Searching for the inner self: evidence against a direct dependence of the self-prioritization effect on the ventro-medial prefrontal cortex

Abstract

The processing of self-referential material is supposed to be located in the medial prefrontal cortex (MPFC) and in particular in the ventro-medial prefrontal cortex (VMPFC). A reliable method to assess effects of self-relevance is the so-called matching paradigm in which the prioritization of newly learned self-associations in comparison to non-self-relevant associations can be measured. To assess the connection of activation in the VMPFC and self-referential processing, we measured the self-prioritization effect (SPE) before and after experimentally manipulating activation in the VMPFC. We applied either excitatory or inhibitory stimulation to the VMPFC via transcranial direct current stimulation (tDCS). In a sample of N = 65 healthy adults, we found a significant SPE before and after both types of stimulation and, remarkably, no systematic change of the SPE due to the stimulation. These results are evidential against a direct dependence of the SPE from activation in the VMPFC, indicating either that the SPE differs from other, more elaborate self-effects, and thereby is processed in different brain areas, or that the connection of SPE and VMPFC is correlational rather than causal.

Appendices

Appendix 1

See Table 1.

Table 1 RTs in milliseconds and absolute error rates as well as the sensitivity measure d’ are presented as a function of VMPFC stimulation condition, time of testing, matching condition, and shape association

Appendix 2

Statistical analyses with the non-adjusted SPE score revealed results which are congruent with the results of the analysis of the adjusted SPE. In RTs, also the non-adjusted SPE was overall significant, t(64) = 9.11, p < .001, d = 1.13. Further, a 2 (test: pre vs. post) × 2 (VMPFC stimulation: anodal vs. cathodal) repeated-measures MANOVA with the non-adjusted SPE revealed a significant main effect of VMPFC stimulation, F(1, 63) = 4.23, p = .044, η_p² = .06, but no significant main effect of test, F(1, 63) = 1.74, p = .192, η_p² = .03. Most important for the hypothesis and comparable to the analysis with the adjusted SPE, no significant interaction was found, F < 1. Supporting this result, a direct comparisons of the SPE before the stimulation (i.e., in the ‘pre’-condition) with the SPE after the stimulation separate for the two stimulation conditions revealed that neither in the anodal, nor in the cathodal condition the SPE varied significantly, both ts < 1 (Fig. 3).

Fig. 3

The non-adjusted SPE as a difference score between mean RTs (left) or d’ (right) in other- and self-relevant trials as a function of VMPFC stimulation and test. Error bars depict the standard error of the mean

In sensitivity rates, the non-adjusted SPE was overall significant, t(64) = 7.40, p < .001, d = 0.92. Further, a 2 (test: pre vs. post) × 2 (VMPFC stimulation: anodal vs. cathodal) repeated-measures MANOVA revealed no significant main effect, F(1, 63) = 1.87, p = .176, η_p² = 0.03, for test, F(1, 63) = 1.87, p = .176, η_p² = 0.03, for VMPFC stimulation. There was also no significant interaction, F(1, 63) = 1.22, p = 0.273, η_p² = 0.02. The comparison of SPE before the stimulation with the SPE after the stimulation separate for the two stimulation conditions revealed—comparable to the results with the adjusted SPE—that the SPE did not vary significantly in the cathodal condition, t < 1, but, however, was reduced in the anodal condition, t(34) = 2.41, p = .021, d = 0.41 (Fig. 3). As with the adjusted SPE, this reduction of the SPE—solely in d’—was not predicted and thus is dicussed.

The corresponding Bayesian repeated-measures ANOVA revealed that, for RTs the data were about 3.9 times as likely to have occurred under a null hypothesis for the interaction than under an alternative hypothesis, BF₀₁ = 3.90, and about 2.4 times as likely, BF₀₁ = 2.41, for d’ (JASP; Love et al. 2015). According to the rules of thumb given by Jeffreys (1961; see also Wagenmakers et al. 2011), this finding is interpreted as “substantial evidence” for the null hypothesis in RTs, and “anecdotal evidence” in d’.