The Effects of Acute Moderate Exercise on Cognitive Functioning — Part 6
Analysis Methodology 3: RIDE & ICA & Average ERP Waveforms
A variety of strategies have been employed to improve the processing of ERPs. Algorithms that decompose ERPs into components, including independent component analysis (ICA) and principal component analysis (PCA), are intended to separate signal from noise but are not specifically designed to handle the variability in latency across trials (Ouyang et al., 2016). Reaction time binning has also been used to combat the issue of latency variability (Poli et al., 2010; Verleger et al., 2005). The downside of these methodologies is that the P3b waveform is not absolutely locked to the stimulus or response (Ouyang et al., 2011; Verleger et al., 2005). Now however, a new algorithm has been proposed to deal with the ISV of ERP components, and specifically P3. This new algorithm is referred to as ‘Residue Iteration Decomposition’ (RIDE) (Ouyang et al., 2011).
According to Ouyang and colleagues (2017), ISV is largely unexplored because the majority of ERP analysis has focused on the established method of finding the mean amplitude and latency — that is, identifying the amplitude and latency from the average waveform. From a methodological point of view, there are flaws with this approach because the mean amplitude of an ERP component is likely to be underestimated when averaging due to variability in the latency across trials (Walhovd et al., 2008; Ouyang et al., 2015). RIDE (which works in conjunction with EEGLAB) enables researchers to use trial-specific amplitudes and latencies to determine ISV, as well as compose a more accurate average ERP waveform (Delorme & Makeig, 2004; Ouyang et al., 2011). The purpose of this algorithm is three-fold: 1) to decompose ERPs into a collection of ERP components and determine their latency variabilities; 2) reconstruct average ERPs to more accurately represent the single trial ERPs; and 3) estimate single trial latency variability (Ouyang et al., 2015). By estimating the single trial latency, ERP components of interest can be aligned so that the average ERP waveform can be more accurately re-constructed (Ouyang et al., 2015).
In this example, ERP data preprocessed using the RIDE algorithm will be prepared for statistical analysis. Some preliminary visualizations will be used for later comparison against the average ERP waveforms created in part 4.
- Load in the RIDE processed data
2. Organize the data and average across participant, condition, time, and flanker task
Similarly to part 4 and 5, we generate the individual average ERP waveforms.
3. Plot one individual’s average ERP waveform across conditions, times, and flankers
MCT here refers to the moderate exercise condition
1 & 2 refer to the congruent and incongruent flanker conditions respectively
4. Plotting the same individual’s data with a smaller time window
Considering that we are primarily interested in the P3 component and the time window it typically falls within, we can reduce our time window to get a better sense of the neural activity within our time frame of interest.
5. Plotting the grand average ERP waveform
For further details on how RIDE works see Ouyang et al.’s paper.
