Although EEG is designed to record cerebral activity, it also records electrical activities arising from sites other than the brain. These unwanted chunks of electrical activity are termed artifacts and may be divided into physiological and non-physiological artifacts. Physiological artifacts are generated from sources other than the brain (i.e., any other part of the body like the eyes, other voluntary and involuntary muscles) and involve head or limb motion and muscular tension. On the other hand, non-physiological artifacts arise from outside the body (i.e., equipment, environment etc.) and involve electromagnetic perturbations from other devices used in the experiment and/or leaking power line contamination, etc.
The chances …show more content…
of detecting and correcting artifacts are higher when the sources of disturbance are monitored by a dedicated measurement. Hence electrophysiological monitoring (ECG, EOG, EMG, etc.) is strongly encouraged in most experimental settings. It is of course always helpful if the artifact is prevented from occurring in the first place by following simple rules such as disallowing contact lenses or giving frequent breaks between tasks to participants.
For steady-state perturbations, which are thought to be independent of the brain processes of interest, approaches like principal or independent component analysis prove effective because of their statistical approach.
ARTIFACT CORRECTION V/S ARTIFACT REJECTION
Most artifacts have frequencies much higher than brain frequencies (e.g. muscle artifacts), which is why they are most likely to be eliminated during filtering. However, this may vary depending largely on the kind of data that is of interest. The problematic artifacts are the ones whose frequencies tend to overlap or match the brain frequencies, especially the ones that appear recurrently and persistently in the spectrum (e.g. ocular artifacts).
Pros of Artifact Correction
While artifact rejection may seem the more feasible procedure as unwanted noise altogether adds to inaccuracy of ERP data, it is also a more crude process as it completely eliminates an entire subset of trials from the ERP averages. As Gratton, Coles and Donchin (1983) discussed, there are three potential problems associated with rejecting trials with ocular artifacts:
1. The Quality of the Data - In some cases, discarding trials with eye blinks and eye movements (even muscular movement) might lead to an unrepresentative sample of trials simply because ocular activity is much too frequent and repetitive and ocular activity may actually overlap with cortical activity.
2. Experimental Design - There are some experimental paradigms in which blinks and eye movements are integral to the tasks (likewise with muscular movements), and rejecting trials with these artifacts would be counterproductive. Whatever the experimental design, completely eliminating confounds is not entirely possible.
3.
Specific Participant Populations – There are some groups of subjects (e.g. children and psychiatric patients) who cannot easily control their blinking, eye movements and/or muscular movements (e.g. epileptic and Parkinson patients) making it difficult to obtain a sufficient number of artifact-free trials. Such populations are in fact more likely to exhibit exaggerated movement.
Cons of Artifact Correction
Researchers have developed several artifact correction procedures (Berg & Scherg, 1991a, 1994; Gratton, Coles, & Donchin, 1983; Lins et al., 1993b; Verleger, Gasser, & Moecks, 1982).
The approach of subtracting voltages from ERP waveforms can be a problematic one. Specifically, the EOG recording contains brain activity in addition to true ocular activity and, as a result, the subtraction procedure ends up subtracting away part of the brain’s response as well as the ocular artifacts.
Although most artifact correction techniques can be useful or even indispensable for certain tasks and certain types of subjects, they have some significant …show more content…
drawbacks:
1.
Some of these techniques can significantly distort the ERP waveforms and scalp distributions, making the data difficult to interpret. A newer and promising approach is the independent components analysis (ICA). This approach is well justified mathematically, and recent studies demonstrated that this technique works very well at removing blinks, eye movements and even electrical noise. However, this approach assumes that the time course of the artifacts is independent of the time course of the ERP activity, which may not always be a correct assumption. Until an independent laboratory rigorously tests this technique, it will be difficult to know whether this sort of situation leads to significant distortions.
2. Artifact correction techniques may require significant additional effort. For example, Lins et al. recommended that recordings should be obtained from at least seven electrodes near the eyes. In addition, one must conduct a set of calibration runs for each subject and carry out extensive signal processing on the data. Thus, it is important to weigh the time saved by using artifact correction procedures against the time required to satisfactorily implement these
procedures.
3. A third problem with these techniques is that they cannot account for the changes in sensory input caused by blinks and eye movements. For example, if a subject blinks at the time of a visual stimulus, then this stimulus may not be seen properly, and this obviously cannot be accounted for by artifact correction techniques. In addition, as the eyes move, the visual world slides across the retina, generating a sensory ERP response. Similarly, eye blinks and eye movements are accompanied by motor ERPs. Artifact correction procedures do not typically address factors, which are especially problematic when task-relevant stimuli trigger the blinks or eye movements.
CONCLUSION
Determining whether artifact rejection should be used or artifact correction essentially depends on the type of ERP analysis that is to be ultimately carried out. Since, artifact correction has quite a few significant limitations, it usually is not the more recommended procedure of the two, unless the nature of the experiment or subject makes artifact rejection impossible. There is also the consideration of sample size, which is an important one. When correction is necessary, it is highly recommended that one of the newer and less error-prone techniques such as ICA or source localization are used, as with simpler techniques (often available in commercial ERP analysis packages) it is difficult to know the extent to which the artifact correction procedures distort the results.
REFERENCES
Pfurtscheller, G., & Lopes da Silva, F. H. (1999). Event-related EEG/MEG synchronization and desynchronization: basic principles. Clinical neurophysiology, 110(11), 1842-1857.
Jung, T. P., Makeig, S., Westerfield, M., Townsend, J., Courchesne, E., & Sejnowski, T. J. (2000). Removal of eye activity artifacts from visual event-related potentials in normal and clinical subjects. Clinical Neurophysiology, 111(10), 1745-1758.
Picton, T. W., van Roon, P., Armilio, M. L., Berg, P., Ille, N., & Scherg, M. (2000). The correction of ocular artifacts: a topographic perspective. Clinical Neurophysiology, 111(1), 53-65.
Joyce, C. A., Gorodnitsky, I. F., & Kutas, M. (2004). Automatic removal of eye movement and blink artifacts from EEG data using blind component separation. Psychophysiology, 41(2), 313-325.
Van Dun, B., Rombouts, G., Wouters, J., & Moonen, M. (2009). A procedural framework for auditory steady-state response detection. Biomedical Engineering, IEEE Transactions on, 56(4), 1098-1107.
http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_mathematik_und_naturwissenschaften/fachrichtung_psychologie/i1/allgpsy/lehre/lehreveranstaltungen/Lehre%20Kroenke%20Seminar%20KN/S7R1EEG.pdf