Data analysis often occurs locally on a researcher’s office computer by accessing files stored in Lagringshotell. Recently, VDI’s have become a popular tool for preprocessing and data analysis.
VDI – Virtual Desktop Infrastructure. A VDI offers users access to a virtual computer with the software and processing power they need. This computer can be used in the same way you use your local computer but can be reached from different devices and operating systems. Which programs that are mounted and can run on the VDI machine is decided together by yourself, your local IT and program managers at the departments. A VDI may offer advantages over using one’s office desktop computer for analysis, as the VDI processing capabilities are more powerful. The lab engineer can be contacted for more information about how to connect and use UiO's VDI's.
Software
Many of the available EEG systems come with analysis software packages with varying levels of detailed descriptions of how the different preprocessing tools are implemented. In addition, several free and commercially available software packages that run on MATLAB/Python/R platforms, offer alternative implementations of data analysis tools. At the psychology department at UiO, MATLAB is the dominant software used for EEG data preprocessing, analysis, and visualization. MATLAB is a high-level programming environment that is relatively easy to learn and use. Several of the most widely used EEG analysis packages are MATLAB toolboxes (e.g. EEGLAB; FieldTrip). MATLAB has a development environment that makes it easy to access large amounts of data. This is advantageous because one can easily inspect data during each stage of processing and analysis. One can also easily inspect and compare results from different subjects. MATLAB is also able to compute plots of your data that are customizable. These can be exported as pixel-based image files (.jpg, .bmp, .png, .tiff), vector files (.eps), or movies, which can be used to make presentations and publication-quality figures. In addition, custom-written software for preprocessing and subsequent analysis is commonly used. In-house software is usually notated by the researcher with descriptions and comments on the script.
Preprocessing and Analysis Considerations
EEG data has a low signal to 
noise ratio. Artifacts can occur due to participant movement, blinking, cardiac function, scalp muscles, swallowing or even the temporary dislocation of electrodes. Filters to remove artifacts should be used with caution as they may also distort the activity of interest (St. Louis & Frey, 2016). Using the data from the derivative channels for eye blinks and vertical eye moment, one can perform independent components analysis (ICA) to clean these artifacts from the data (Acheson, 2019).
The primary derivatives of EEG are evoked potentials and event-related potentials. Evoked potentials (EP) is a electrical potential in following a stimulus recorded/seen in spontaneous EEG and is NOT time-locked to the event. In the case of EP, one averages the electrical activity within a specific period in which a stimulus was presented. Time-locked/event-related potentials (ERP) are measured brain responses that are a direct result of stimulus presentation and are time-locked to the event.
The next step in preparing EEG data for analysis is to filter it to remove low and high signal frequencies. Using a bandpass filter, one can achieve a more smooth, defined signal. The data must then be segmented into epochs that correspond to the onsets of events of interest in the experiment. This is done using the event codes from the data file (Acheson, 2019). Epochs can also include an entire trial, made up of multiple events, if the data analysis strategy calls for it.
Next, these epochs are put on the same scale by averaging the signal from a short time prior to the onset, or the ‘baseline period’ and subtracting the average from every time-point in the epoch. You will also want to remove any epochs or signals that carry values too extreme, which are likely due to artifacts or a dysfunctional electrode. These will otherwise distort the effect (Acheson, 2019).
The exact details of analysis will vary by study, but typically the next steps in analysis entail averaging signals in the events of interest within subjects, followed by averaging these averages across subjects (ibid).
In most cases, when you have EEG data in several different file formats, loading the data into the analysis software converts the files so they conform to the standard format that is native for the program. However, in some cases you may have to convert these files manually.
Behavioral Data
Additional behavioral data can be acquired from the stimulus presentation software. This data can be used in conjunction with the EEG data during analysis. Their use tends to vary according to the study’s needs. For example, you may require information about a participant's responses or response times in order to understand the patterns of activation in your EEG data. Presentation, Psychtoolbox for MATLAB and E-Prime tend to be the most popular programs for experiments in labs, however open source programs like PsychoPy are coming more and more into use. Paradigm, PEBL, and Inquisit are additional programs for creating neuroscience experiments that you may encounter.