Personality and Augmenting/Reducing (A/R) in auditory event-related potentials (ERPs) during emotional visual stimulation

An auditory augmenting/reducing ERP paradigm recorded for 5 intensity tones with emotional visual stimulation was used, for the first time, to test predictions derived from the revised Reinforcement Sensitivity Theory (rRST) of personality with respect to two major factors: behavioral inhibition system (BIS), fight/flight/freeze system (FFFS). Higher BIS and FFFS scores were negatively correlated with N1/P2 slopes at central sites (C3, Cz, C4). Conditional process analysis revealed that the BIS was a mediator of the association between the N1/P2 slope and the FFFS scores. An analysis of covariance showed that lower BIS scorers exhibited larger N1/P2 amplitudes across all tone intensities while watching negative, positive and neutral pictures. Additionally, lower FFFS scorers compared to higher FFFS scorers disclosed larger N1/P2 amplitudes to the highest tone intensities and these differences were even more pronounced while watching positive emotional pictures. Findings were explained assuming the operation of two different, but related processes: transmarginal inhibition for the BIS; the attention/emotional gating mechanism regulating cortical sensory input for the FFFS trait. These findings appear consistent with predictions derived from the rRST, which traced fear and anxiety to separate but interacting neurobehavioural systems.

--S2 --Results for RST traits and LORETA source localization. Since we have found that N1/P2 complex is sensitive to individual differences in RST traits, we have also tested these differences on standardized sLORETA scores to find regional cortical localizations of significant difference on current density within the N1 and P2 time windows.
In terms of individual differences in BIS, we found that high BIS, compared to low BIS, had a significantly lower activity at 101 ms (i.e., a maximal positive t value in the time window of N1 wave) in the sublobar insula, fusiform and parahippocampal gyri (BA13, BA31, BA23) of the left hemisphere ( Supplementary Fig. S1-(a) and Supplementary Table S2). Interestingly, at 226 ms from auditory stimulus onset (P2 wave) high BIS had lower activity than low BIS participants in the left superior parietal lobule (BA7), a higher activity in the right orbital frontal cortex (BA11, OFC), and a reduced activity in the right inferior parietal lobule (BA40; Supplementary Fig. S1-(b) and Supplementary Table S2).
For the FFFS we found that high FFFS, compared to low FFFS, had a significantly lower activity at 226 ms (P2 wave) in the right posterior cingulate (BA23 and BA30) and in the left postcentral gyrus in the parietal lobe (BA7; Supplementary Fig. S2-(a) and Supplementary Table  S2).
-- Supplementary Fig. S1 --Supplementary Fig. S1. Panel (a): statistical parametric maps of sLORETA differences comparing high vs. low BIS groups (N1 wave). Note that a lower current density difference (yellow color) occurred in high BIS in the left sub-lobar insula, limbic parahippocampal gyrus, and right temporal fusiform gyrus (BA13, BA31, BA23, and BA6) at a time-frame of 101 ms (corrected threshold, p < .01). Panel (b): sLORETA difference maps comparing high vs. low BIS groups at 226 ms (P2 wave). A lower current density difference (light blue color) occurred in high BIS in the superior and inferior parietal lobule (BA7 and BA40), while a higher activity (yellow color) occurred in high BIS participants in the orbitofrontal gyrus (BA11).  Table S2--Supplementary Table S2. MNI coordinates and Brodmann areas (BA) of significant differences in current density of N1 and P2 ERP waves between high BIS (N=19) vs Low BIS (N=19), high FFFS (N=15) vs Low FFFS (N=18). ERPs were elicited by auditory tones delivered using the augmenting/reducing paradigm.  Table 4 online).

