper bound performance of five methods for detecting various types of artifacts by separately optimizing and then applying them to artifact-free EEG data into which we had added simulated artifacts of several types, ranging in size from thirty times smaller (-50 dB) to the size of the EEG data themselves (0 dB). Of the methods tested, those involving spectral thresholding were most sensitive. Except for muscle artifact detection where we found no gain of using ICA, all methods proved more sensitive when applied to the ICA-decomposed data than applied to the raw scalp data: the mean performance for ICA was higher and situated at about two standard deviations away from the performance distribution obtained on raw data. We note that ICA decomposition also allows simple subtraction of artifacts accounted for by single independent components, and/or separate and direct examination of the decomposed non-artifact processes themselves.

Download full-text

Full-text

Available from: Arnaud Delorme, Aug 30, 2015
1 Follower
 · 
197 Views
    • "During autocalibration , optimization aimed at collecting N=7 artifact free trials for each mental task. Artifact and outlier rejection was based on statistical metrics of EEG features (thresholding of amplitude, kurtosis and probablity of 4-40 Hz band pass filtered EEG, [14]) and spectral components [7]. BP C ch,f b were extracted from clean imagery EEG segments and grouped according to mental task pair combinations C. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Non-stationarity and inherent variability of the noninvasive electroencephalogram (EEG) makes robust recognition of spontaneous EEG patterns challenging. Reliable modulation of EEG patterns that a BCI can robustly detect is a skill that users must learn. In this paper, we present a novel online co-adaptive BCI training paradigm. The system autonomously screens users for their ability to modulate EEG patterns in a predictive way and adapts its model parameters online. Results of a supporting study in seven first-time BCI users with disability are very encouraging. Three of 7 users achieved online accuracy > 70% for 2-class BCI control after 24 minutes of training. Online performance in 6 of 7 users was significantly higher than chance level. Online control was based on one single bipolar EEG channel. Beta band activity carried most discriminant information. Our fully automatic co-adaptive online approach allows to evaluate whether user benefit from current BCI technology within a reasonable timescale.
    37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC’15), Milano, Italy; 08/2015
  • Source
    • "Epochs with irregular noise were identified and rejected using a computer algorithm based on abnormal statistical distribution, as well as by inferences from visual inspection (Delorme et al., 2007). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Recent studies have demonstrated neural overlap between resting state activity and self-referential processing. This "rest-self" overlap occurs especially in anterior cortical midline structures like the perigenual anterior cingulate cortex (PACC). However, the exact neuro-temporal and biochemical mechanisms remain to be identified. Therefore, we conducted a combined EEG-MRS study. EEG focused on pre-stimulus (e.g., prior to stimulus presentation or perception) power changes to assess the degree to which those changes can predict subjects' perception (and judgment) of subsequent stimuli as high or low self-related. Magnetic resonance spectroscopy (MRS) measured resting state concentration of glutamate, focusing on PACC. High pre-stimulus (e.g., prior to stimulus presentation or perception) alpha power significantly correlated with both perception of stimuli judged to be highly self-related, and with resting state glutamate concentrations in the PACC. In sum, our results show (i) pre-stimulus (e.g., prior to stimulus presentation or perception) alpha power and resting state glutamate concentration to mediate rest-self overlap which (ii) dispose or incline subjects to assign high degrees of self-relatedness to perceptual stimuli.
    Social neuroscience 07/2015; DOI:10.1080/17470919.2015.1072582 · 2.87 Impact Factor
  • Source
    • "Many traditional approaches have been proposed to remove or attenuate artifacts from recorded EEG signals [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. The most widely used methods for attenuating artifacts in EEG signals are based on blind source separation such as independent component analysis (ICA) and canonical correlation analysis (CCA) [20], [11], [21], [22], [23], [14], [15], [24]. The BSS-based algorithms assume that the observations are linear mixing of the sources and the number of sources is equal or less than the number of observations. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a method to reduce artifacts from scalp EEG recordings to facilitate seizure diagnosis/detection for epilepsy patients. The proposed method is primarily based on stationary wavelet transform and takes the spectral band of seizure activities (i.e. 0.5 - 29 Hz) into account to separate artifacts from seizures. Different artifact templates have been simulated to mimic the most commonly appeared artifacts in real EEG recordings. The algorithm is applied on three sets of synthesized data including fully simulated, semi-simulated and real data to evaluate both the artifact removal performance and seizure detection performance. The EEG features responsible for the detection of seizures from non-seizure epochs have been found to be easily distinguishable after artifacts are removed and consequently the false alarms in seizure detection are reduced. Results from an extensive experiment with these data-sets prove the efficacy of the proposed algorithm, which makes it possible to use it for artifact removal in epilepsy diagnosis as well as other applications regarding neuroscience studies.
    IEEE Journal of Biomedical and Health Informatics 07/2015; DOI:10.1109/JBHI.2015.2457093 · 1.98 Impact Factor
Show more