3.4 Do trial-to-trial measures display systematic change?
The Curve Fitting app and functions of the Curve Fitting Toolbox in
MATLAB (2020b) were used to perform exploratory analyses to determine if
any of the components’ ST amplitude or latency variability exhibit
systematic change across successive trials. The curve fitting process
for each component measure (e.g., ST baseline-to-peak P1 amplitude)
started with averaging the measure across those participants with
correct responses for the first trial and then averaging separately for
each successive trial in a similar manner through trial number 480. This
averaging procedure was also completed for participant response times
(RT) for each correct trial. Scatterplots were then made with
cross-participant mean amplitude (or latency) plotted on the y-axis as a
function of trial number on the x-axis separately for each block (i.e.,
Trials 1-240 and Trials 241-480) and then with both blocks combined
(i.e., Trials 1-480). Then to determine if these distributions of means
contained any change in values as a function of trial number, the
scatterplot of each ST component was evaluated using four Sum of Sine
models where in each successive model the phase constant for the model’s
sine wave term was incremented from 1 to 4 unless there was evidence of
overfitting of the curve as depicted in Figure 1. Overlays of the single
trial mean amplitudes derived from
STP baseline-to-peak, STP
peak-to-peak, and STW approaches are shown in Figure 2 for each
component. For comparison purposes, the curve fit of the RT are also
shown in Figure 2. Non-linear patterns were found in all components of
STP baseline-to-peak and STW approaches with the most dramatic changes
in those from the STP peak-to-peak approach. The single trial mean
latencies for STP approach are shown in Figure 3 along with mean RT.
The amount of variance (i.e., R 2 values)
accounted for by each fitted curve depicted in the figures are reported
in Table 11. For the
baseline-to-peak and peak-to-peak measures from STP approach, all were
successfully fitted to sine wave value of 4 except for one. TheR2 values of the peak-to-peak measures were
generally larger than those of the baseline-to-peak measures. In
contrast, only two of the amplitude measures from the STW approach could
be fitted to sine wave value of 4 and all but two of theR2 values of measures from the STW approach
were lower than those from the STP approach.