Applications of EEG Neuroimaging Data: Event-related Potentials, Spectral Power, and Multiscale Entropy

The overall goal of this procedure is to quantify the variability of an EEG time series and relate that variability to the information processing capacity of the underlying neural system. This is accomplished by first acquiring high quality EEG recordings of the brain's response. The second step is to pre-process the data to remove any artifacts.

Next, the statistics of interest are extracted. Here we will contrast the novel application of multi-scale entropy with more traditional methods of mean amplitude and spectral power. The final step is to analyze the statistical significance of the results and interpret the data.

This step can be facilitated by using data-driven multi-variate approaches like partial lease square's analysis. Ultimately, multi-scale entropy is used to show how a sequence of changes in the spatial temporal pattern across multiple times scales contributes to specific cognitive operations. The advantage to using MSC over existing methods like mean amplitude or spectra power, is that MSC is sensitive to the non-linearities in the data.

These non-linear dynamics reflect transitions or bifurcations between the micro states of a network, which is important for information exchange across a distributed network of brain regions. Demonstrating this procedure will be Christina Backer from the ERP lab at the Rotman Research Institute. First, explain the experimental procedures to the participant and obtain informed consent.

Clean the area where the dropdown electrodes will be placed using an alcohol swab. Place some gel on the electrode. Take off the papers from the skin side and place the electrodes on the participant to identify eye movement artifacts.

Place an electrode on the lateral junction of the upper and lower of the eyelid. Place another electrode on the center of the orbital ridge, approximately one centimeter below the eye and in line with the pupil. Repeat for the other eye.

Measure the participant's head circumference and choose the appropriate electrode cap size following the internationally recognized 10 20 system. For electrode placement, measure the distance from Indian Ian along the midline and divide by 10%Using that number. Measure up from NAS and mark.

Align the electrode cap position FP with this mark and pull the cap back. Make sure that the center of the cap is in line with the nose. Measure nasn to cz and confirm that this distance is half the distance from NAS to Indian.

Then tighten the chin strap and place gauze under the strap for comfort if necessary. Now, put the gel filled blunt point syringe in the electrode holders to create a conductive column of gel. Start in contact with the scalp, then squeeze and pull back.

Note that the application of too much gel may bridge the signals of neighboring electrodes. Next, fix active electrodes into the electrode holders. Then position the subject in front of the monitor at the appropriate distance.

For the experiment, ask the participant to remain still emphasizing the importance of minimizing eye movements and blinks. For a clean recording, examine the electrode connections and EEG signal quality on the acquisition computer. If there is a problem with a particular electrode, take out that electrode and reapply gel to adjust impedances at that site after experimentation, but before extracting the particular statistic of interest.

Pre-process the continuous EEG data to remove artifacts using standard procedures of filtering and artifact Rejection. Event related potential analysis captures the synchronous brain activity that is phase lock to the onset of a stimulus Time lock the brain's response to the onset of a salient event, and then average over many like events. To increase the signal to noise ratio, identify the peak amplitude and latency of the ERP component for each subject, Spectra power quantifies the relative contribution of a frequency to a particular EEG signal.

Use Fourier analysis to transform the EEG signal from the time domain to the frequency domain and decompose the signal into its component sine waves of varying frequencies. Multi-scale entropy is an information theoretic metric that captures the variability of neuro electrical signals over time and across multiple times.Scales. Use the algorithm at PhysioNet to compute multi-scale entropy in two steps.

In the first step, progressively downs. Sample the signal for each trial and condition timescale. One represents the original signal.

Create subsequent timescales by first dividing the original signal into non-overlapping windows of the timescale length. Then average the data points within each window. For example, to create timescale two average together.

The first two points, the next two points, and so on. To create timescale three average together the first three points, the next three points and so on. By representing the original signal at various timescales, neural processes that may be unfolding at different rates can be analyzed.

The second step calculates sample entropy for each course. Grained time series. This provides an estimate of the complexity of the brain's response at the different timescales.

Regular signals have lower sample entropy than more stochastic signals. In this example, the pattern length M is set to two. This means that the time series will be represented as a ratio of two versus three point sequence matches.

Parameter R is the similarity criterion. Data points that lie within this amplitude range have similar values and thus are said to match. For details on setting parameters, consult the text protocol to calculate sample entropy for this simulated time series.

Begin with the first two components. Sequence pattern, red orange, first count the number of times the red orange sequence pattern occurs in the time series. There are 10 matches for this.

Two components sequence. Second count the number of times the first three components. Sequence pattern, red, orange, purple occurs in the time series.

There are five matches for this. Three component sequence. Continue with the same operations for the next two component sequence, orange, purple, and the next three component sequence, orange, purple, green of the time series.

Add the number of two component matches and three component matches for these sequences to the previous values. Repeat for all other sequence matches in the time series to determine the total ratio of two component matches to three component matches. Sample entropy is the natural logarithm of this ratio.

These data show condition differences in ERP spectral power and entropy contrasting the initial versus repeated presentation of facial photographs. In this example, all measures converged to reveal the same effect of decrease in sample entropy that accompanies face repetition. This decrease in complexity suggests that the functional network engaged is simpler and processing less information.

These statistical results are derived from the multi-variate analysis of partial lease squares of ERP spectral power and multi-scale entropy for faces associated with different levels of familiarity. The contrast shows that the ERP amplitude distinguished new faces from familiar faces, but not among the familiar faces that varied in amount of prior exposure. Spectral power distinguished faces according to the familiarity, but did not accurately distinguish between faces of low and medium familiarity.

Multi-scale entropy was most sensitive to the condition differences. Sample entropy values increased with increasing face familiarity. These image plots capture the spatial temporal distribution of the condition effect.

Interestingly, Multiscale entropy revealed unique information that was not obtained by the more traditional analyses of ERP or spectral power. This divergence of multi-scale entropy suggests that the conditions differ with respect to non-linear aspects of their network dynamics possibly involving the interactions between frequency bands. This novel analytical tool helps us to capture new information about the neural network dynamics, helping to move us away from characterizing mental function in terms of static states and towards understanding the fluid unfolding of processes that link to human cognition.