Studies Included in the Analysis
For the purpose of the planned analyses, and to minimize potential
confound effects, all the data used in the main analyses were recorded
in the Rosenfeld Lab, with the same hardware and with the same version
of P300-CIT – the Complex Trial Protocol, which is the most modern and
countermeasure-resistant version of the P300-based CIT (Rosenfeld
et.al., 2004; Rosenfeld et. al. 2008). Since 2008, many studies have
demonstrated high sensitivity and specificity in the CTP (over 90% for
autobiographical information and over 80% for incidentally acquired
information; with the AUC typically over .9) (see Rosenfeld et.al, 2013
for review).
In the Complex Trial Protocol (CTP), contrary to the traditional
three-stimulus protocol, the probe, irrelevant and target are separated
into two parts (Rosenfeld, et.al., 2008). In the first phase, probes and
irrelevants are presented and the participants are asked to make a
button response to indicate that they saw the stimulus (same button for
both stimuli). In the second phase, targets and non-targets are
presented during which the participants are instructed to respond
differently to targets compared to non-targets. These targets and
non-targets are usually a series of digits that are not relevant to the
first phase (e.g., “11111” is the target, and “22222”, “33333”,
“44444”,”55555”, “66666” are non-targets), such that more
cognitive resources can be dedicated to phase 1, where the diagnostic
analysis takes place. This protocol is to be used as an example
throughout the paper. So, data collected from the CTP can further be
analyzed through statistical tools like Bootstrapping to make
classification results more accurate and effective.
Critically, we reanalyzed the data from Rosenfeld et al. (2017b) using
our rBS technique, specifically regarding the following two studies:
1) Rosenfeld, et.al (2017a) (Experiment 2 - n=52 guilty participants
with semantic stimuli) – this study explored the possibility of a
memory suppression effect on P300 amplitude while using the Complex
Trail Protocol with semantic stimuli. No differences in amplitude or
latency were found between suppression versus non-suppression groups, so
all participants were pooled in Rosenfeld et al. (2017b). Consequently,
we also merged suppression and non-suppression groups in the current
analysis.
2) Ward, Rosenfeld (2017) (n=29 guilty participants with episodic
stimuli) – this experiment also verified the memory suppression effect
on P300 amplitude in the CTP, however with episodic memory. Since there
were no differences in P300 amplitude or latency between experimental
groups, data were pooled in Rosenfeld et al. (2017b) and also in the
current analysis.
Please refer to Rosenfeld et al., (2017a) and Ward & Rosenfeld (2017)
for details on participant recruitment and EEG/ERP data processing
methods.