With this strategy, pieces of data contaminated by artifacts are identified and removed from further analysis. Having the data in memory after segmenting is an efficient way of visually browsing through the data segments of interest, especially if you know that there are artifacts in between the trials (e.g., EOG artifacts in the blink period).ĭetecting the artifacts without reading the complete data into memory allows you to work with datasets that are too large to fit in memory all at once.
segmented with time gaps between the segments) and whether your data is stored on disk or already read into memory.ĭetecting the artifacts in continuous data prior to segmenting allows you to apply filters (e.g., a band-pass filter to zoom in on the muscle artifacts on the temporal channels) without having to worry about edge effects due to the filter (i.e. The functions in FieldTrip that you can use for artifact detection depend on whether your data is continuous or trial-based (i.e. subtracting the spatio-temporal contribution of the artifact from the data, e.g., using a filter or ICA.rejecting the piece of data containing the artifact, e.g., for a short-lived artifact or poorly attached EEG electrode, or.After you know what the artifacts are, they are removed by either Detection of artifacts can be done visually, or using automatic routines, or a combination of both. How does FieldTrip manage artifacts?įieldTrip deals with artifacts by first identifying them, and subsequently removing them. There is not a single most optimal manner to detect the artifacts: it depends on the data properties, the type of artifacts that you anticipate to be present (given your recording setup, your task, and your participants) and your own preferences. Behavioral artifacts are typically short-lived, whereas environmental and instrumentation artifacts are typically more persistent. The corresponding EEG channel will show a flat line, or potentially a lot of noise.īesides considering the physiological or non-physiological aspect of the artifact, you can also think about whether the artifact is caused by the behavior of the participant (e.g., an eye movement), whether it is caused by something in the environment (e.g., 50Hz line noise) or whether it is caused by a malfunction of the equipment (e.g., a poorly attached electrode). An example of a non-physiological artifact is an EEG electrode that has poor contact with the scalp. The contribution of this so-called electrooculogram is mainly visible on frontal electrodes, but when looking carefully enough, you can see it on all electrodes. The retina is electrically charged and the movement of the eye causesa a deflection of the scalp potential. Artifacts can be physiological or non-physiological in origin.Īn eyeblink is an example of a physiological artifact that shows up in the EEG. Generally speaking, an artifact (American English spelling) or artefact (British English spelling) is some unexpected or unwanted feature in the data that we acquired with our EEG or MEG system.
Therefore you should be aware of the different approaches and of the variability of artifact rejection (automatic/manual) procedures described onwards.Īt the end of an automated procedure, consider always to visual inspect your data, after rejection. Since FieldTrip supports the data of many different acquisition systems, the particular artifacts in your data might behave very different from the examples demonstrated and discussed here.
0 kommentar(er)