ERP X(MNI) Y(MNI) Z(MNI) t(max) BA
Finally, using sLORETA source localization method we found some overlapping significant differences in current densities within the N1 and P2 time windows between high and low BIS and  Table 4). On this basis, it is reasonable to conclude that the reduced activity we observed in the above mentioned structures reflects the reduced sensitivity (or enhanced avoidance) to intense stimuli. This reduced cortical activity, observed in high BIS and FFFS individuals is consistent with the RST 1 that predicts, enhanced avoidance to intense/negative stimulations in high BIS and FFFS individuals. In line with this, high levels of stress have been associated with a reduction of both prefrontal cortex (PFC) and rostral anterior cingulate cortex (ACC) activity 2 and with our prior observation using auditory startle stimuli that lower self-report fear was associated with larger P200 amplitude, and enhanced current density in the medial and superior frontal gyrus (BA6) 3 . The present observations are also consistent with previous fMRI findings in a study by Bishop,et al. 4 designed to evaluate the role of anxiety in the processing of threat-related distractors. These authors found that anxiety is associated with reduced activity in the rostral ACC and lateral PFC indicating a loss in top-down control over threat-related distractors. Interesting, we also found a higher activity in the right OFC in high BIS that low BIS participants (Supplementary  Table 4). We think that the activation of OFC reflects the activation of neurophysiologic processes geared to enhanced cognitive control in these individuals 5 .
Although LORETA provides good localization accuracy, a major limitation of the present study is in the fact that only 30 electrodes were used for source analysis. This reduces the spatial resolution, and with impaired spatial resolution, there is a smaller chance that LORETA will be able to separate two closely spaced sources. Thus, a greater number of recording electrodes is recommended in future studies to enhance spatial resolution. In conclusion, the present study served to test RST constructs of BIS, FFFS in terms of emotional modulation of auditory A/R and source localization of the ERP components, and indicated that these temperamental traits involve distinctive neurophysiological systems.

Procedure for the Selection of Emotional Pictures
In a pilot study, a total of 30 female (22-36 age range, M = 24.6, SD = 2.6 yrs) Psychology students rated each image on valence and arousal to verify the initial selection. The ratings for positive, negative and neutral valence were obtained using a 9-point Likert Scale, ranging from 1 to 9 that varies from negative to positive with a neutral point (five). A similar scale was used to rate arousal levels, ranging from 1 (calm) to 9 (aroused). We selected from the IAPS all the positive pictures that had a score equal or higher than 7 in valence and 7 in arousal. We followed the same procedure for the negative images. We took the neutral images in a small interval around the 5-value for the valence (4.5 to 5.5). Since neutral images are typically rated lower in arousal relative to positive or negative images, this was quite difficult to select, from IAPS, a set of neutral images with a high arousal level such as that of emotional images. To remedy this problem, we selected from the IAPS the neutral pictures with the highest score on arousal, but we had to add 18 neutral surrealistic, potentially high arousable pictures downloaded from the web, as suggested in a previous study by Mourao-Miranda, et al. 6  sLORETA is a method for estimating the localization of cortical generators at specific time windows by providing a solution to the inverse problem by assuming that neighboring grid points are more likely to be synchronized than grid points that are far from each other and to find the best solution that is consistent with the scalp distribution 7 . sLORETA is mainly used to compute statistical maps from EEG data that indicate the locations of the putative underlying source generators 8 . These maps are derived by performing a location-wise inverse weighting of the results of a minimum norm least squares analysis with their estimated variances. This tool performs source localization in 6239 cortical gray matter voxels sized 5 mm 3 rather than 7 mm 3 offered by the previous LORETA version, and localization inference is based on standardized values of the current density estimates 9 . The solution space of LORETA is restricted to cortical and some hippocampal and amygdala gray matter defined via a reference brain from the Brain Imaging Center at the Montreal Neurological Institute 10,11 (MNI). The sLORETA implementation incorporates a 3-D shell spherical head model registered to a recognized anatomical brain atlas 12 . Individual 3-D electrodes are positioned by the Talairach coordinate system according to the spatial association between anatomical brain landmarks and scalp positions 13 . sLORETA enables the computation of statistical maps from ERP components data that indicate the locations of the underlying source processes with low error 8 and does not require a priori hypotheses regarding field distribution of active sources. In the present experiment the coordinates of the 30 electrode positions were applied to a probabilistic anatomical template of the Talairach Atlas (McConnell Brain Imaging Centre, Montréal Neurological Institute, McGill University). These Talairach coordinates were used to compute the LORETA transformation matrix and then to transform ERPs of each subject into sLORETA forms. This resulted in the corresponding 3D cortical distribution of the electrical neuronal generators for each subject. For source reconstruction, subtractions of ERP traces between high and low personality traits were assessed, for the N1 and P2 waves, respectively within time intervals of 100-140 ms and 200-225 ms. Statistical significance was assessed using a non-parametric randomization test 14 . sLORETA maps of high vs low personality levels were compared using a t test for independent samples with the aforementioned nonparametric permutation test. It is important to note that this localization is not a complete listing of all significantly different cortical areas, but a listing of the local maxima of these differences. Although simulations have shown that LORETA localization performed better than some other localization methods, LORETA, as with all ERP localization algorithms, has accuracy limitations 15 . Consequently, the significant brain regional differences reported in the present study between high and low levels of personality traits were examined only when N1/P2 differences between those traits were seen statistically significant.